Decreased frontal serotonin 5-HT 2a receptor binding index in deliberate self-harm patients

UNIVERSITEIT GENT

FACULTY OF MEDICINE & HEALTH SCIENCES

Functional Neuroimaging in Psychiatry:

a Functional Psychopathological approach

K. Audenaert

Thesis submitted in fulfilment of the requirements for the degree of

doctor in medical sciences

2001

Promotors: Prof. Dr. R.A. Dierckx

Prof. Dr. C. van Heeringen

3

Dankwoord

Ik had graag mijn erkentelijkheid uitgesproken naar mijn promotors, Prof. Dr.

Dierckx en Prof. Dr. van Heeringen, voor hun blijvende aanmoedigingen en

vertrouwen in mijn onderzoek. Het was een voorrecht te mogen werken in een sfeer

van grote openheid en wetenschappelijke vrijheid. Moge dit werk slechts het begin

zijn van een boeiende samenwerking. Graag ook een woord van dank naar het

personeel van de afdelingen Nucleaire Geneeskunde en Psychiatrie voor de

ondersteuning bij mijn onderzoekswerk. In het bijzonder wil ik Dr. Van de Wiele,

Dr. Van Laere, Dr. Brans en Dr. Ir. Lahorte bedanken voor hun geduld bij mijn

initiële pogingen tot pootje-baden in de technische materie van de nucleaire

geneeskunde en Prof. Dr. M. Vervaet voor de hulp bij het werk rond patiënten met

eetstoornissen. Uiteraard ook mijn erkentelijkheid naar Prof. Dr. Slegers en zijn

medewerkers van de afdeling Radiofarmacie voor de materiële ondersteuning en

prettige samenwerking. Ik heb ook goede herinneringen aan de aangename uitleg

van Prof. Dr. Mertens (VUB Cyclotron) en de fijne samenwerking met Prof. Dr.

Verschooten en zijn medewerkers van de Dienst Medische Beeldvorming uit de

Faculteit Diergeneeskunde, waarvoor mijn dank. Een bijzonder woord van dank aan

mijn grote vriend Dr. De Bacquer (Dienst Maatschappelijke Gezondheidkunde) voor

zijn inspanningen omtrent methodologie en supervisie van statistiek. Alleen

Bruichladdich 18 years old kan dit nog in evenwicht brengen, hoop ik...

Uiteraard mijn dank aan Dr. Peremans voor haar toewijding. Zelden beginnen mooie

verhalen met het voorstel tot hersenonderzoek-bij-bavianen…

In het bijzonder wil ik mijn ouders, familie en vrienden bedanken voor hun blijvende

steun en geduld tijdens de lange maanden waarin ik vaak afwezig was.

Ik wens dit werk op te dragen aan mijn patiënten.

5

Promotors: Prof. Dr. R.A. Dierckx

Prof. Dr. C. van Heeringen

Begeleidingscommissie : Dr. B. Brans Prof. Dr. R.A. Dierckx Dr. Apoth. F. Dumont Prof. Dr. C. van Heeringen

Dr. K. Van Laere Examencommissie: Prof. Dr. M. Mareel (voorz.) Prof. Dr. W. Buylaert

Prof. Dr. L. Crevits Prof. Dr. R.A. Dierckx Prof. Dr. P.P. De Deyn Prof. Dr. D’Haenen Prof. Dr. L. Pilowsky Prof. Dr. Apoth. Slegers Prof. Dr. C. van Heeringen

Prof. Dr. P. Van Rijk

ISBN 90-805-957-3-X

Vakgroep Psychiatrie en Medische Psychologie De Pintelaan, 185

B-9000 Gent België / Belgium

7

Table of contents

Introduction……………………………………………………………….……….. p. 9-22

Chapter 1: The classical Stroop Interference task as a prefrontal

activation probe: a validation study using 99mTc-ECD brain SPECT………………p. 23-46

Chapter 2: Verbal Fluency as a prefrontal activation probe: a

Validation study using 99mTc-ECD brain SPECT…………………………………..p. 47-68

Chapter 3: SPECT neuropsychological activation procedure with

the Verbal Fluency Test in depressed suicide attempters…………………………..p. 69-94

Chapter 4: Decreased frontal serotonin 5-HT2a receptor binding

index in deliberate self-harm patients………………………………………………p. 95-116

Chapter 5: Prefrontal 5-HT2a receptor binding index, hopelessness

and personality characteristics in attempted suicide……………………………….p. 117-136

Chapter 6: Decreased 5-HT2a receptor binding in patients with

anorexia nervosa…………………………………………………………………...p. 137-154

Epilogue……………………………………………………………….…………..p. 155-174

Samenvatting van de thesis………………………………………………………..p. 175-176

Résumé de these……….………………………………………………….……….p. 177-178

Summary of the thesis…………………………………………………….……….p. 179-180

9

We must recollect that all of our provisional ideas in psychology will presumably one day be based on an organic substructure.

Sigmund Freud, “On Narcissism” (1914)

INTRODUCTION

Biological research in psychiatry: not a success-story.

During the first decades of the twentieth century, the work of Freud was founding

the basis for our insights on the processes of mental life. But most important of

all, as is stated by Kandel, the year 2000 Nobel Prize of Medicine laureate, it

offered us concepts to understand the irrationality of human motivation [1].

Indeed, the work of Freud still represents the most exciting and intellectually

satisfying view of the mind. In his early writings, Freud argued that biology was

not advanced enough to be helpful to psychoanalysis at that time and that it was

premature to bring the two together. However, he expected physiology and

chemistry “to be of a kind that will blow away the whole of our artificial structure

of hypothesis” [2].

From the second half of the previous century on, the emerging biological and

psychopharmacological psychiatric research has sought for underlying organic

substructures for these mental processes and corresponding diseases. But, against

expectations, this has not been a success-story. And, to our belief, not only

because research tools were and still are too premature. Neuropsychological

research has offered us tools from the thorough evaluation of brain damaged

10

patients, through the designing of well validated test batteries, to the exciting

neuropsychological brain-activation functional imaging paradigms. A long

tradition of shallow neuropathology has nowadays been evolved to a precise

instrument that allows us to unravel the ultra-structure of the brain.

Electrophysiology has, thanks to the revolution in computer sciences, contributed

to a powerful tool that gives access to the functional processes of the brain.

Molecular genetics have recently revealed the structure of the human genome and,

in the slipstream of this large project, has given us access to the evaluation of

deviant alleles. Indirect biochemical research, such as the measurement of

neurotransmitter metabolites on cerebrospinal fluid or pharmacological challenge

studies, gave us the opportunity to estimate central nervous neurotransmitter

function. Post-mortem research with autoradiographic techniques permits us to

evaluate the neurotransmitter receptor status on the brains of patients with a

psychiatric disease. Functional neuroimaging techniques, such as SPECT/PET

and fMRI gave us direct access to the perfusion, metabolism and neurotransmitter

receptor status of the living brain.

But, despite all these well validated and powerful techniques, findings on the

biological substrate were far from unanimous and very few biological hypotheses

in psychiatry could stand thorough retesting. Therefore, reasons of this failure

must be sought elsewhere. At least partially, they might be related to the

inflexibility of the mind of the scientist, fostered by the lack of interdisciplinary

training, and by the scientific categorical method in psychiatric research itself.

At first, the inflexibility of the scientist’s mind led to a rigid way of tackling a

problem. Instead of searching for similarities between psychological and

biological hypotheses, researchers from both disciplines engaged in fiery debates

on the nature or nurture determinism on theories of the mind. Happily, in recent

years, the “either-or” model, evolved through a combined “constitution and

environment”-model into an “interactional” model. This interactional paradigm

interprets mental processes and diseases as a series of transactions between

organism and environment over time. By applying this model, neuroscience has

been able to fit valid models on, for example, post-traumatic stress disorder, that

combine research issues from psychoanalysis, cognitive psychology,

11

neuropsychological, neuropathology, neurochemistry and genetics. However, on

other psychiatric disorders, for instance mood disorders, no robust unifying

psychobiological models could be generated. This leads us to the second, the

scientific categorical method in psychiatric research. This will be discussed in the

next paragraph.

Diagnosis in psychiatry: matter of concern

A possible explanation for this failure could lie in the categorical nosological

model psychiatrists are trained for. In medical school, we learned to categorize

pathophysiological processes following a diagnostic system of discrete and

separable disorders, each with its own etiology, symptomatology and course. This

led to the oversimplified research hypothesis that each mental process and each

disorder has its specific psychological and biological substrate. It was the merit of

Van Praag to confront us with the curtailments of this categorical nosological

method in psychiatry and to present us a radically different diagnostic approach to

abnormal behaviour [3].

Diagnosis in psychiatry nowadays largely relies on a categorical classification of

diseases following the criteria of the Diagnostic and Statistical Manual of Mental

Disorders [4]. Undoubtedly, the introduction of DSM has contributed to a better

definition of nosological concepts and has offered a world-wide accepted

framework that enables a better communication between psychiatrists. However,

due to its rigid categorical definitions following a choice-principle (“five out of

nine symptoms”), it does not fit the reality of diagnosis in the psychiatric field.

For instance, in patients considered to suffer from the same psychiatric disease, a

diversity in symptoms is often striking. Furthermore, there is much symptom

overlap between patients, classified to be suffering from different diseases. We

try to circumvent this problem by introducing the terms “co-morbidity” and

“syndromal heterogeneity”. If we engage our biological research within this

nosological framework, van Praag confronts us with the thorny problem “How

can one study the biology of a given disorder, its epidemiology, its course, the

12

results of biological and psychological therapeutic interventions, if it more than

not appears hand in hand with several other diagnoses? Which of them are we

actually studying?” [3].

As a possible solution, van Praag stated that a comprehensive psychiatric

diagnosis should consist of a three-tiered structure. Tier one is the nosological

diagnosis representing no more than a broad general diagnostic outline, tier two

depicts the syndromal composition of the disorder and an assessment of other

relevant variables, such as severity, duration, course, and etiology on independent

axes and tier three represents the dissection of the syndrome into its basic

components - the psychological dysfunctions - and includes a detailed assessment

of functions that are disturbed and those that are still intact [3].

Can we, bearing in mind the proposition of Van Praag, apply this transnosological

and functional psychopathological approach to functional neuro-imaging in

psychiatry?

Within the framework of this thesis, we tried to apply this strategy in research to

patients with symptoms of cognitive and behavioural dysfunctions of impaired

planning and behavioural execution and impulsivity. In literature, we can find clues

on both the neuro-anatomical substrate and the biochemical processes that are

involved in these dysfunctions. This matter is shortly overviewed in the next two

paragraphs.

Issues of neuropsychological dysfunctions in brain functional imaging: a

functional-psychopathological approach and the formulation of the objectives of

the study.

In clinical practice, we are confronted with patients that, across diagnostic

boundaries, often show with disorders that are related to cognitive and behavioural

dysfunctions such as impaired planning and behavioural execution and impulsivity.

From the categorical point of view, based on the DSM classification, those patients

13

belong to a broad variety of nosological entities such as mood disorders, anxiety

disorders, eating disorders and personality disorders. From a neuropsychological

point of view, these afore-mentioned impairments have a basic component, i.e. they

can, at least partially, be attributed to dysfunctions in prefrontal cortex circuits.

This knowledge is based on several lines of evidence. In the early fifties, ablation

experiments in monkeys demonstrated a relationship between prefrontal damage and

increased impulsivity [5], inordinately behaviour or behavioural rigidity. But also

imaginative classical human case reports of patients with damage of the prefrontal

cortex described changes in behaviour towards impulsivity and absence of

behavioural insight and planning. One famous case was described by Damasio

(1984) where he commented on “Phineas Gage”, a nineteenth century docile and

industrious railway worker that was struck by a huge steel rod and thereby lost a

large chunk of forebrain. He miraculously survived the accident but turned into an

impulsive, drunken drifter that could no longer plan his life [6]. More systematic

studies in populations of violent offenders [7] or neurological patients with

irritability and impulsivity [8] demonstrated that there was a significant amount of

patients with lesions of the prefrontal cortex. More refined research demonstrated

that the prefrontal portion of the frontal lobes is subdivided, with different functions

mediated in different cortical regions. Defects in the control, regulation, planning

and integration of cognitive activities tend to predominate in patients with

dorsolateral lesions, i.e. lesions on the convexity of the frontal lobes, or in its

connections to other cortical or subcortical structures. Orbitomedial (cingulate)

lesions or lesions to its connections are most apt to affect drive and motivation.

Orbitofrontal lesions, i.e. lesions to the basal portion of the prefrontal cortex, or

lesions to its connections, play a key role in disinhibition or loss of impulse control

[9].

The integrity of the prefrontal functioning has been investigated by

neuropsychological research and by functional imaging modalities.

Neuropsychological research has developed tests that can be used to evaluate

prefrontal functioning. Among others, the Stroop Colour Word Test is used to

evaluate the capacity to suppress a habitual response in favour of an unusual one

[10]. A disturbed performance on this test can, at least partially, be attributed to

14

disinhibition. Reduced testing performance was found across the boundaries of

psychiatric and neurological diseases, such as subpopulations of patients with

depression, anxiety, Parkinson’s disease and Huntington’s disease [11]. Another

test, used to evaluate prefrontal functioning is the Verbal Fluency Test. This test is

used to evaluate initiation of action and planning of strategies. Reduced ability to

perform this test has also been demonstrated across the boundaries of psychiatric and

neurologic nosological entities, such as subpopulations of patients with depression,

psychotic disorders, Parkinson’s disease and Huntington’s disease [12]. Functional

imaging studies, with SPECT or PET, study the prefrontal cortex function through

the measurement of cortical blood perfusion or glucose metabolism. Both perfusion

and metabolism were demonstrated to be closely related to neuronal activity. Again,

across the boundaries of nosological categorical diagnosis, prefrontal

hypometabolism or hypoperfusion was found in subpopulations of depressed patients

[13,14], patients with schizophrenia [15] and eating disorders [16,17].

Neuropsychological research and functional imaging studies were combined in

neuropsychological activation studies with PET, fMRI or SPECT as imaging tools.

Hereby, subjects are scanned under two or more conditions, a basal or resting

condition and one or more activated conditions. Comparison of the images that were

acquired during the resting condition with those acquired during the activation (here:

a neuropsychological test) can give information about the prefrontal activity elicited

during the task performance, and hence offer a tool to evaluate the prefrontal reserve

capacity. Patients with impaired prefrontal functions are expected to have a blunted

increase in prefrontal activity, measured through a blunted increase in prefrontal

blood perfusion.

The vast majority of these neuropsychological activation studies were carried out

with PET or fMRI. These techniques offer an excellent image quality due to their

high spatial or time resolution and offer the opportunity of repeat-condition

measurements due to the physical properties of the tracers. Up to now, SPECT

neuropsychological activation studies were sparse due to their lower spatial

resolution and to their operator-prone bias in image processing. However, if one can

partially circumvent these problems, SPECT neuroactivation has one large advantage

on PET or fMRI paradigms. Due to the physical properties of the perfusion tracers

15

used in SPECT, one can present the task in standardized neuropsychological test

conditions i.e. sitting at a table in a quiet room, hereby introducing as less as possible

testing bias, especially in patient subjects. Conventional PET and fMRI studies do

require testing in supine position, with the head of the patient lying in the camera and

hereby requiring changes in the test procedure.

The first general objective of this thesis is to evaluate the contribution of functional

brain imaging with SPECT in disorders where impulsivity and impaired executive

functioning are core symptoms. This resulted in the first major corpus of the thesis,

comprising the first three chapters dealing with more specific objectives:

Firstly, we evaluated the feasibility of the Stroop Colour Word Test as a

neuropsychological activation paradigm in healthy volunteers. The results of

this study are described in chapter 1.

Secondly, we evaluated the Verbal Fluency test, both in the letter fluency as

in the category fluency modality in healthy volunteers. The results of this

study are described in chapter 2.

Thirdly, we evaluated the Verbal fluency test in a population of depressed

patients that recently attempted suicide. The results of this study are

described in chapter 3.

Psychopathological dysfunctions and the consequence on biochemical

parameters and psychopharmacology: serotonergic dysfunctions and the

objectives of the study

Van Praag states that in “biological psychiatry, the limitations of the nosological

approach were felt most dramatically since biological data have yielded little, if any,

practical diagnostic significance, despite years of intensive research efforts”. He

considered the categorical approach as the most plausible cause of the lack in

uniformity in biological psychiatric research [3]

16

Indeed, deficiencies in different neurotransmitter systems are present in different

subpopulations of patients of the same diagnostic category and similar deficiencies of

neurotransmitter systems are present in different nosological entities. Maybe this

matter is best illustrated through the classification of psychotropic drugs which is

nosologically based. For instance, anti-depressants are indeed effective in

depression, but likewise in anxiety disorders, eating disorders and impulsivity

disorders. Moreover, within one diagnostic category such as depression, anti-

depressants acting on differing neurotransmitter systems have variable outcome

results in differing subpopulations of the diagnostic category.

Impulsivity and reduced planning activity is an aspect of borderline and antisocial

personality disorders, of the attention and hyperactivity disorder (ADHD) in children

and is associated with alcoholism, substance abuse, anorexia nervosa, bulimia and

suicidal and violent behaviours [18] and is present in a subpopulation of depressed

patients and patients with anxiety disorders.

Different approaches to investigate the involvement of serotonin as a modulating

factor in impulsive and dysexecutive behaviour are available. Indirect techniques

focus on the estimation of the serotonergic central brain function through dietary

alterations, through measurements on blood or cerebro-spinal fluid or through

neuropharmacological challenge tests. Direct techniques allow a straightforward

evaluation of the serotonin metabolism or serotonin receptor binding status through

post-mortem research or through functional neuroimaging studies and ligands with 5-

HT receptor affinity.

In 1976, Asberg et al. reported that depressed patients with reduced cerebrospinal

fluid concentrations of 5-hydroxyindoleacetic acid (CSF 5-HIAA), the major 5-HT

metabolite, were significantly more likely to have had histories of suicide attempts

than similar depressed patients with CSF 5-HIAA concentrations in the same range

as normal, healthy volunteers [19]. A remarkable study on male navy recruits

described a trivariate relationship between a life history of aggression and

impulsivity, a life history of suicidal behavior, and reduced CSF 5-HIAA

concentration, suggesting that individuals with reduced CSF 5-HIAA concentration

were at risk for both impulsive and suicidal behavior [20]. These findings were

extended in the early 80s by the Finnish group of Linnoila et al. (1983) to violent

17

criminals [21] and impulsive arsonists [22] and to patients with anorexia nervosa

[23].

Psychopharmacological challenge studies, using serotonin agonists such as

fenfluramine and m-CPP, demonstrated a blunted rise in prolactin and growth

hormone in patients with impulsive suicide attempts [24], with impulsive behavior

towards others [25] and with anorexia nervosa [26], indicating a disturbance in

central serotonergic neurotransmission. The dysfunction in the serotonergic system

could be demonstrated in direct ex vivo research by demonstrating a reduced number

of pre-synaptic 5-HT transporter sites in frontal cortex in violent suicide victims

compared with accident victims [27].

Looking further into possible disturbances on serotonergic receptor level, shows that

animal research demonstrated abnormalities in the numbers of 5-HT-2a receptors

associated with aggressive behaviour in primates [28], as well as in post-mortem

studies of suicide victims [29].

Van Praag et al suggests to interpret signs of diminished serotonin as not being

disorder specific, but rather related to psychopathological dimensions such as

aggression and anxiety, independent of the nosological framework in which these

dysfunctions occur [30].

In addition to this suggestion, another striking observation, in favour of a functional

psychopathological approach to the pathology of impulsive and non-planning

behaviour, lies in the uniform therapeutic strategy suggested in diverse nosological

entities such as depression, anxiety disorders, impulsivity disorders, eating disorders,

etc. They are all treated with drugs that have serotonergic properties.

Functional nuclear imaging with PET or SPECT, using receptor-specific

radioligands, allows an in vivo estimation of brain receptor binding index. The

choice of the radiotracer was directed by literature. The involvement of the

serotonergic system in the pathophysiology of some neuropsychiatric disorders in

general, and in disorders related to impulsivity in particular, is largely demonstrated

through indirect studies. Studies on serotonergic receptors were mostly directed to

the sertotonin-2a receptors and to the serotonin-1a receptor. These were studied

18

through serotonin-2a blood platelet receptors and through neuroendocrine challenge

tests that mediate post-synaptic endocrine effects, merely through binding of

serotonin on the serotonin-1a and 2a receptor. Since these studies only indirectly

estimated central nervous serotonin receptor function, we planned to directly and in

vivo evaluate the aforementioned receptors through functional imaging research.

Since a radio-iodine labelled tracer for the 5-HT2a receptor was available at the Free

University in Brussels (Prof. Mertens) and at the University of Gent (Prof. Slegers),

we choose this tracer for research purposes.

There only exist few radioligands that are highly specific to the serotonin-2a

receptor. 18F-setoperone and 18F-altanserine, designed as PET-tracers, were

validated in healthy volunteers and tested in patients with major depression [31,32].

A highly specific serotonin-2a radioligand was synthetised at the VUB Brussels [33],

tested in non-human primates [34] and human primates [35], and investigated

concerning its gender and age specific characteristics [36].

The second general objective of this thesis is to evaluate the contribution of

functional brain imaging with SPECT using 123I-5-I-R91150 in clinical populations

and to search for possible correlations between regional binding indices with

psychological personality variables [37,38]. This resulted in the second major

corpus of the thesis, comprising the chapters four to six dealing with more specific

objectives:

Firstly, we evaluated the serotonin-2a binding index in patients that recently

attempted suicide. We further evaluated possible differences in binding index

between patients with deliberate self-poisoning versus deliberate self-injury.

Secondly, we evaluated a possible relation between personality variables and

the serotonin-2a binding index in patients that recently attempted suicide.

The results of this study are described in chapter 5.

Thirdly, we evaluated the serotonin-2a binding index in patients with

anorexia nervosa. The results of this study are described in chapter 6.

19

References

1. Kandel E. Biology and the future of psychoanalysis: an new intellectual framework for

psychiatry revisited. Am J Psychiatry 1999; 156: 505-524.

2. Freud S. Beyond the pleasure principle (1920). London: W. W. Norton & Company,

1975.

3. Van Praag H. Make-Believes in Psychiatry or The perils of progress. New York:

Brünner-Mazel, 1992.

4. Diagnostic and Statistical Manual of Mental Disorders: DSM-IV. Washington: American

Psychiatric Association, 1994.

5. Pribram K, Bragshaw M. Further analysis of the temporal lobe syndrome utilizing

frontotemporal ablation. J Comparative Neurology 1953; 99: 347-375.

6. Damasio H, Grabowski T, Frank R, Galaburda A, Damasio A. The return of Phineas

Gage: The skull of a famous patient yields clues about the brain. Science 1994; 264:

1102-1105.

7. Raine A, Buchsbaum M, Stanley J, Lottenberg S, Abel L, Stoddard J. Selective

reductions in prefrontal glucose-metabolism in murderers. Biol Psychiatry 1994; 36:

365-373.

8. Bear D. Neurological perspectives on aggression. J Neuropsychiatry Clin Neurosc 1991;

3 (suppl. 1): 3-8.

9. Lezak M. Neuropsychological assessment. New York : Oxford University Press, 1995.

10. Spreen O, Strauss E. A compendium of neuropsychological tests. New York / Oxford:

Oxford University Press, 1997, pp. 171-218.

11. Lezak M. Neuropsychological assessment. New York : Oxford University Press, 1995.


Oxford University Press, 1997, pp. 447-464.

13. Baxter L, Schwartz J, Phelps E, Mazziota J, Guze B, Selin S, Gerner R, Sumida R.

Reduction of prefrontal cortex glucose metabolism common to three types of depression.

Arch Gen Psychiatry 1989; 46: 243-250.

14. Drevets W. Functional neuroimaging studies of depression: the anatomy of melancholia.

Ann Rev Med 1999; 49: 837-844.

20

15. Sabri O, Erkwoh R, Schreckenberger M, Dickmann C, Schulz G, Kaiser H, Buell U,

Sass H. 99m-Tc HMPAO brain SPECT and psychopathology in drug-naive

schizophrenic patients during the active phase and after treatment. In: De Deyn P,

Dierckx R, Alavi A, Pickut B, eds. SPECT in neurology and psychiatry. London: John

Libbey & Company Ltd.; 1997: pp. 133-137.

16. Delvenne V, Goldman D, De Maertelaer V, Wikler D, Damhaut P, Lotstra F. Brain

glucose metabolism in anorexia nervosa and affective disorders: influence of weight loss

or depressive symptomatology. Psychiatry Res 1997; 74: 83-92.

17. Delvenne V, Goldman D, Simon Y, De Maertelaer V, Lotstra F. Brain hypometabolism

of glucose in bulimia nervosa. Int J Eating Dis 1997; 21: 313-320.

18. Plutchik R, Van Praag H. The nature of impulsivity: definitions, ontology, genetics, and

relations to aggression. In: Hollander E, Stein D, eds. Impulsivity and aggression.

1995: pp. 7-24.

19. Asberg M, Traskman L, Thoren P. 5-HIAA in the cerebrospinal fluid: a biochemical

suicide predictor? Arch Gen Psychiatry 1976; 33: 1193-1197.

20. Brown G, Goodwin F, Ballenger J, Goyer P, Major L. Aggression in humans correlates

with cerebrospinal-fluid amine metabolites. Psychiatry Res 1979; 1: 131-139.

21. Linnoila M, Virkkunen M, Scheinin M, Nuutila A, Rimon R, Goodwin FK. Low

cerebrospinal fluid 5-hydroxyindolacetic acid concentration differentiates impulsive

form nonimpulsive violent behavior. Life Sci 1983; 33: 2609-2614.

22. Virkkunen M, Nuutila A, Goodwin F, Linnoila M. Cerebrospinal fluid metabolite levels

in male arsonists. Arch Gen Psychiatry 1987; 44: 241-247.

23. Kaye W, Gwirtsman H, George D, Ebert M. Altered serotonin activity in anorexia

nervosa after long term weight restoration: Does elevated cerebrospinal fluid 5-

hydroxyindolacetic acid level correlate with rigid and obsessive behaviour? Arch Gen

Psychiatry 1991; 48: 556-562.

24. Coccaro E, Siever L, Klar H, Maurer G, Cochrane K, Cooper T, Mohs R, Davis K.

Serotonergic studies in patients with affective and personality disorders: correlates with

suicidal and impulsive aggressive behavior. Arch Gen Psychiatry 1989; 46: 587-599.

25. Coccaro E, Kavoussi R, Cooper T, Hauger R. Central serotonin activity and aggression:

inverse relationship with prolactin response to d-fenfluramine, but not CSF 5-HIAA

concentration, in human subjects. Am J Psychiatry 1997; 154: 1430-1435.

21

26. Brewerton T, Jimerson D. Studies of serotonin function in anorexia nervosa. Psychiatry

Res 1996; 62: 31-42.

27. Stanley M, Virgillio S, Gershon S. Tritiated imipramine binding sites are decreased in

the frontal cortex of suicides. Science 1982; 216: 1337-1339.

28. Raleigh M, Brammer G. Individual differenced in serotonin-2 receptors and social

behavior in monkeys. Society for Neuroscience Abstracts 1993; 19: 592.

29. Arango V, Ernsberger P, Marzuk PM. Autoradiographic demonstration of increased

serotonin 5HT2 and beta-adrenergic receptor binding sites in the brain of suicide

victims. Arch Gen Psychiatry 1990; 47: 1038-1044.

30. Van Praag H, Asnis G, Kahn RS, Brown S, Korn M, Harkavy-Friedman J, Wetzler S.

Monoamines and abnormal behavior: a multi-aminergic perspective. Br J Psychiatry

1990; 157: 723-734.

31. Biver F, Wikler D, Lotstra F, Damhaut P, Goldman S, Mendlewicz J. Serotonin 5-HT2

receptor imaging in major depression: focal changes in orbito-insular cortex. Br J

Psychiatry 1997; 171: 444-448.

32. Attar-Lévy D, Martinot J, Blin J, Dao-Castellana M, Crouzel C, Mazoyer B, Poirier M,

Bourdel M, Aymard N, Syrota A, Féline A. The cortical serotonin-2 receptors studied

with positron-emission tomography and 18F-setorperone during depressive illness and

antidepressant treatment with clomipramine. Biol Psychiatry 1999; 45: 180-186.

33. Mertens J, Terriere D, Sipido V, Van Gommeren W, Janssen P, Leysen J.

Radiosynthesis of a new radio-iodinated ligand for serotonin-5HT(2)-receptors, a

promising tracer for gamma-emission tomography. J Lab Comp Radiopharm 1994; 34:

795-806.

34. Abi-Dargham A, Zea Ponce Y, Terriere D, Al Tikriti M, Baldwin R, Hoffer P, Charney

D, Leysen J, Laruelle M, Mertens J, Innis R. Preclinical evaluation of [I-123] R93274 as

a SPECT radiotracer for imaging 5-HT2a receptors. Eur J Pharmacol 1997; 321: 285-

293.

35. Busatto G, Pilowsky L, Costa D, Mertens J, Terriere D, Ell P, Mulligan R, Travis M,

Leysen J, Gacinovic S, Waddington W, Lingford-Hughes A, Kerwin R. Initial evaluation

of 123I-5-I-R91150, a selective 5-HT2A ligand for single-photon emission tomography,

in healthy subjects. Eur J Nucl Med 1997; 24: 119-124.

36. Baeken C, D'Haenen H, Flamen P, Mertens J, Terriere D, Chavatte K, Boumon R,

Bossuyt A. 123I-5-I-R91150, a new single-photon emission tomography ligand for 5-

22

HT2a receptors: influence of age and gender in healthy subjects. Eur J Nucl Med 1998;

25: 1617-1622.

37. Cloninger C, Svrakic D, Pryzbeck T. A psychobiological model of temperament and

character. Arch Gen Psychiatry 1993; 30: 975-990.

38. Cloninger C. The genetics and psychobiology of the seven-factor model of personality.

In: Silk K, eds. Biology of Personality Disorders. Washington: American Psychiatry

Press; 1998: 66-93.

23

CHAPTER 1:

THE CLASSICAL STROOP INTERFERENCE TASK AS A PREFRONTAL

ACTIVATION PROBE: A VALIDATION STUDY USING 99MTC-ECD BRAIN

SPECT.

Kurt Audenaert, Philippe Lahorte , Boudewijn Brans, Koen Van Laere, Ingeborg

Goethals, Kees van Heeringen, Rudi Dierckx

(Nuclear Medicine Communications, 2001; 22(2): 145-150)

Summary

This study aims at investigating the feasibility of brain SPECT functional imaging in

a neuropsychological test setting, following a single-day protocol with a split-dose

paradigm.

The Stroop Color Word Test (SCWT) is an example of a well-documented prefrontal

activation task. In a split-dose protocol, ten right-handed healthy volunteers were

injected twice with 370 MBq Technetium-99m-ethyl cysteinate dimer (ECD) while

performing consecutively both series of card-reading of the SCWT. Images were

reconstructed using filtered back-projection and normalized to a standard template in

Talairach coordinates. Statistical Parametric Mapping (SPM96) was used to

determine voxelwise significant changes. A first activation cluster was found in the

left medial prefrontal cortex, consisting of the gyrus cinguli anterior and the gyrus

frontalis medius and superior. A second activation cluster included the right gyrus

frontalis dorsalis and medius. These findings confirm to a large extent the results of

previous fMRI and PET studies of Stroop-like tasks. The choice and validity of

various methodological characteristics of the experimental design leading to these

24

results is critically discussed. It is concluded that brain SPECT activation with the

Stroop Color Word Test under standard neuropsychological conditions in healthy

volunteers, is both technically and practically feasible.

Introduction

Focused attention is the capacity to highlight one or two important stimuli being

dealt with, while simultaneously suppressing awareness of competing distractions

[1]. This ability helps higher organisms to select relevant information for processing

and to ignore irrelevant material. In a theoretical model on human attention, Shallice

postulated two adaptive mechanisms to regulate the selection mechanism: contention

scheduling for routine behavior and supervisory attentional control for non-routine

situations [2]. The latter may be viewed as a voluntary indirect modulation of the

excitability of schemas, most likely by inhibition [3]. Protection against interference

from both inside and outside the organism is thought to be mediated by the frontal

cortex [4].

However, it is possible for unwanted schemas to intrude in spite of the supervisory

attentional control, as may be seen in the classical Stroop test by errors and by the

increase in reaction time. In 1935 Stroop reported that it took subjects longer to name

the color of the ink that color words (more particularly, the words red, green, yellow

and blue) were written in when the ink color and the color word did not match (e.g.,

the word red written in blue ink, the correct answer being blue) than it did for them

to simply name the color of colored squares [5]. The Stroop Color Word Test

(SCWT) was developed as a neuropsychological test to assess the ease with which a

person can shift his or her perceptual set to conform to changing demands and

suppress a habitual response in favor of an unusual one [6]. The classic Stroop

interference effect can be defined as a decrease in performance efficiency, and hence

an increase in reaction time, when subjects are required to suppress a concurrent

competing response [5]. The inability to perform the SCWT has been described in

patients with head injury [7], with both left [8] and right [4] frontal lesions, in

Parkinson’s disease and in Huntington’s disease [9], in HIV-infection [10] and in

25

patients suffering from schizophrenia [11], depression [12] and obsessive-

compulsive disorder [13].

Functional imaging techniques such as functional magnetic resonance imaging

(fMRI), positron emission tomography (PET) and also single photon emission

computed tomography (SPECT) allow investigation of motor, sensory, visual,

auditory and cognitive brain systems [14]. PET and fMRI research strategies have

included “activation” by means of a cognitive task during the scanning procedure

[15]. By consequence, the conditions in which the test is performed using PET or

fMRI do not fit the classical neuropsychological test conditions of sitting at a table or

before a computer screen in a quiet room. Because of this incompatibility, the

classical and well-validated test presentations commonly need to be adapted in order

to fit PET or fMRI test conditions.

The characteristics of perfusion tracers used in SPECT allow the use of the original

and habitual paper version of the SCWT. Distribution of Technetium-99m-ethyl

cysteinate dimer (ECD) intracellular uptake takes place almost pro rata the

microperfusion over approximately a two minutes’ time window. Subsequent

scanning of the patient can be performed over four to six hours in a flexible manner

and will reflect the cognitive activation state at the time of injection.

The current experiment is part of a larger project aimed at developing a

neuropsychological prefrontal activation probe, that is suitable for research purposes

in normal volunteers and for the clinical evaluation of patients with neuropsychiatric

diseases using the widely available SPECT imaging technique. Thus, the objectives

of this experiment were twofold: first, the study aimed to evaluate the feasibility of

SPECT brain perfusion following a split-dose paradigm in a neuropsychological

prefrontal activation test. As it was demonstrated that patterns of activation crucially

depend upon experimental parameters, a second aim was to investigate the functional

imaging of the Stroop Color Word Test in neuropsychological test conditions in

healthy subjects and compare the results of activation with those in the literature on

PET or fMRI functional imaging with Stroop-like paradigms.

26

Methods

Subjects

Ten healthy volunteers (9 females, 1 male) with a mean age of 24.8 years (SD 7.5,

range 21 to 37 years) were included in the study. None of the participants had a

history of major medical or neuropsychiatric diseases, including attention deficit

disorder or previous head injury. No subject took any medication with the exception

of oral contraceptives. Research was compliant with the Code of Ethics of the World

Medical Association (Declaration of Helsinki). Informed consent was given by each

subject following the guidelines of the local ethics committee.

All subjects were right-handed as assessed by the Edinburgh Handedness Inventory

[16]. All of them had normal or corrected-to-normal vision and were not color-blind.

Subjects had a mean 13,9 years of education (SD 1,9). No subject had any previous

experience with the SCWT.

Stroop Color Word Test (SCWT)

The original Stroop test, as designed by Stroop in 1935, consisted of three cards,

each with ten rows of five items [5]. The Dutch version of the SCWT [17] consists of

three cards (21 cm x 29.5 cm), each card showing 10 rows of 10 items.

Card I consists of 100 names of the colors red, green, blue and yellow in the Dutch

language printed in black ink and distributed at random. The patient is asked to read

aloud the 100 words of card I as quickly and with as few errors as possible. The

reading of this series of words is considered a solicitation of the reading reflex

preceding the baseline (card II) and activation (card III) condition. Card II displays

100 rectangles, printed in red, green, blue or yellow ink, also distributed at random,

while card III shows 100 names of colors printed in a conflicting ink color as

described above.

27

Figure 1

Stroop Color Word Test. Excerpt of the Colored card (II) and the Colored Word card (III)

green red blue yellow

red green yellow blue

green blue blue red

29

The only modification to the classical neuropsychological test conditions in our

experiment, is the length of time subjects had to read both cards. In the classical

Stroop test, subjects have to read a 100 items card and the required time is measured.

In the present experiment, subjects had to read as many items as possible in a 240 s

period from a 100 item card. Two experimental factors could potentially influence

the subjects’ tempo in an opposite way. On the one hand, the effect of tiredness

might result in a gradual deceleration of reading speed towards the end of the 4 min

reading period. In contrast, the effect of practice could lead to a higher tempo in the

reading of items in the complete 240 s period versus the reading of the first 100

items. The global tempo effect of the 240 s reading period of our experiment versus

the 100 item reading in the classical Stroop conditions was assessed by comparing

the measured mean time needed to read the first 100 items versus the calculated time

needed to read 100 items, assuming a constant tempo in the 240 s item reading.

Paradigm design

The split-dose paradigm was used since it was demonstrated that a brain blood flow

tracer can be given in a split-dose protocol enabling at least one repeat scan within a

short time period [18, 19]. Ten healthy volunteers were injected intravenously two

times 370 MBq (10 mCi) 99mTc-ECD following a split-dose paradigm.

30

Figure 2

Single-day split dose paradigm design. Along a time axis, time of reading card II and card

III, injection time and dose and time of acquisition 1 and 2 are marked

According to standard neuropsychological testing conditions, subjects sat at a table

in a quiet room. An intravenous line was installed and after approximately ten

minutes, the SCWT Color Card (SCWT-Card II) was presented and subjects were

instructed to read aloud the colors of the 100 rectangles for a 240 s period as quickly

as possible, however at such a pace that the correct answer was produced as often as

possible. Subjects were asked if they understood the procedure since they could not

ask any questions during the 240 s experiment. After the 30 s period of reading, 370

MBq 99mTc-ECD was injected via the intravenous line and subjects kept on reading

for another 210 s period. Immediately following the reading session, the first image

acquisition was performed during a 20 min period. Subsequently, subjects were

presented the SCWT Color Word Card (SCWT-Card III) and the aforementioned

procedure was repeated.

Corrected and uncorrected errors in the two test conditions were assessed by the

experimenter, since errors decrease the reaction time and hence decrease the

cognitive effort of the subject. To avoid auditory activation, subjects were not

Reading Card II

Acquisition 1

Reading Card III

30” 3’ 30”

370 MBq

99m Tc ECD

370 MBq

99m Tc ECD

30” 3’ 30”

Acquisition

2

20’ 20’

Total time ± 60’

31

corrected when “reading” a wrong color. A Stroop test performance was “normal”

when the calculated time for reading 100 items was below 61s for the SCWT-Card

II and below 97 s for the SCWT-Card III, i.e. below the reading of 395 items in 240 s

for the SCWT-Card II and below 247 items for the SCWT-Card III. These cut-off

values were set on the 20th percentile of normative scores in a population aged 22-37

y, tested with the Dutch version of the SCWT. The difference in numbers of named

colors of card II minus card III, called the interference score, is an indication for the

ability of the subject to adapt to changing task demands and is a measure of

resistance to perceptual interference when accomplishing a task.

Image acquisition and reconstruction

Image acquisition was performed with a three-headed gamma camera (Toshiba GCA

9300 A) equipped with parallel-hole, low energy high resolution collimators. A 360°

arc consisted of 6° steps, taking 60” each, in a step-and-shoot mode with a 1.5 zoom.

Projections were reconstructed using filtered back-projection with a Butterworth

filter of order 8. Uniform attenuation correction was applied with automatic

definition of the sinogram border (Sörensen correction, mean attenuation coefficient

0.12 cm-1). Reconstructed images were transferred in Interfile format onto a central

image processing system (HERMES, Nuclear Diagnostics, Sweden). The

reconstructed data were fitted automatically onto an in-house constructed database

template positioned in Talairach co-ordinates [20]. This template consisted of 21

healthy volunteers with the same age distribution as the participants. The template

was constructed with measurements on the same equipment, total dose and under the

same standard circumstances and was reconstructed as given above. The fitting

procedure was performed with 9 parameters (shift, rotation and scale) using a

principal axis transform and a count-difference minimization algorithm (fit threshold

of 0.50; BRASS, Nuclear Diagnostics, Sweden) [21].

Statistical Parametric Mapping

To apply statistical parametric mapping within SPM96 (Wellcome Department of

Cognitive Neurology, Institute of Neurology, London, UK), images were converted

32

into ANALYZE format by means of an in-house file format conversion program

(MEDCON). All SPM calculations were performed in Matlab 4.2 (The MathWorks,

Natick, MA) on a SUN SPARC 10 computer (Sun Microsystems Europe Inc,

Brussels, Belgium). Normalization was performed by applying the transformation

parameters of the SPECT perfusion template onto a PET SPM96 template smoothed

isotropically with a 14-mm gaussian kernel. This normalization was done using a 9-

parameter transformation (shift, rotation and scale). No affine (shear) parameters

were used because of the relatively poor spatial resolution of SPECT data.

Differences in global activity between scans were removed by scaling the activity in

each pixel proportional to the global activity. Mean global activity of each scan was

adjusted to 50. The threshold value for gray matter was put at 0.40 and a resulting

voxel size of 3x3x3 mm was used. The normalized studies were smoothed with an

isotropic 14-mm kernel to account for individual variability in structure-function

relation and to improve signal-to-noise ratio [22]. For determination of significant

sites of increased or decreased perfusion, a categorical multi-subject, multiple

condition model was used.

Planned comparisons between conditions were performed on a pixel-by-pixel basis

by t statistics, generating SPM(t) maps subsequently transformed to the unit normal

distribution SPM(Z) maps. We investigated areas at p=0.05, setting the height

threshold to p = 0.01 , without correction for multiple comparisons (anatomical pre-

existing hypothesis in prefrontal lobe).

Statistical analysis

The expected Stroop interference effect and possible timing interference were

measured in a two-factor within-subject analysis using a non-parametric Wilcoxon

Signed Rank Test (SPPS v 7.5 for Windows, SPSS Inc, Heverlee, Belgium).

33

Results

Behavioral results

A significantly (z=-2.80; p =0.005) larger number of items was read in 240 s when

reading the color card (mean 461 items, SD 52) than when reading the color word

card (mean 336 items, SD 49), indicating a Stroop interference effect when reading

the color word card.

There was no statistically significant difference between the measured (mean 52 s;

SD 5) and the calculated time (mean 52 s; SD 6) needed for the 100 item reading of

the color card (z=-0.87; p=0.39) and between the measured (mean 72 s; SD 10) and

the calculated time (mean 73 s; SD 10) needed for the 100 item reading of the color

word card (z=-0.91; p=0.36), suggesting no altered tempo effect in any of the two

cards in our experiment versus the classical Stroop test conditions.

Brain SPECT flow results

For the activation contrast, two clusters were found, indicating an increase in

regional blood flow. The first cluster was situated in the left medial wall of the

prefrontal cortex and comprised the anterior part of the cingulate gyrus (BA 32)

(Talairach coordinates = (-21, 21, 30)), the medial frontal gyrus (BA 8/9) (-15, 27,

36) and the superior frontal gyrus (BA 10) (-18, 39, 18). The second cluster was

situated in the right dorsal frontal cortex and comprised part of the supplementary

motor area (BA 6) (6, 6, 57) and the more anteriorly situated BA 8 (33, 27, 39) No

significant deactivation clusters were found at the specified thresholds.

34

Figure 3

Glass-brain image indicating projections of activation clusters in the Stroop test

(reading color word card versus reading color card)

Talairach coordinates

Cluster level P value x y z Region

0.067 -21 21 30 L. gyrus cinguli ant. (BA 32)

-15 27 38 L. gyrus frontalis med. (BA 8)

-18 39 18 L. gyrus frontalis sup. (BA 10)

0.054 6 6 57 R. gyrus frontalis dorsalis (BA 6)

21 9 51 R. gyrus frontalis dorsalis (BA 6)

33 27 39 R. gyrus frontalis medius (BA 8)

Table 1

Statistical Parametric Mapping: Cluster level P values and activated regions during Stroop card reading

35

Discussion

Findings of the neuropsychological activation experiment

As presented in Table 2, previous functional imaging studies report a great diversity

in activation foci resulting from Stroop-like activation paradigms. These studies

represent a large variety in imaging modalities, experimental paradigms and

processing and analysis methods. Common in all studies however, is an activation of

the gyrus cinguli anterior and other prefrontal regions.

The results presented in this work are to a large extent in accordance with previous

findings in both lesion studies in humans and studies with PET and fMRI using

alternative interference tasks.

Frontal cortex Study Methods N

Gyrus Cinguli Other frontal regions Other Brain regions

Pardo et al

(1990)

H2[15O] – PET

Stroop-like

8

R/L G Cing ant

L Promotor area

L Suppl Motor Area

R & L Peristriatum

R Gyr Temp sup

L Sulcus Postcentralis

L Putamen

Bench et al

(1993)

(Exp 1/2)

[15C]O2 – PET

Stroop-like

12 R Gyr Cing

-------------------

R Gyr Cing ant

R Orbitofrontal

-------------------------

R Frontal polar

R/L Gyr Pariet inf

--------------------------

Taylor et al

(1993)

H2[15O] – PET

Stroop-like

8 L Sulc Cing R Gyr Front inf

Bush et al

(1998)

f MRI

Counting Stroop

9 R Gyr Cing ant R/L Gyr Front med

L Gyr Precentralis

L Premotor cortex

R Gyr Temp inf

L Gyr Pariet sup

This work 99mTc-ECD– SPET

Stroop

10 L Gyr Cing ant R/L Gyr. Front med

R Gyr Front dors

L Gyr Front sup

Table 2

Regions activated in Stroop and Stroop-like activation studies

36

Neuropsychological evaluation in brain-damaged patients points unequivocally

towards involvement of the prefrontal cortex as the critical brain structure involved

in the Stroop task. Perret reported that the interference effect in the Stroop test was

greater for patients with left frontal damage than for other patients or control groups

[8]. Using MRI images of patients with selective frontal lesions, Vendrell et al.

concluded that the right prefrontal lateral cortex was the only region consistently

related to the Stroop effect [4].

In our study, a first reported activation cluster is located in the left medial prefrontal

cortex, corresponding to three functionally separate regions (see Table 1 and Figure

3): left gyrus cinguli anterior (BA 32), left gyrus frontalis medius (BA 8) and left

gyrus frontalis superior (BA 10).

Of these, the gyrus cinguli is discussed first. Using perfusion H215O PET with a

stimulus-response conflict paradigm related to the Stroop task, Taylor et al. found

significant activation in the left cingulate gyrus (BA32) and a non-significant

activation occurring to the right of the midline [23]. The stimuli consisted of four

visually-presented letters to which subjects had to respond both in a congruent

manner (naming the letters) and in an incongruent manner (responding to each letter

with the name of the other letters). In a H215O PET study, Pardo et al. found that the

right anterior cingulate gyrus was involved in the ‘response selection process’ of the

Stroop attentional conflict paradigm comparing the reading of words printed in a

congruent color versus words printed in an incongruent color [24]. Bush et al.

reported activation of anterior cingulate in the Counting Stroop, a specialized

interference task designed for fMRI [25]. All the aforementioned findings strengthen

the hypothesis that the anterior cingulate cortex plays a central role in interference

and attentional tasks by mediating response selection and/or by allocating attentional

resources when confronted with competing information processing streams.

The second significant activation in the left medial frontal cortex was found in the

left gyrus frontalis medius (BA8). The third region found to be involved was the

gyrus frontalis superior (BA 10) This region was also found by Bench et al. in a

C15O2 PET experiment with a Stroop-like paradigm in which the naming of ink color

in incongruent color nouns was compared with the naming of ink color in neutral

words [26].

37

The second activation cluster corresponds to the right gyrus frontalis dorsalis,

comprising the right supplemental motor area (SMA) (BA 6) and the right frontal eye

field (BA 8). Pardo et al. [1990] found the supplemental motor area (BA 6) on the

left side to be involved in ‘response selection’. In a task where subjects were asked to

read words aloud, SMA activation was also reported [27]. Bush et al. found a

network of attention/motor cortical regions that were activated during the

interference versus the neutral conditions [25]. This network also comprised the right

and left middle frontal gyrus (BA 9), the left precentral gyrus (BA 4) and premotor

cortex (BA6). Bench et al. reported involvement of the primary motor cortex and

supplementary motor area when comparing the reading of colored crosses, neutral

words and Stroop incongruent words [26].

A possible explanation for the activation in the right SMA can be found in the fact

that in the present study some subjects tended to move their left hand in the Stroop

Coloured Word condition and point with their finger to the card in order to follow,

although they were instructed not to do so. They could not move their right hand

because this was less mobile due to fixation of an intravenous line. This observation

is in accordance with reported PET finding of Kawashima et al. since they used

finger movements in a visual go/no-go task and found a network of regions including

prefrontal, premotor, precentral and insular cortex [28].

Methodological considerations

BRAIN SPECT

The vast majority of neuropsychological activation studies are performed either with

fMRI or PET. Here we will only briefly mention those features that are of direct

relevance as to explain the specific pros and cons of SPECT neuroactivation studies.

fMRI offers the possibility of almost unlimited repeatability of studies due to the

absence of any radiation burden, in combination with a spatial resolution unmatched

by any other functional imaging technique. PET, due to the short half-life of the

38

perfusion tracers, allows for the performance of a limited number of repetitions of

both baseline and activation tasks, and its optimal spatial resolution is still superior to

that of SPECT.

Inherent to the use of fMRI and PET is the almost real-time functional imaging of the

brain in combination with a rigid experimental set-up. Indeed, subjects are

performing a test with their heads positioned in the camera and lying in the supine

position. In contrast, brain SPECT allows one to freeze a mental state with a duration

of minimally two to three minutes due to the short accumulation period for HMPAO

and ECD. This opens up the possibility of applying the tracer outside the nuclear

medicine department, or at least away from the arousing conditions linked to the

camera. Hence, the tracer can be injected intravenously while the subject is

comfortably sitting at a table in the research room of the neuropsychologist and is

performing the neuropsychological test. Evidently, this condition approaches the

classical neuropsychological test conditions much more closely than the

aforementioned experimental conditions linked to fMRI or PET.

The demand in neuropsychological research to adhere to the prescribed test

conditions as strictly as possible in order to validate and evaluate a

neuropsychological test performance, also implies that certain tests may have to be

adapted substantially in order to be suitable under fMRI and PET conditions. The

well-defined neuropsychological testing conditions and the need for only a minor

adaptation to the Stroop test (240 s reading period instead of naming 100 items) were

the main reasons for using SPECT in this study as a tool in the design of

neuropsychological activation paradigms that are clinically applicable in the routine

evaluation of both normal volunteers and patients with neuropsychiatric disease. In

conclusion, even though for a large majority of neuroactivation paradigms, fMRI and

PET would be the imaging techniques of choice, SPECT offers substantial

advantages when performing neuropsychological activation studies.

99MTC-ECD 99mTc-ECD was used as a brain perfusion tracer. When compared to regional cerebral

blood flow (rCBF) measurements with PET and 15O-octanol, it was demonstrated

39

that ECD and HMPAO-SPECT do not completely reflect rCBF. ECD-SPECT

results in an underestimation of high rCBF for activations [29]. Taking this into

account, a decreased sensitivity may especially affect occipital regions where blood

flow as measured by ECD is highest. Since in our study only frontal activations

were studied, this effect might be of less importance. Perfusion tracers such as 99mTc

ECD show a swift blood-brain barrier passage and a relatively large extraction in the

brain. Some reports have shown that for activation studies by a short-interval split-

dose technique, some caution should be applied for washout [19, 30]. Corrections for

inhomogeneous neuronal washout are only necessary when scan-to-scan intervals

surpass a few hours [31]. Hence, the distribution of 99mTc-ECD reflects the rCBF

weighted during approximately the first three minutes post-injection.

Both for the rest and activation task, data acquisition can be performed 5-10 min

after injection of the tracer, as was demonstrated by Koyama et al. [29].

SPLIT-DOSE PARADIGM

The consequences of using SPECT tracers with their relatively long half-life (e.g.

6.02 h for 99mTc) is that, within the time-frame of a typical brain SPECT activation

experiment, the second image is substantially contaminated with the remaining

activity distribution of the previous injection. The number of image acquisitions in

an 99mTc brain SPECT activation is practically limited to two. This is in contrast to

fMRI and PET activation experiments that allow for multiple repeats of various

conditions, even presented in a counterbalanced order within the same subject [32].

For the present study, preference was given to a single-day split-dose activation

paradigm. The main motivation to do so is inspired by our search for a subject-

friendly experimental design. Compared to a two-day protocol, our one-day design

is experienced as much more comfortable by out-patients. A second advantage of

the one-day paradigm is the fact that the physiological variability in brain perfusion

on different days or due to spontaneous activations and changes in mental activity, is

minimized. This can be especially important for patients with mood disorder or

schizophrenia. Most of the neuropsychological activation studies with SPECT, using

a 99mTc-labeled tracer, are applying a two-day protocol [33, 34] and hence, are not

40

confronted with the spill-over activity in the second scan. As one of the first

cognitive activation studies with Tc-labeled SPECT ligands, Philpot et al. used a

single-day split-dose protocol in a verbal fluency task in late-life depressed patients

and in normal volunteers, but the analysis was performed with a less sensitive

regions of interest (ROI) approach [35].

STATISTICAL PARAMETRIC MAPPING (SPM)

Very little work has been performed towards rigorously validating the use of the

SPM method with regard to SPECT neuroactivation studies [36]. Evidently, a valid

statistical analysis of SPECT data requires that the proper system parameters are

accounted for, that the experimental design is optimized and that an analysis is made

of the performance of the statistical tests used. A first study of these issues was

recently undertaken by some of the present authors [37]. SPM is now being applied

to SPECT studies which are far from optimal in terms of the numbers of subjects and

the constraints on multiple repetitions of tasks. Most neuropsychological activation

studies with SPECT used ROI analysis, which is prone to operator bias and when

manually applied, is time consuming and subjective. Moreover, ROI analysis has less

power to accurately distinguish regional variations between a set of images, due to

the dilution effect inherent to the ROI size. Some older neuropsychological

activation SPECT studies have used ROI analysis. The use of more rigorous pixel-

based statistical tests lead to the implementation of SPM, which has now become the

gold standard in the analysis of brain activation experiments [38]. Recently, Tien et

al., in a two-day split-dose activation paradigm with the Wisconsin Card Sorting Test

(WCST) found reliable results, using SPM, despite a sample group of only five

patients [39]. The small number of subjects limits the statistical power, increasing

the chance of a type II error. Our study results, despite the small sample size, were

significant and fitted literature findings.

THE STROOP TEST: BASELINE AND ACTIVATION CONDITION AND

EXPERIMENTAL SET-UP

As in all neuroactivation studies, the design of the appropriate rest and activation

41

task is crucial in neuropsychological activation experiments. The lack of an adequate

estimation of within-subject variability of activation patterns in brain SPECT

activation studies also adds to the importance of this issue. In the SCWT, the choice

of an appropriate activation task is simple since the neuropsychological task itself

consists of a baseline condition, the color card (SCWT card II) and the activation

condition, the colorword card (SCWT card III). Ideally, the only task specific to the

activation condition consists of the suppression the reading reflex. Nonspecific

activations in both tasks are card watching, card item scrolling, color recognizing and

color naming. In other neuropsychological tests, the choice of an appropriate rest

task is far more difficult and is a major key-point in finding the specific activation

patterns in the paradigm. This issue is addressed by some researchers [40], while a

simple baseline scan (sitting in a quiet room with eyes open) is preferred as rest

condition by others [35, 41]. Evidently, additional non-specific activation patterns

are included in the latter case.

Up to now neuroimaging studies used Stroop-like paradigms in which items were

presented on a computer screen, one after one, at fixed time intervals [23, 24, 26].

With real-time imaging modalities, additional drawbacks are associated with the use

of the traditional SCWT. The process of overt speech during data acquisition

produces head movements that potentially hamper fMRI and PET image acquisition

and processing. On the other hand, the alternative of arbitrarily color-labeling four

buttons to be pushed by the subject, adds an undesired layer of cognitive complexity

and requires subjects to have training in order to learn the designated button-color

combinations [25]. To our knowledge, the current study is the first in which the

Stroop test is presented in a commonly used neuropsychological research format,

offering subjects the possibility to set their own pace while reading as quickly and as

accurately as possible without interval times between the items. The only

modification to the classical format consists of the 240 s reading period instead of the

naming of 100 items. This implies that most subjects had to read the Stroop cards

three to four times in order to keep up performance during the full 240 s period,

which is the time required for the bulk of the tracer activity to be captured

intracellularly. Since this presents a change in the classical procedure in which each

card has to be read only once, we wondered if practice would improve the

42

performance and hence lessen the Stroop interference effect and consequently the

perfusion linked with it. From that viewpoint we measured the number of items read

after 1, 2, 3 and 4 minutes and found no shortening of time and constancy of pace.

Moreover, MacLeod and Dunbar found that improvement of the Stroop performance

can occur with extensive practice, but that this generally occurs over a period of

hours to days, if at all [42].

Conclusions

Single-day split-dose 99mTc SPECT allows for the identification of activation regions

during performance of the SCWT under standard neuropsychological conditions.

The main findings, activation in the prefrontal cortex and in the anterior cingulate

cortex are in concordance with previous PET and fMRI findings during Stroop-like

interference tasks.

Due to the physiological properties of the flow tracers, SPECT can offer major

advantages in the experimental design of a neuropsychological activation study.

Patients can perform the test in patient-friendly neuropsychological conditions and

only minor adjustments to the neuropsychological tests have to be done. Although

SPECT is more widely available, its major drawbacks are a lower spatial resolution

and a more stringent limit on the number of possible conditions.

As this study clearly demonstrates the feasibility of a prefrontal activation probe with

SPECT, this approach will be further validated with studies in patient populations.

43

References

1. Johnston M, Keister M. Selective attention. Ann Rev Psychol 1986; 37: 43-75.

2. Shallice T. Specific impairments in planning. In: Broadbent D, Weiskrantz L, eds. The

neuropsychology of cognitive function. London: The Royal Society, 1982: 199-209.

3. Van Zomeren A, Brouwer W. Theories and concepts of attention. In: Clinical

neuropsychology of attention. New York - Oxford: Oxford University Press, 1994: 23-

25.

4. Vendrell P, Junqué C, Pujol J, Angeles Jurado M , Molet J, Grafman J. The role of

prefrontal regions in the Stroop Task. Neuropsychologia 1995; 33: 341-352.

5. Stroop J. Studies of interference in serial verbal reactions. J Exp Psychol 1935; 18:

643-662.


Oxford University Press , 1997: 213-218.

7. Batchelor J, Harvey A, Bryant R. Stroop Colour Word Test as a measure of attentional

deficit following mild head injury. The clinical Neuropsychologist 1995; 9 : 180-186.

8. Perret E. The left frontal lobe of man and the suppression of habitual responses in

verbal categorical behavior. Neuropsychologia 1974; 12: 323-330.

9. Meco G, Gasparini M, Doricchi F. Attentional functions in multiple system atrophy

and Parkinson's disease. J Neurol Neurosurg Psychiatry 1996; 60: 393-398.

10. Martin E, Robertson L, Edelstein H, Jagust W, Sorensen D, San Giovanni D, Chirurgi

V. Performance of patients with early HIV-1 infection on the Stroop task. J Clin Exp

Neuropsychol 1992; 14: 857-868.

11. Cohen J, Servan-Schreiber D. Context, cortex, and dopamine: a connectionist

approach to behavior and biology in schizophrenia. Psychol Rev 1992; 99: 45-77.

12. Lemelin S, Baruch P, Vincent A, Laplante L, Everett E, Vincent P. Attention

disturbance in clinical depression. Deficient distractor inhibition or processing

resource deficit? J Nerv Ment Dis 1996; 184: 114-121.

44

13. Martinot J, Allilaire J, Mazoyer B, Hantouch E, Huret J, Legaut-Demare F,

Deslauriers A, Hardy P, Pappata S, Baron J, Syrota A. Obsessive-compulsive disorder:

a clinical, neuropsychological and positron emission tomography study. Acta Psychiatr

Scand 1990; 82: 233-242.

14. Bigler ED. Frontal lobe damage and neuropsychological assessment. Arch

Neuropsychol 1988; 3: 279-297.

15. Boivin J, Giordani B, Berent S, Amato D, Lehtinen S, Koeppe R, Buchtel H, Foster N,

Kuhl D. Verbal fluency and positron emission tomographic mapping of regional

cerebral glucose metabolism. Cortex 1992; 28: 231-239.

16. Oldfield R. The assessment and analysis of handedness: The Edinburgh Inventory.

Neuropsychologia 1971; 9: 97-113.

17. Hammes J. Stroop Kleur-Woordtest: Handleiding (Manual of the Stroop Colour-Word

Test). Amsterdam: Swets & Zeitlinger, 1971.

18. Shedlack K, Hunter R, Wyper D, McLuskie R, Fink G, Goodwin G. The pattern of

cerebral activity underlying verbal fluency shown by split-dose single photon emission

tomography (SPET or SPECT) in normal volunteers. Psychol Med 1993; 21: 687-696.

19. Orlandi C, Crane P, Platts S, Walovitch R. Regional cerebral blood flow and

distribution of 99m Tc ethyl cysteinate dimer in nonhuman primates. Stroke 1990; 21:

1059-1063.

20. Petersen S, Fox P, Posner M, Mintun M, Raichle M. Positron emission tomographic

studies of the cortical anatomy of single-word processing. Nature 1988; 331: 585-589.

21. Koole M, d’Asseler Y, Van Laere K, Van De Walle R, Vandewiele C, Dierckx R,

Lemahieu I. Evaluation of the accuracy of MR-SPECT and SPECT-SPECT

coregistration using 8 different algorithms. Nucl Med Comm 1999; 20:659-669.

22. Friston K, Ashburner J, Frith CD, Poline J, Heather J, Frackowiak RS. Spatial

registration and normalisation of images. Human Brain Map 1995; 2: 165-189.

23. Taylor S, Kornblum S, Minoshima S, Oliver L, Koeppe R. Changes in medial cortical

blood flow with a stimulus-response compatibility task. Neuropsychologia 1994; 31:

249-255.

45

24. Pardo J, Pardo P, Janer K, Raichle M. The anterior cingulate cortex mediates

processing selection in the Stroop attentional conflict paradigm. Proceedings of the

National Academy of Sciences of the United States of America 1990; 87: 256-259.

25. Bush G, Whalen P, Rosen B, Jenike M, McInerney S, Rauch S. The Counting Stroop:

an interference task specialized for functional neuroimaging-validation study with

functional MRI. Human Brain Mapping 1998; 6: 270-282.

26. Bench C, Frith CD, Grasby P, Friston KJ, Paulescu E, Frackowiak RS, Dolan RJ.

Investigations of the functional anatomy of attention using the Stroop test.


27. Holm S, Madsen P, Sperling B, Lassen N. Use of the 99m Tc-bicisate in activation

studies by split-dose technique. J Cereb Blood Flow Metab 1994; 14 Suppl 1: S115-

120.

28. Kawashima R, Satoh K, Itoh H, Ono S, Furumoto S, Gotoh R, Koyoma M, Yoshioka

S, Takahashi T, Takahashi K, Yanagasiwa T, Fukuda H. Functional anatomy of go/no-

go discrimination and response selection - A PET study in man. Brain Res 1996; 728:

79-89.

29. Koyama M, Kawashima R, Ito H, Ono S, Sato K, Goto R, Kinomura S , Yoshioka S,

Saro T, Fukuda H. SPECT Imaging of Normal Subjects with Technetium-99m-

HMPAO and Technetium-99m-ECD. J Nucl Med 1997; 38: 587-592.

30. Mild Traumatic Brain Injury Committee of the Head Injury Interdisciplinary Special

Interest Group of the American Congress of Rehabilitation Medicine. Definition of

mild traumatic brain injury. J Head Trauma Rehab 1993; 8: 86-87.

31. Leveillé J, Demonceau G, Deroo M, Rigo P, Taillefer R, Morgan R, Kupranick D,

Walovitch R. Characterization of Technetium-99m L,L-ECD for brain perfusion

imaging.2. Biodistribution and brain imaging in humans. J Nucl Med 1989; 30: 1902-

1910.

32. Frith CD, Friston KJ, Herold S, Silbersweig D, Fletcher P, Cahill C, Dolan RJ,

Frackowiak RS, Liddle PF. Regional brain activity in chronic schizophrenic patients

during the performance of a verbal fluency task. Br J Psychiatry 1995; 167: 343-349.

33. Parellada E, Catafau AM, Bernardo M, Lomena F, Gonzalez-Monclus E, Setoain J.

46

Prefrontal dysfunction in young acute neuroleptic-naive schizophrenic patients: a

resting and activation SPECT study. Psychiatry Research: Neuroimaging 1994; 55:

131-139.

34. Keilp J, Herrera J, Stritzke P, Cornblatt B. The continuous performance test, identical

pairs version (CTP-IP): III. Brain functioning during performance of numbers and

shapes subtasks. Psychiatry Research: Neuroimaging 1997; 74: 35-45.

35. Philpot MP, Banerjee S, Needham-Bennett H, Costa DC, Ell PJ. 99mTc-HMPAO single

photon emission tomography in late life depression: a pilot study of regional cerebral

blood flow at rest and during a verbal fluency task. Journal of Affective Disorders

1993; 28: 233-240.

36. Rimel R, Giordani B, Barth J, Boll T, Jane J. Disability caused by minor head injury.

Neurosurg 1981; 9: 221-228.

37. Rutherford W, Merrett J, McDonald J. Symptoms at one year following concussion

from minor head injuries. Injury 1979; 10: 225-230.

38. Acton PD, Friston KJ. Statistical parametric mapping in functional neuroimaging:

beyond PET and fMRI activation studies. Eur J Nucl Med 1998; 25: 663-667.

39. Tien AY, Schlaepfer TE, Orr W, Pearlson G. SPECT brain blood flow changes with

continuous ligand infusion during previously learned WCST performance. Psychiatry

Research: Neuroimaging 1998; 82: 47-52.

40. Marenco S, Coppola R, Daniel DG, Zigun JR, Weinberger D. Regional cerebral blood

flow during the Wisconsin Card Sorting Test in normal subjects studied by Xenon-133

Dynamic SPECT: comparison of absolute values, percent distribution values and

covariance analysis. Psychiatry Research: Neuroimaging 1993; 50: 177-192.

41. Warkentin S, Risberg J, Nilsson A, Karlson S, Graae E. Cortical activity during speech

production. A study of regional cerebral blood flow in normal subjects performing a

word fluency task. Neuropsychiatry Neuropyschology and Behaviour Neurology 1991;

4: 305-316.

42. MacLeod C, Dunbar K. Training and Stroop-like interference: evidence for a

continuum of automaticity. J Exp Psychol 1988; 10: 304-315.

47

CHAPTER 2:

VERBAL FLUENCY AS A PREFRONTAL ACTIVATION PROBE: A

VALIDATION STUDY USING 99MTC-ECD BRAIN SPECT.

Kurt Audenaert, Boudewijn Brans, Koen Van Laere, Philippe Lahorte, Jan Versijpt,

Kees van Heeringen, Rudi Dierckx

(European Journal of Nuclear Medicine, 2000; 27(12):1800-1808)

Summary

This study aimed to investigate the feasibility of brain single-photon emission

tomography (SPET) in the letter and Category Fluency paradigm of the Controlled

Oral Word Association (COWA) test in healthy volunteers. Two groups each

comprising ten right-handed healthy volunteers were injected twice with 370 MBq

Tc-99m ECD following a split-dose paradigm (resting and activation condition).

Statistical parametric mapping (SPM96) was used to determine voxelwise significant

changes. The Letter Fluency and the Category Fluency activation paradigm had a

differential brain activation pattern. The posterior part of the left inferior prefrontal

cortex (LIPC) was activated in both paradigms, with the Category Fluency task

having an extra activation in the anterior LIPC. In the Category Fluency task, but not

the Letter Fluency task, an activation in the right inferior prefrontal cortex was

found. These findings confirm to a large extent the results of previous functional

magnetic resonance imaging and positron emission tomography studies in semantic

and phonological activation paradigms. The choice and validity of various

methodological characteristics of the experimental design leading to these results are

critically discussed. It is concluded that brain SPECT activation with the Letter

Fluency and Category Fluency paradigm under standard neuropsychological

conditions in healthy volunteers is both technically and practically feasible.

48

Introduction

Tests of Verbal Fluency, such as the Controlled Oral Word Association (COWA)

test, are designed to evaluate the spontaneous production of words beginning with a

given letter (Letter Fluency; syntactic or phonetic association) or words belonging to

a given class (Category Fluency; semantic association), within a limited amount of

time [1]. These tests are frequently used in the assessment of patients with suspected

prefrontal disorders.

Neuropsychological investigation of patients with brain damage [2,3] or degenerative

diseases [4] has identified the prefrontal cortex as being critically involved in

language function in general, and word fluency in particular. Some authors report

high sensitivity of the COWA test to frontal lobe damage regardless of the side of the

lesion [2,5], while others have found more impairment in patients with left and

bifrontal lesions [3,6,7]. Neuropsychological investigations have demonstrated that

brain damage or disease can differentially affect phonological and semantic

processing [8, 9], suggesting that distinct brain structures are involved in these

language processing components [10]. It is of interest from a clinical

neuropsychological viewpoint to know which prefrontal regions are involved in the

semantic and the phonological association subtest of the COWA test. This was the

first aim of this study.

Regional brain activity during phonological and semantic processing has been

studied using functional magnetic resonance imaging (fMRI) and positron emission

tomography (PET). These studies have included "activation" procedures by means of

a cognitive task during the scanning procedure [11]. Thus, the conditions under

which the test is performed using PET or fMRI do not fit classical

neuropsychological test conditions, i.e. sitting at a table or before a computer screen

in a quiet room. Due to this incompatibility, classical and well-validated test

presentations commonly need to be adapted in order to fit PET or fMRI test

conditions, in that the subject has to carry out the test in the supine position with his

or her head positioned in a scanner.

The characteristics of perfusion tracers used in single-photon emission

tomography (SPET) allow the use of classical neuropsychological test conditions.

49

Distribution of technetium-99m ethyl cysteinate dimer (ECD) intracellular uptake

takes place almost in proportion to the microperfusion over approximately a 2-min

time window after intravenous injection. Subsequent scanning of the patient can

thus be performed from 10 min to 6 h in a flexible manner and will reflect the

physiological and cognitive state at the time of injection.

Brain perfusion SPET studies with the Verbal Fluency paradigm have already been

carried out [12,13]. However, differentiation of phonological versus semantic

processing in the COWA test under neuropsychological test conditions has never

been studied.

Although advances have been made in the instrumentation, leading to a spatial

resolution of 7 mm, SPET resolution is still limited compared with current PET and

fMRI. Therefore, our second aim was to compare the SPET functional imaging

results of the COWA activation procedure with the results from PET or fMRI

functional imaging and COWA or other Verbal Fluency paradigms.

In order to compare statistical parametric mapping (SPM) with the conventional

region of interest approach, an automated volume of interest (VOI)-based analysis

was also conducted and compared with the SPM analysis.

Methods

Subjects

Twenty healthy volunteers (8 males, 12 females) with a mean age of 22.7 years (SD

2.66; range 19-28 years) and a mean duration of education of 14.9 years (SD 1.48;

range 12-16 years) were included in the study. All subjects were right-handed as

assessed by the Edinburgh Handedness Inventory [14] and Dutch-speaking. Since

two paradigms were tested, subjects were divided at random into two groups of ten

subjects each. None of the participants had a history of major medical or

neurological or psychiatric disease. Subjects were drug-free with the exception of

oral contraceptives. Research was compliant with the Code of Ethics of the World

Medical Association (Declaration of Helsinki). Each subject gave informed consent,

50

following the guidelines of the local ethics committee. No subject had any previous

experience with the COWA test.

COWA test

The subjects underwent two regional cerebral blood flow measurements, one during

a resting task and one during an activation task, i.e. one of the subtests of the Verbal

Fluency test (Letter Fluency or Category Fluency). For the activation task, the

subjects were instructed to say aloud as many words as they could think of beginning

with a specified target letter ("N", "A", K" and "B") or a specified target category

("animals", "jobs", "fruit and vegetables" and "interior and furniture"). Target letters

or categories were changed at 1-min intervals. The test score was defined by the

number of words produced in the 240-s period, excluding perseverations. For the

resting task, subjects were instructed to say aloud the letters of the alphabet and

numbers from 1 to 100 alternately, during a 240-s period. The test score was defined

by the number of words produced in the 240-s period. In this resting task, input

(auditory target) and output (saying words) were essentially the same as in the

activation task. However, as this task is executed "automatically" and with minor

effort, semantic and phonological processing is minimized. The measurements were

performed in the same order (resting task, activation task) in all subjects. The tasks

commenced 30 s prior to tracer injection and continued throughout the 240-s period.

Subjects who were included in the Letter Fluency paradigm completed the COWA

test after the imaging procedure by performing the Category Fluency test, and vice

versa.

Paradigm design.


tracer can be given in a split-dose protocol, enabling at least one repeat scan within a

short period [13]. The study subjects were injected intravenously twice with

370 MBq (10 mCi) 99mTc-ECD following a split-dose paradigm. In accordance with

standard neuropsychological testing conditions, subjects sat at a table in a quiet room

and were informed about the testing procedure. Subsequently, an intravenous line

51

was installed and after approximately 10 min, subjects were instructed to start the

resting task. After a 30-s period of citing words by the subject, 370 MBq 99mTc-ECD

was injected via the intravenous line and subjects kept on generating words for

another 210 s period. The injection was performed manually in bolus over a period of

less than 3 s using a three-way valve with subsequent flushing with 10 ml saline. Ten

minutes following the reading session, the first image acquisition was performed

during a 20-min period. Subsequently, subjects were instructed to start the activation

task and the aforementioned procedure was repeated. The number of words

generated by the subject was assessed by the experimenter.

Figure 1

Split-dose paradigm with injected dose and testing sequence

Image acquisition, reconstruction and SPM.

Image acquisition was performed with a three-headed gamma camera (Toshiba CGA

9300 A) equipped with parallel-hole low-energy high-resolution collimators. The

full-width at half-maximum (FWHM) of this system as measured in-house was

10.0 mm for 99mTc. Data were acquired over 90 projections, 40 s per projection, in a

single 20% 99mTc window.

Projections were reconstructed using filtered back-projection and uniform

attenuation correction was applied with an effective attenuation coefficient of

52

0.12/cm. Filtered back-projection was used with 3D Butterworth post-filtering (order

8, cut-off 0.12 cycles/cm) The reconstructed data were transferred to Interfile 3.3

format onto a HERMES processing system (Nuclear Diagnostics, Hagerstad,

Sweden) and fitted automatically onto an in-house constructed database template

positioned in Talairach co-ordinates [15].

Statistical parametric mapping with SPM96 (Wellcome Department of Cognitive

Neurology, Institute of Neurology, London, UK) was applied to determine

significant regions of increased or decreased activity. Prior to statistical analysis, data

were smoothed with a Gaussian filter of 14 mm FWHM. Proportional scaling was

used with a grey matter threshold of 0.4. Statistical comparisons between conditions

were performed on a pixel-by-pixel basis using t statistics, generating SPM(t) maps

subsequently transformed to the unit normal distribution SPM(Z) maps.

We investigated areas at P=0.05 setting the height threshold to P=0.01, without

correction for multiple comparisons since there was an anatomical pre-existing

hypothesis to find activation in the prefrontal lobe.

For the VOI analysis, the reconstructed data were fitted automatically onto an in-

house constructed database template positioned in Talairach co-ordinates. This

template was based on 20 normal, healthy volunteers and was measured with the

same equipment, total dose and under the same standard circumstances and

reconstructed as given above. The fitting procedure was performed with nine

parameters (scale, shift and rotation) using a principal axis transform and a count-

difference minimization algorithm (fit threshold of 0.50; BRASS, Nuclear

Diagnostics, Sweden). A set of 37 VOIs comprising predefined Brodmann areas was

delineated on the standard perfusion template mentioned above by direct reference to

the Talairach atlas. For each individual subject and scan, the VOI activity counts

were calculated per voxel and normalized onto the total number of counts in the

complete VOI set of the scan.

Statistical analysis.

Age and years of education were compared in a between-group design using the

Mann-Whitney U test. Results of the Category Fluency task and the Letter Fluency

53

task were compared in a within-group design using the Mann-Whitney U test.

Correlations were calculated using Spearman's correlation test.

Results

Demographic variables

The mean age and years of education of the volunteers were 22.7 (SD 2.66) and

14.9 years (SD 1.48), respectively. Since two paradigms were tested, subjects

were divided into two groups of ten subjects. There was no difference in gender

since each group contained four men and six women. There were no significant

differences in mean age and mean number of years of education between the two

groups.

Group 1 Group 2

Mean SD Mean SD U P

Demographic variables

Age 21.8 2.0 23.6 2.9 31.5 0.1

Years of Education 14.7 0.8 15.1 1.97 42.5 0.5

Neuropsychological Test results

Category Fluency task 55.7 9.5 54.7 12.2 46.0 0.7

Letter Fluency task 49.1 10.6 47.1 11.2 47.0 0.8

Table 1

Demographic variables and neuropsychological test results of the subjects included in the

Letter Fluency paradigm (group 1) and the Category Fluency paradigm (group 2). Mann

Whitney U and P values are given

54

Test results

There were no statistically significant differences in number of cited items in the

tasks between the study groups (Table 1). When comparing the scores on the Letter

Fluency task (range 39-92; mean 64.1; SD 14.2) with those on the Category Fluency

task (range 47-91; mean 73.7; SD 14.2) in an intra-subject design, significantly more

items were cited in a 240-s period in the Category Fluency task (Z=-3.33; P=0.001).

The scores were significantly correlated (r=0.86; P<0.001).

Brain SPET flow results

The Letter Fluency activation paradigm (phonological association) showed three

clusters of activation, indicating an increase in regional blood flow. The first cluster

was found in the left inferior frontal gyrus (BA 45 and BA 47).

The second cluster was situated in the left medial wall of the prefrontal cortex and

comprised the anterior part of the cingulate gyrus (BA 32 and BA 24). The third

cluster was found in the right medial and inferior temporal gyrus (BA 37 and BA 21)

(Fig. 2). No significant deactivation clusters were found at the specified thresholds.

Cluster

P

Talairach

coordinates Region

0.296 (-39, 27, -6) L G Front Inf (BA 47)

(-51, 9, 6) L G Front Inf (BA 45)

0.488 (-6, 24, 27) L G Cinguli (BA 24)

(-6, 9, 21) L G Cinguli (BA 24)

(18, 15, 39) R G Cinguli (BA 32)

0.394 (54, -60, -9) R G Temp Inf (BA 37)

(60, -51, 3) R G Temp Med (BA 21)

55

Figure 2

Comparison of means (ten subjects) o

Figure 2

Comparison of means (ten subjects) on the Letter Fluency test versus citing the alphabet and

counting as a resting task. Statistical parametric maps in three whole-brain projections.

Cluster level P values and activated regions are given

Regarding the Category Fluency activation paradigm (semantic association), a first

activation cluster was found in the left inferior frontal gyrus (BA 44/45) and left

medial frontal gyrus (BA 10/46). A second cluster of activation comprised two

distinct regions of which the first was situated in the anterior cingulate gyrus, both

left and right sided (BA 24 and BA 32), and the second in the right medial frontal

gyrus (BA 10) (Fig. 3). No significant deactivation clusters were found at the

specified thresholds.

56

Cluster

P

Talairach

coordinates Region

0.032 (-42, 45, 3) L G Front Med (BA 10/46)

(-48, 15, 3) L G Front Inf (BA 45)

(-54,15, 24) L G Front Inf (BA 44)

0.031 (21, 18, 18) R G Cinguli (BA 24)

(-6, 24, 30) L G Cinguli (BA 32)

(36, 54, 0) G Front Med (BA 10)

Figure 3

Comparison of means (ten subjects) on the Category Fluency test versus citing the

alphabet and counting as a resting task. Statistical parametric maps in three whole-brain

projections. Cluster level P values and activated regions are given

A differential activation pattern between the Letter Fluency task and the Category

Fluency task was found. Phonological processing but not semantic processing

activated the right medial and inferior temporal gyrus. Semantic but not phonological

processing activated the right medial frontal gyrus. Phonological and semantic

processing were both associated with an activation of the left inferior prefrontal

cortex (LIPC) and the anterior cingulate cortex (Fig. 4).

57

Figure 4

Schema of areas activated in the semantic and phonological tasks

The frontal lobes and those regions showing significant (de)activations on the SPM

analysis were analysed by the VOI method. The areas encompassed in the VOI and

quantification results are shown in Table 2.

Common areas, activated in phonetic and

semantic task

Areas activated in phonetic, but not in

semantic task

Areas activated in semantic, but not in

phonetic task

58

Phonological activation

paradigm

Semantic activation

paradigm

Regio VOI

size

Basal

condition

Activation

condition P

Basal

condition

Activation

condition P

Gyrus cinguli ant 374 109.6 108.3 0.18 107.9 106.7 0.12

Gyrus cinguli post 279 114.1 113.0 0.26 113.6 111.6 0.12

L lat frontal 619 96.6 97.2 0.25 93.9 94.9 0.08

L orbito-frontal 168 94.9 98.0 0.26 99.2 99.4 0.91

L med frontal 697 91.2 92.1 0.21 89.5 89.9 0.47

L sup frontal 473 93.9 93.8 0.96 91.4 92.4 0.21

L ant temporal 166 87.5 87.6 0.90 86.5 88.7 0.06

L middle temporal 610 92.1 93.2 0.24 92.2 92.0 0.71

L sup temporal 261 104.4 103.7 0.28 104.0 103.1 0.40

R lat frontal 609 100.9 100.6 0.56 99.9 99.9 0.90

R orbito-frontal 132 93.7 97.2 0.17 100.9 94.3 0.10

R med frontal 641 93.2 95.0 0.02 94.3 94.2 0.84

R sup frontal 384 95.4 94.9 0.65 92.4 93.3 0.28

R ant temporal 163 86.7 85.7 0.50 88.7 88.8 0.32

R middle temporal 600 97.3 96.7 0.12 98.2 97.5 0.11

R sup temporal 281 104.3 103.2 0.14 104.3 103.7 0.24

ant.= anterior; post.= posterior; lat.= lateral; sup.= superior; med= medial

Table 2

Conventional VOI analysis in the phonological and semantic activation paradigms

Discussion

Regional blood flow distribution during Verbal Fluency

Task performance depends on multiple cognitive processes, including sensory

processing of the cue, retrieval of words from memory, the selection of a word

appropriate for the cue, and covert or overt articulation of the word [16].

59

Despite variability in the design of previous studies, our results tended to be

consistent with those from other studies on semantic or phonological processing in

general and COWA neuropsychological activation in particular. Thus the

involvement of the left-sided prefrontal cortex, and more specifically of the LIPC, in

our Category Fluency paradigm was in keeping with previous findings using

different semantic processing paradigms. These included semantic decision tasks

[17], a task based on generating words in semantic relationships [18, 19], selection of

information among competing alternatives from semantic memory [20] and

generating verbs to a presented concrete noun [21]. Neuropsychological studies of

patients with lesions of the LIPC have shown a reduced ability to take semantic

decisions when compared to patients with lesions in the left superior prefrontal

cortex or the right prefrontal area [22]. The function of the LIPC can thus be

described as a semantic executive system that mediates on-line retrieval of long-term

conceptual knowledge necessary for guiding task performance [23].

Our finding that the Letter Fluency task also challenged the LIPC was expected,

based on previous functional imaging studies using phonological processing

paradigms and on neuropsychological studies in patients with brain lesions. In a

Letter Fluency paradigm using the xenon inhalation technique with multiple

scintillation detectors, several authors found activation in the left prefrontal cortex

[24,25]. More specifically, activation of the left-sided inferior prefrontal cortex was

found in an fMRI Letter Fluency design [16], in a phonological syllable judgment

paradigm with fMRI [17], in a rhyming judgment task with PET [26] and in phonetic

monitoring, again with PET [27,28].

Neuropsychological examination of patients with brain lesions showed a reduced

performance on fluency tasks in association with left prefrontal cortex injury

[2,18,29,30]. When comparing the LIPC activation in the Category Fluency versus

the Letter Fluency task in our experiment, both tasks induced an activation of the

posterior part of the LIPC (BA 45 and 47) but only the semantic task was associated

with activation of the anterior part of the LIPC (BA 46/10). This finding is in

keeping with the observations of Poldrack et al. [17] and Warburton et al. [31].

This part of the LIPC may thus serve as a semantic memory system or semantic

executive system [30, 32, 33]. The role of such a system would be to access,

60

maintain and manipulate semantic representations which are localized elsewhere in

the cortex [17].

Activation of the anterior part of the right-sided inferior frontal cortex was found in

our semantic activation paradigm and not in the phonological activation paradigm.

Thus, the Category Fluency task, and not the Letter Fluency task, resulted in a

bilateral prefrontal activation. This was previously demonstrated in a SPET

activation study by Cardebat et al. [34], who found increased rCBF in the right and

medial frontal region for the semantic (category) fluency task, but no increase

compared with baseline measures for formal (letter) fluency [34]. Other groups,

using PET or fMRI, have also found a bilateral prefrontal activation in a semantic

task [17,31].

We found a significant activation in the anterior cingulate in both the letter and the

Category Fluency task. However, the activation of the cingulate in the two conditions

was unequal (Letter Fluency task: P=0.49 vs P=0.03 in the Category Fluency task).

Activation of the cingulate cortex was also found in a semantic noun generation to

supraordinate noun stimulus (BA 24) [35] and in a semantic and in a phonological

processing task (BA 24/32) [17]. Functions attributed to these regions include the

formulation and initiation of a strategy to identify the appropriate words from

memory and attention to the task following motivation processes [18, 31]. Activation

of the anterior cingulate is not specific since it occurs in other prefrontal activation

tasks, such as the Stroop test [36,37] and the Wisconsin Card Sorting Test [38].

The activation of the inferior and middle right temporal cortex was found in the

phonological task but not in the semantic task. Parks et al. [39] found a bilateral

activation of temporal and frontal lobes.


Brain SPET. The vast majority of cognitive neuropsychological activation studies are

performed with either fMRI or PET. Here we will only briefly mention those features

that are of direct relevance to explanation of the specific pros and cons of SPET

neuroactivation studies.

61

fMRI offers the possibility of almost unlimited repeatability of studies owing to the

absence of any radiation burden, in combination with a spatial resolution unmatched

by any other functional imaging technique. PET, owing to the short half-life of the

perfusion tracers, allows for the performance of a limited number of repetitions of

both baseline and activation tasks, and its optimal spatial resolution is still superior to

that of SPET.

Inherent to the use of fMRI and PET is the almost real-time functional imaging of the

brain in combination with a rigid experimental set-up. Indeed, subjects are


position. In contrast, brain SPET allows one to freeze a mental state with a duration

of 2-3 min owing to the short accumulation period for HMPAO and ECD. This opens

up the possibility of applying the tracer outside the nuclear medicine department, or

at least away from the arousing conditions linked to the camera. Hence, the tracer

can be injected intravenously while the subject is comfortably sitting at a table in the

research room of the neuropsychologist and is performing the neuropsychological

test. Evidently, this condition approaches the classical neuropsychological test

conditions much more closely than the aforementioned experimental conditions

linked to fMRI or PET.

99mTc-ECD. 99mTc-ECD was used as a brain perfusion tracer. Comparison with regional cerebral

blood flow (rCBF) measurements with PET and oxygen-15 octanol has demonstrated

that ECD SPET and HMPAO SPET do not completely reflect rCBF. In this respect,

it should be emphasized that ECD SPET results in an underestimation of high rCBF

for activations [40]. Non-linear uptake of ECD is seen in high-flow areas such as the

occipital (visual) cortex, thereby limiting detection sensitivity in these areas.

Taking this into account, a decreased sensitivity may especially affect occipital

regions, where blood flow as measured by ECD is highest. Since in our study only

frontal and temporal activations were studied, this effect might be of limited

relevance to our findings. Perfusion tracers such as 99mTc-ECD show a swift blood-

brain barrier passage and a relatively large extraction fraction in the brain. Some

62

reports have shown that in activation studies using a short-interval split-dose

technique, some caution should be applied for washout [41]. Corrections for

inhomogeneous neuronal washout are only necessary when scan-to-scan intervals

surpass a few hours [42]. Hence, the distribution of 99mTc-ECD reflects the rCBF

weighted during approximately the first 3 min post-injection. For both the rest and

the activation task, data acquisition can be performed at the earliest 10 min after

injection of the tracer, as was demonstrated by Koyama et al. [40] and is mentioned

by the supplier.

Split-dose paradigm

The consequences of using SPET tracers with their relatively longer half-life is that,

within the time-frame of a typical brain SPET activation experiment, the second

image is substantially contaminated with the remaining activity distribution of the

previous injection. The number of image acquisitions in a 99mTc brain SPET

activation is practically limited to two. This is in contrast to fMRI and PET activation

experiments, which allow for multiple repeats of various conditions, even presented

in a counterbalanced order within the same subject [43].


paradigm, the main motivation being our search for a subject-friendly experimental

design. Compared with a 2-day protocol, our 1-day design is experienced as much

more comfortable by out-patients. A second advantage of the 1-day paradigm is the

fact that the physiological variability in brain perfusion on different days or due to

spontaneous activations and changes in mental activity is minimized. This can be

especially important for patients with mood disorder or schizophrenia.

SPM

Very little work has been devoted to rigorously validating the use of the SPM

method with regard to SPET neuroactivation studies. Evidently, a valid statistical

analysis of SPET data requires that the proper system parameters are accounted for,

that the experimental design is optimized and that an analysis is made of the

63

performance of the statistical tests used. A study of these issues was undertaken by

some of the present authors [44]. Lahorte et al. have shown in a simulation study that

for the current study size and image processing characteristics (filtering, filtered

back-projection), spherical activation foci of about 2 and 7 ml can be detected using

an SPM approach if their "activation level" (= the percentage increase in rCBF)

exceeds 10% and 5%, respectively.

SPM is now being applied to SPET studies which are far from optimal in terms of

the numbers of subjects and the constraints on multiple repetitions of tasks. Most

older neuropsychological activation studies with SPET used ROI analysis, which is

time-consuming, prone to operator bias and, when manually applied, subjective.

Moreover, ROI analysis has less power to accurately distinguish regional variations

between a set of images owing to the dilution effect inherent to the ROI size. The use

of more rigorous pixel-based statistical tests has resulted in the implementation of

SPM, which has now become the gold standard in the analysis of brain activation

experiments [45]. Recently, Tien et al., in a 2-day split-dose activation paradigm

with the WCST, using SPM, found reliable results despite a sample group of five

patients [38]. The small number of subjects limited the statistical power, increasing

the chance of a type II error. Our study results, despite the small sample size, were

significant and fitted literature findings.

The VOI procedure failed to show the significant activation areas, except for the

right medial frontal cortex in the phonological activation paradigm. When comparing

automated VOI and SPM analyses, it is known that differences can partly be

attributed to a dilution over larger volumes of focal activation effects.

Inhomogeneous focal deactivation and activation over the VOI area or adjacent to

the artificial separation of two VOIs can lead to decreased sensitivity of the VOI

analysis [46].

The size of the VOIs on the one hand determines the resolving capability of localized

focal changes between images, but on the other hand an increase in false-negatives

can be expected due to the low degrees of freedom and multiple comparison

corrections.

64

Conclusions

Single-day split-dose 99mTc SPET allows for the identification of activation regions

under standard neuropsychological conditions. The main findings, a differential

activation in the prefrontal cortex and an activation in the anterior cingulate cortex,

are in concordance with previous PET and fMRI findings during tasks that require

semantic and phonological processing.

Owing to the physiological properties of the flow tracers, SPET can offer advantages

in the experimental design of a neuropsychological activation study. Patients can

perform the test in patient-friendly neuropsychological conditions and only minor

adjustments to the neuropsychological tests have to be done. Although SPET is more

widely available, its major drawbacks are a lower spatial resolution and a more

stringent limit on the number of possible conditions.

This study clearly demonstrates the feasibility of a prefrontal activation probe with

SPET; studies in different patient populations will need to be undertaken to further

validate this approach.

65

References

1. Spreen O, Strauss EA. Compendium of neuropsychological tests. New York Oxford:

Oxford University Press, 1997.

2. Miceli G, Caltagirone C, Gainotti G, Masullo C, Silveri M. Neuropsychological

correlates of localized cerebral lesions in non-aphasic brain-damaged patients. J Clin

Neuropsychol 1981; 3: 53-63.

3. Perret E. The left frontal lobe of man and the suppression of habitual responses in verbal

categorical behaviour. Neuropsychologia 1974; 12: 323-330.

4. Pachana N, Boone K, Miller B, Cummings J, Berman N. Comparison of

neuropsychological functioning in Alzheimer's disease and frontotemporal dementia. J

Int Neuropsychol Soc 1996; 2: 505-510.

5. Bruyer R, Tuyumbu B. Fluence verbale et lésions du cortex cérébrale: performances et

types d'erreurs. Encephale 1980; 6: 287-297.

6. Ramier A, Hecaen H. Rôle respectif des atteintes frontales et la latéralisation lésionelle

dans les déficits de la "fluence verbale". Rev Neurol 1970; 123: 17-22.

7. Ruff R, Allen C, Farro C, Niemann H, Wylie T. Differential impairment in patients with

left versus right frontal lobe lesions. Arch Clin Neuropsychol 1994; 9: 41-55.

8. Lacy M, Gore P, Pliskin N, Henry G. Verbal Fluency task equivalence. Clin

Neuropsychol 1996; 10: 305-308.

9. Barr A, Brandt J. Word-list generation deficits in dementia. J Clin Exp Neuropsychol

1996; 18: 810-822.

10. Bayles K, Salmon D, Tomoeda C, Jacobs D. Semantic and letter category naming in

Alzheimer's patients: a predictable difference. Dev Neuropsychol 1989; 5: 335-347.

11. Boivin J, Giordani B, Berent S, Amato D, Lehtinen S, Koeppe R, Buchtel H, Foster N,

Kuhl D. Verbal fluency and positron emission tomographic mapping of regional cerebral

glucose metabolism. Cortex 1992; 28: 231-239.

12. Philpot MP, Banerjee S, Needham-Bennett H, Costa DC, Ell PJ. 99mTc-HMPAO single

photon emission tomography in late life depression: a pilot study of regional cerebral

blood flow at rest and during a verbal fluency task. J Affect Disord 1993; 28: 233-240.

13. Shedlack K, Hunter R, Wyper D, McLuskie R, Fink G, Goodwin G. The pattern of

66



14. Oldfield R. The assessment and analysis of handedness: the Edinburgh Inventory.


15. Talairach J, Tournoux P. Co-planar stereotactic atlas of the human brain. Stuttgart:

Thieme Medical, 1988.

16. Curtis V, Bullmore E, Brammer M, Wright I, Williams S, Morris R, Sharma T, Murray

R, McGuire P. Attenuated frontal activation during a verbal fluency task in patients with

schizophrenia. Am J Psychiatry 1998; 155: 1056-1063.

17. Poldrack R, Wagner A, Prull M, Desmond J, Glover G, Gabrieli J. Functional

specialization for semantic and phonological processing in the left inferior prefrontal

cortex. Neuroimage 1998; 10: 15-35.

18. Petersen S, Fox P, Posner M, Mintun M, Raichle M. Positron emission tomographic

studies of the cortical anatomy of single-word processing. Nature 1988; 331: 585-589.

19. Klein D, Milner B, Zatorre R, Meyer E, Evans A. The neural substrates underlying word

generation: a bilingual functional-imaging study. Proc Natl Acad Sci U S A 1995; 92:

2899-2903.

20. Thompson-Schill S, D'Esposito M, Aguirre G, Farah M. Role of the left inferior

prefrontal cortex in retrieval of semantic knowledge: a reevaluation. Proc Natl Acad Sci

U S A 1997; 94: 14792-14797.

21. Wise R, Chollet F, Hadar U, Friston K, Hoffner E, Frackowiak R. Distribution of

cortical neural networks involved in word comprehension and word retrieval. Brain

1991; 114: 1803-1817.

22. Swick D, Knight R. Is prefrontal cortex involved in cued recall? A neuropsychological

test of PET findings. Neuropsychologia 1996; 34: 1019-1028.

23. Wagner A, Buckner R, Koutstall W, Schacter D, Gabrieli J, Rosen B. An fMRI study of

within- and across-task item repetition during semantic repetition. Cog Neurosci Soc

Proc 1997; 35: 35.



word fluency task. Neuropsychiatry Neuropyschol Behav Neurol 1991; 4: 305-316

25. Cantor-Graae E, Warkentin S, Franzén G, Risberg J. Frontal lobe challenge: a

67

comparison of activation procedures during rCBF measurements in normal subjects.

Neuropsychiatry Neuropyschol Behav Neurol 1993; 6: 83-92.

26. Sergent J, Ohta S, Macdonald S. Functional neuroanatomy of face and object processing:

a positron emission tomographic study. Brain 1992; 115: 15-36.

27. Zatorre R, Evans A, Meyer E, Gjedde A. PET studies of phonetic processing of speech:

review, replication, and reanalysis. Science 1992; 256: 846-849.

28. Stromswold K, Caplan D, Alpert N, Rauch S. Localization of syntactic comprehension

by positron emission tomography. Brain Lang 1996; 52: 452-473.

29. Benton A. Differential behavioral effects in frontal lobe disease. Neuropsychologia

1968; 6: 53-60.

30. Fiez J. Phonology, semantics, and the role of the left inferior prefrontal cortex. Human

Brain Map 1997; 5: 79-83.

31. Warburton E, Wise R, Price C, Weiller C, Hadar U, Ramsay S, Frackowiak R. Noun and

verb retrieval by normal subjects Studies with PET. Brain 1996; 119: 159-179.

32. Gabrieli J, Desmond J, Demb J, Wagner A. The role of left prefrontal cortex in language

and memory. Proc Natl Acad Sci U S A 1998; 95: 906-913

33. Wagner A, Desmond J, Demb J, Glover G, Gabrieli J. Semantic repetition priming for

verbal and pictorial knowledge: a functional MRI study of the left inferior prefrontal

cortex. J Cog Neurosci 1997; 9: 714-726.

34. Cardebat D, Demonet J, Viallard G, Faure S, Puel M, Celsis P. Brain functional profiles

in formal and semantic fluency tasks: a SPECT study in normals. Brain Lang 1996; 52:

305-313.

35. Price C, Wise R, Warburton E, Moore C, Howard D, Patterson K, Frackowiak R, Friston

K. Hearing and saying. The functional neuro-anatomy of auditory word processing.

Brain 1996; 119: 919-931.

36. Hock C, Villringer K, Müller-Spahn F, Wenzel R, Heekeren H, Schuh-Hofer S,

Hofmann M, Minoshima S, Schwaiger M, Dirnagel U, Villringer A. Decrease in

partietal cerebral hemoglobin oxygenation during performance of a verbal fluency task

in patients with Alzheimer's disease monitored by means of near-infrared spectroscopy

(NIRS) - correlation with simultaneous rCBF-PET measurements. Brain Res 1997; 755:

293-303.

37. Pardo J, Pardo P, Janer K, Raichle M. The anterior cingulate cortex mediates processing

68

selection in the Stroop attentional conflict paradigm. Proc Natl Acad Sci U S A 1990;

87: 256-259.



Res: Neuroimaging 1998; 82: 47-52.

39. Parks RW, Loewenstein DA, Dodrill KL, Barker WW, Yoshii F, Chang JY, Emran A,

Apicella A, Sheramata WA, Duara R. Cerebral metabolic effects of a verbal fluency test:

a PET scan study. J Clin Exp Neuropsychol 1988; 10: 565-575.

40. Koyama M, Kawashima R, Ito H, Ono S, Sato K, Goto R, Kinomura S, Yoshioka S,

Saro T, Fukuda H. SPECT imaging of normal subjects with technetium-99m-HMPAO

and technetium-99m-ECD. J Nucl Med 1997; 38: 587-592.

41. Holm S, Madsen P, Sperling B, Lassen N. Use of the 99mTc-bicisate in activation studies

by split-dose technique. J Cereb Blood Flow Metab 1994; 14 Suppl 1: S115-S120.

42. Leveillé J, Demonceau G, Deroo M, Rigo P, Taillefer R, Morgan R, Kupranick D,

Walovitch R. Characterization of technetium-99m-ECD for brain perfusion imaging.2.

Biodistribution and brain imaging in humans. J Nucl Med 1989; 30: 1902-1910.




44. Lahorte P, Vandenberghe S, d'Asseler Y, Koole M, Audenaert K, Dierckx R. Statistical

parametric analysis of activation foci detection in brain SPECT imaging. Proceedings of

the 20th Annual International Conference of the IEEE Engineering in Medicine and

Biology Society 1998; 20: 771-774.

45. Acton PD, Friston KJ. Statistical parametric mapping in functional neuroimaging:

beyond PET and fMRI activation studies. Eur J Nucl Med 1998; 25: 663-667.

46. Van Laere K, Vonck K, Boon P, Brans B, Van De Kerckhove T, Dierckx R. Vagus

nerve stimulation in refractory epilepsy: a SPECT activation study. J Nucl Med 2000;

41: 1145-1154.

69

CHAPTER 3:

SPECT NEUROPSYCHOLOGICAL ACTIVATION PROCEDURE WITH

THE VERBAL FLUENCY TEST IN DEPRESSED SUICIDE ATTEMPTERS.

Kurt Audenaert, Koen Van Laere, Ingeborg Goethals, Philippe Lahorte, Boudewijn

Brans, Luc Beelaert, Kees van Heeringen, Rudi Dierckx

(European Journal of Nuclear Medicine, submitted)

Summary

Performance on the Verbal Fluency Test, as a measure of the ability of initiating

processes, is reduced in depressed suicide attempters. The hampered results in this

prefrontal executive task parallel the reduction in prefrontal blood perfusion and

metabolism in depressed subjects. A neuropsychological activation study with the

Verbal Fluency paradigm could evaluate a possible blunted increase in perfusion in

the prefrontal cortex in depressed suicide attempters.

Twenty depressed patients that recently attempted suicide and twenty healthy

volunteers were included in a SPECT split-dose activation study following a Verbal

Fluency paradigm. Statistical Parametric Mapping was used to determine voxelwise

significant changes.

Differences in regional cortical activation between the Letter Fluency and the

category fluency task in depressed suicidal patients were found. Depressed suicide

attempters showed a blunted increase in perfusion changes in the prefrontal cortex.

Methodological restrictions concerning group uniformity, medication bias and

subjective effort of the participants are discussed.

Our findings point at a blunted increase in prefrontal blood perfusion changes as a

possible biological underpinning of reduced drive and loss of initiative in depressed

suicide attempters.

70

Introduction

Loss of interest or apathy and poverty of speech are core aspects of depressive

symptomatology that have a major impact on the daily functioning and quality of life

of the patients. Considering these symptoms from a cognitive viewpoint, they have

an impairment of initiation processes as a common and underlying deficit and they

form part of a so-called dysexecutive syndrome [1,2]. By definition, the initiation

process can be described as the planning activity that leads to the accomplishment of

an action oriented towards an aim that requires control processes [3]. The Verbal

Fluency Test (VFT; Controlled Oral Word Association Test; COWAT) is one of the

tests used most frequently to assess initiation processes [2] and comprises two

subtests: the Letter Fluency test (phonological association) and the Category Fluency

test (semantic association) [4].

Indeed, neuropsychological assessment has revealed depression-related cognitive

deficits in executive functions in general [5,6], and impairment in initiation

processes, as measured by the Verbal Fluency Test, in particular [7,8]. These

impairments have been explained by a subnormal capacity to initiate willed action

[9], more particular, a suboptimal initiation of search strategies [2]. Very recently,

Keilp et al. described neuropsychological dysfunctions in depressed suicide

attempters and demonstrated that deficits in executive functions, comprising Verbal

Fluency, were a risk factor for severe suicide attempts [51]. Previously, reduced

Verbal Fluency was also demonstrated in suicide attempters [10].

The involvement of the frontal cortex in Verbal Fluency tasks was initially

demonstrated through studies in brain injured patients [11]. Later on, functional

neuroactivation studies with the Verbal Fluency Test paradigm in healthy subjects

demonstrated increased blood perfusion, related to an increased cortical activity, in

the prefrontal cortex, more particular in the left DLPFC and anterior cingulate

[12,13,14].

Keeping in mind the impaired Verbal Fluency Test performance in depressed suicide

attempters, these findings seem to fit results from functional imaging studies with

PET or SPECT in non-activated conditions that indicate reduced frontal cortex

perfusion and metabolism, mainly in the left dorsolateral prefrontal cortex and

71

anterior cingulate, in depressed subjects [15,16].

Recently, our group validated a neuropsychological activation paradigm with the

Verbal Fluency Test in a subject-friendly one-day split-dose paradigm setting

SPECT with subsequent SPM analysis [17]. When normal volunteers performed the

VFT in both its modalities, an associated increase in frontal blood flow was found

However, there was a differential activation in the Letter Fluency versus the

Category Fluency test since the Category Fluency modality also demonstrated

activation in right prefrontal cortex, besides the activation in the left prefrontal cortex

and the anterior cingulate which was present in both modalities [12].

Neuropsychological activation studies following these paradigms were already

carried out in schizophrenic patients and demonstrated attenuated increase in frontal

and temporal cortex [18]. One prior SPECT neuropsychological activation study in

late life depression patients was carried out and demonstrated no advantages on the

use of a neuropsychological activation paradigm [19]. However, major shortcomings

of the study, such as the mixed paradigm design since both Letter Fluency and

category fluency were demanded simultaneously, the lack of an appropriate basal

task and an operator-dependent ROI-based analysis in stead of a SPM based analysis,

make a comparable study worthwhile.

This study firstly aimed at investigating possible differences in brain perfusion on the

subtests of Verbal Fluency in depressed patients that recently attempted suicide,

comparable to the investigation in healthy volunteers [12]. Since neuropsychological

investigations demonstrated that depressed suicide attempters were impaired in both

the Letter Fluency and the Category Fluency test [51], both subtests were included in

the study.

Secondly, we wanted to test the hypothesis that depressed suicide attempters, in

comparison to healthy volunteers, show a blunted increase in cerebral blood flow in

prefrontal structures in both the subtest modalities in the Verbal Fluency Test.

72

Methods

Subjects

Twenty patients, aged between 19 and 49 years, that were admitted to the

Department of Psychiatry at Ghent University were enrolled in the study. Inclusion

criteria for the patients were a recent suicide attempt (< 7 days) and DSM-IV criteria

[20] for major depressive disorder (DSM IV 296.2x and 296.3x), a score of 21 or

more on the Hamilton Depression Rating Scale (HAMD) [21], right-handedness

[22], Dutch as mother tongue and a Mini-Mental Examination Score of 28 or more

[23]. Exclusion criteria were other major psychiatric, neurological or medical

disorder, electro convulsion therapy in the previous 12 months, a history of alcohol

or drug abuse and the use of psychotropic drugs with anti-cholinergic properties or

benzodiazepines or other minor tranquillizers in doses exceeding the equivalent dose

of lorazepam 1 mg per day within six weeks prior to the study. Moreover, the

inclusion of patients that were on other than the aforementioned psychotropic

medication was restricted to those who were on a stable dose for at least 6 weeks.

Twenty normal control subjects, within the 18-50 age-range were enrolled in the

study. Control subjects were excluded if they showed any evidence of psychiatric or

neurologicial disorder or were taking any concurrent medication that could influence

cognition.

No subject had any previous experience with the Verbal Fluency Test.

Research was compliant with the Code of Ethics of the World Medical Association

(Declaration of Helsinki). Each subject gave informed consent, following the

guidelines of the local ethics committee.

Clinical assessment procedure

All patients underwent a complete physical and mental state examination by a board-

certified psychiatrist. The Mini-Mental State Examination was carried out in order to

rule out major cognitive failure related to confusional states or dementia [23]. The

Edinburgh Handedness Scale was used to evaluate predominant handedness [22].

Premorbid intelligence was estimated using the NLV, a Dutch version of the NART

73

[24].

Controlled Oral Word Association Test (COWA)

The subjects underwent two regional cerebral blood flow (rCBF) measurements, one

during a basal task and one during an activation task, i.e. one of the subtests of the

Verbal Fluency Test. For the activation task, the subjects were instructed to say

aloud as many words as they could think of beginning with a specified target letter

(“N”, “A”, K”, “B”) (Letter Fluency; phonological association) or a specified target

category (“animals”, “jobs”, “fruit and vegetables” and “interior and furniture”

(Category Fluency; semantic association). Target letters or categories were changed

at 1-minute intervals. The test score was defined by the number of words produced

in the 240 s period, excluding perseverations. For the basal task, subjects were

instructed to say aloud the letters of the alphabet from A to Z and numbers from 1 to

100 consecutively, during a 240 s period.

Concerning this resting task, input (auditory target) and output (saying words) was

essentially the same as in the activation task. However, as this task is executed

“automatically” and with minor effort, semantic and phonological processing is

minimized. The measurements were performed in the same order (resting task,

activation task) in all subjects. The tasks commenced 30 s prior to tracer injection

and continued throughout the 240 s period.

Healthy volunteers and depressed suicide attempters were randomly assigned to the

Letter Fluency paradigm or the category fluency SPECT activation paradigm

resulting in four groups of 10 subjects each. Subjects who were included in the

Letter Fluency brain activation SPECT paradigm completed the Verbal Fluency Test

after the imaging procedure by the Category Fluency test and vice versa (see Figure

1).

Paradigm design


tracer can be given in a split dose protocol enabling at least one repeat scan within a

74

short time period [25,26]. Tracer injection conditions were described elsewhere [12].

According to standard neuropsychological testing conditions, subjects sat at a table

in a quiet room and were informed about the testing procedure. Ten minutes

following the reading session, the first image acquisition was performed during a 20

min period. Subsequently, subjects were instructed to start the activation task and

the aforementioned procedure was repeated. The number of words generated by the

subject was assessed by the experimenter.

Image acquisition, reconstruction and SPM

All patients were injected with 2 times 555 MBq 99mTc-ethyl cysteinate dimer

(ECD, Dupont Pharmaceuticals, Brussels, Belgium). Image acquisition was

performed with a three-headed gamma camera (Toshiba CGA 9300 A) equipped

with parallel-hole low-energy high-resolution collimators. Further camera

specifications and reconstruction parameters were described elsewhere [12]. The

reconstructed data were fitted automatically onto an in-house constructed database

template positioned in Talairach co-ordinates [27]. Statistical parametric mapping

with SPM99 (Wellcome Department of Cognitive Neurology, Institute of Neurology,

London, UK) was applied to determine significant regions of increased or decreased

activity. Statistical comparisons between conditions were performed on a pixel-by-

pixel basis using t statistics, generating SPM(t) maps subsequently transformed to the

unit normal distribution SPM(Z) maps. A 'multistudy with replication of conditions'

design was used without covariates or confounders. Contrasts were defined to

evaluate activation and deactivation within the groups of volunteers and depressed

patients for both activation tasks, while also the differential activation/deactivation

differences were assessed in combined contrasts. Unless mentioned otherwise, areas

were investigated at the p height = 0.05 and p ext = 0.05 level, uncorrected for

multiple comparisons. Further specificities on the SPM analysis were described

elsewhere [12].

75

Statistical analysis

In order to evaluate differences in the distributions of demographic and clinical

measures according to diagnostic groups and Verbal Fluency paradigms, the Mann-

Whitney test was used for age, years of education, verbal IQ estimate, results of the

‘Counting + Alphabet’ baseline task and the scores on the Hamilton Depression

Rating Scale within the depression group. Gender differences were statistically

evaluated using Fisher’s exact test. In comparing verbal IQ and the ‘Counting +

Alphabet’ results between diagnostic groups, adjustment for age and years of

education was done according to analysis of covariance. The independent effects of

diagnostic group and paradigm on letter and category fluency scores were studied in

the framework of multiple regression modelling with correction for age and years of

education. A possible synergistic effect of diagnosis and paradigm on each of these

scores was studied by fitting the appropriate interaction term. Model assumptions

were visually evaluated using Pearson residuals. P-values were indicating statistical

significance if lower than α=0.05.

76

Figure 1

Design of the study. Twenty healthy volunteers and twenty depressed suicidal patients were tested with the Verbal Fluency test, resulting in testing scores on the Letter Fluency test and the category fluency test. Subjects of both diagnostic groups were randomly assigned to a Letter Fluency or a category fluency SPECT activation paradigm.

77

Results

Demographic and clinical measures

A description of the demographic and clinical measures under study in both

depressed subjects that attempted suicide and healthy volunteers in their respective

paradigms, is given in Table 1.

Healthy volunteers Depressed Subjects

in Letter Fluency

SPECT activation

paradigm

(N=10)

in Category Fluency

SPECT activation

paradigm

(N=10)

in Letter Fluency

SPECT activation

paradigm

(N=10)

in Category Fluency

SPECT activation

paradigm

(N=10)

D

e

m

o

g

r

a

p

h

i

c

&

c

l

i

n

i

c

a

l

Gender 4 M / 6F 4 M / 6F 3 M / 7F 5M / 5M Age 23.6 (3.0)a 21.8 (2.0) 31.5 (7.6) 31.9 (9.1) Years of education 15.1 (2.0) 14.7 (0.8) 14.1 (3.0) 13.5 (2.4) Verbal IQ estimate (NLV)b 102.8 (9.5) 103.5 (11.6) 98.9 (10.1) 98.7 (7.8) Counting+Alphabet (240") 413.2 (24.1) 428.1 (31.8) 395.8 (28.5) 393.8 (27.2) HAMDc -- -- 27.0 (3.8) 27.8 (3.0) Verbal Fluency Test Letter Fluency (4 items)d 62.8 (14.9) 65.4 (14.1) 42.8 (13.0) 43.3 (10.5) Category Fluency (4 items)e 73.0 (16.2) 74.3 (12.7) 55.2 (10.3) 55.4 (9.2) a Cell entries are mean (SD)

b NLV: Nederlandse Leestest voor Volwassenen (Dutch Version of the NART) c HAMD: Hamilton Depression Rating Scale d Items in Letters Fluency: “N”, “A”., K” and “B” e Items in Category Fluency: “animals”., “occupations, jobs”, “fruit and vegetables” and “interior, furniture”

Table 1

Description of demographic and clinical measures in 20 depressed suicidal patients and 20 healthy volunteers and their results on the Verbal Fluency test (VFT; COWAT).

78

Combining the subject categories proved that the distributions of gender (P=0.75),

age (P=0.46), years of education (P=0.43), verbal IQ score (P=0.82), the ‘Counting +

Alphabet’ baseline task result (P=0.66) and, within the subgroup of depressed suicide

attempters, the HAMD scores (P=0.73) were not significantly different between both

Verbal Fluency paradigm groups. The average ages in the letter and category fluency

paradigms were respectively 26.8 and 27.6 years. Obviously, the comparability of

demographic and clinical information is the natural result of the randomization

procedure performed and allows us to study the differences in demographic and

clinical characteristics between diagnostic groups in the combined sample of subjects

in both Verbal Fluency paradigms. Doing so, proved that ages were significantly

(P<0.001) higher in depressed suicide attempters (31.7 years) compared to the group

of healthy controls (22.7 years). Moreover, after correction for age, depressed

suicide attempters were found to be less educated (P=0.02) than the volunteers. The

distributions of verbal IQ scores (P=0.43) and ‘Counting + Alphabet’ baseline task

results (P=0.08) were not significantly different between diagnostic groups after

adjustment for age and educational level.

The Controlled Oral Word Association Test (COWAT)

All subjects in our study underwent both subtests, the Letter Fluency and the

category fluency part of the Verbal Fluency Test. Comparison of the Verbal Fluency

Test scores (4 items per subtest) revealed significant differences for the Letter

Fluency scores (P< 0.001) as well as the Category Fluency scores (P<0.001) between

depressive patients that attempted suicide and healthy volunteers. No significant

differences were observed in both subtests between the two paradigms. Studying the

interaction between diagnosis and paradigm on both Verbal Fluency Test scores in a

multivariate model proved that both these effects acted additive rather than

multiplicative or synergistic. Hence, the observed differences between both

depressed suicide attempters and healthy subjects in Verbal Fluency Test scores were

independent of the paradigm, and vice versa. Table 2 shows the results of multiple

regression analyses for both outcomes. After adjustment for age, educational level

and paradigm, the differences in both Verbal Fluency Test scores between depressed

79

patients and healthy controls remained highly significant. After adjustment, scores

for the Letter Fluency subtest and the category subtest were found to be respectively

26.4 and 22.5 units lower in the group with a clinical diagnosis of depression and

recent suicide attempt. (Figures 2 and 3)

Letter Fluency (4 items)a Category Fluency (4 items)b

β (SE)c td P β (SE)c td P

Intercept

Depressed vs. Healthy

CF vs. LF paradigm

Age

Years of education

37.41 (15.33)

-26.37 (5.33)

+2.36 (3.98)

+0.67 (0.35)

+0.68 (0.99)

--

-4.95

0.59

1.95

0.69

--

P<0.001

P=0.56

P=0.06

P=0.49

43.80 (14.17)

-22.54 (4.92)

+1.69 (3.68)

+0.59 (0.32)

+1.04 (0.91)

--

-4.58

0.46

1.85

1.14

--

p<0.001

p=0.65

p=0.07

p=0.26

a Items in Letters Fluency: “N”, “A”., K” and “B” b Items in Category Fluency: “animals”., “occupations, jobs”, “fruit and vegetables” and “interior, furniture” c Estimated regression coefficient and its standard error d T-statistic for testing H0: β=0

Table 2 Letter and Category Fluency test scores in depressed suicidal patients and healthy volunteers. Results of multiple regression analyses.

80

Figure 2 Means of the Letter Fluency Test scores corrected for age and education according to diagnosis and paradigms

Figure 3 Means of the Category Fluency Test scores corrected for age and education according to diagnosis and paradigms

Healthy volunteers

Healthy volunteers

Depressed

Suicidal

Depressed Suicidal

81

Evaluation of the SPM results

In answer to the first hypothesis, the comparison of the Category Fluency versus the

Letter Fluency activation in the depressed suicide attempters, regional differential

activation foci between the two tasks were evaluated. Depressed patients showed, in

the Category Fluency condition, an extra activation in the right prefrontal cortex

(Right Gyrus Frontalis Medius ) and in the left gyrus hippocampalis, compared to the

Letter Fluency condition. There was no extra prefrontal activation in the Letter

Fluency Test.

In answer to the second hypothesis, the comparison of the Category Fluency and the

Letter Fluency activation between healthy volunteers and depressed suicide

attempters, regional differential activation foci between the two populations were

evaluated.

When comparing the activated brain regions between healthy volunteers and

depressed suicide attempters in the Category Fluency paradigm, we found a

statistical significant (SPM) blunted perfusion in the depressed subjects in the left

gyrus frontalis inferior, in the right gyrus parietalis inferior and in the left and right

gyrus cinguli anterior. These regions are visually presented in the glass brain

projections in fig. 4. There are no significant regions with increased perfusion in the

depressed suicide attempters as compared to the healthy volunteers.

82

Figure 4 Comparison of means of the difference in brain perfusion (Category Fluency task versus citing the alphabet and counting as a resting task) between ten healthy volunteers and ten depressed suicidal subjects. Statistical Parametric Maps (SPM) in 3 glass brain projections. T values, uncorrected P values and activated regions are given.

When comparing the activated brain regions between healthy volunteers and

depressed suicide attempters in the Letter Fluency paradigm, we found a statistical

significant blunted perfusion in the left and right gyrus temporalis medius, in the

right gyrus cinguli anterior and in the hypothalamic region. These regions are

visually presented in the glass brain projections in fig. 5. There are no significant

regions with increased perfusion in the depressed suicide attempters as compared to

the healthy volunteers.

T * Talairach

coordinates Region

4.76 (-42, 42, 0) L G Frontal Inf

4.46 ( 21, 0, 33) R G Cinguli

4.13 (24, 12, 12) R G Cinguli

4.07 (36, -30, 39) R Lob Parietal Inf

3.75 (-6,12,-9) L G Cinguli

3.72 (3, 39, -6) R G Cinguli

* All P< 0.001

83

Figure 5 Comparison of means of the difference in brain perfusion (Letter Fluency task versus citing the alphabet and counting as a resting task) between ten healthy volunteers and ten depressed suicidal subjects. Statistical Parametric Maps (SPM) in 3 glass brain projections. T values, uncorrected P values and activated regions are given.

Within the population of healthy volunteers, the Category Fluency activation

paradigm, as compared to the Letter Fluency paradigm, showed a significant

increased perfusion in the left gyrus frontalis inferior and in the right gyrus cinguli

anterior. Within the population of depressed suicide attempters, the Category

Fluency activation paradigm, as compared to the Letter Fluency paradigm, showed a

significant increased perfusion in the gyrus frontalis medius and in left medial

temporal structures nearby the gyrus hippocampalis (Figure 6).

T Talairach

coordinates Region

4.50 * (-48, -57, -9) L G Temporal med

4.13 * ( 33, -57, 15)

3.39 ** (45, -42, 15) R G Temporal med

4.12 * (60, -40, 3) R G Temporal med

4.04 * (3, -3, -9) Hypothalamic region

3.47 ** (9, 27, 3)

3.41 ** (3, 36, -3) R G Cinguli

* P<0.001; ** P=0.001

84

Figure 6

Comparison of means of the difference in brain perfusion between the Letter Fluency task and the Category Fluency task within a group of twenty depressed suicidal subjects. Statistical Parametric Maps (SPM) in 3 glass brain projections. T values, uncorrected P values and activated regions are given.

Discussion

This study examined the functional neuroanatomy of Verbal Fluency, both the Letter

Fluency (phonological association) and the Category Fluency (semantic association)

paradigm, in a group of depressed patients that very recently attempted suicide and in

healthy volunteers.

The depressed suicide attempters, compared to healthy volunteers, showed

significant poorer word-production performance on both the Letter Fluency and the

T Talairach

coordinates Region

3.85 * (39, 27, 33) R G Frontalis Med

3.45 ** (-9, -39, 0)

2.53 ** (-15, -30, -21) L G Hippocampalis

* P<0.001; ** P<0.01

85

Category Fluency tests.

This deficit in general Verbal Fluency in depressed suicide attempters is in

agreement with most authors [28, 5, 2]. In line with literature, both healthy

volunteers and depressed suicide attempters produced fewer words in the Letter

Fluency subtest as compared to the Category Fluency subtest. Although retrieval

strategies in the Letter Fluency and Category Fluency subtests largely differ, the

scores were highly correlated in the healthy volunteers (r=0.84) as well as in the

depressed suicide attempters (r=0.77). A possible explanation for the correlation of

the score between the subtests may lie in a common non-specific factor of

psychomotor speed [29].

There were no relationships between the Verbal Fluency scores and a general

measurement of severity of depression (results not shown), a finding that is

consistent with others [5,29]. Ages were significantly different between the

subgroups. This is not likely to influence the results since age effects have been

shown only for subjects in the higher age range [4] or were not found at all in the 16-

65 age range [30]. There were no gender differences concerning the scores on Letter

Fluency and Category Fluency in both the healthy volunteers and depressed suicidal

subjects population (results not shown), a finding which is keeping with the major

normative studies [4,31].

In the evaluation of our first hypothesis, a differential activation between the

Category Fluency and the Letter Fluency in the depressed suicide attempters could

be described. The first focus of extra activation was demonstrated in the right

inferior prefrontal cortex. This region was comparable to the differential activation

between the subtests in healthy volunteers [12]. On the contrary, there was no extra

activation in the left inferior prefrontal cortex in the semantic task versus the

phonological task, although this was present in healthy volunteers. One could put

forward that depressed suicide attempters, contrary to healthy volunteers, have no

augmentation in left prefrontal blood perfusion when this region is challenged by a

task. This finding is in line with regional blood flow and metabolism studies that

suggest a dysfunction of the left inferior prefrontal cortex is involved in the

expression of depression [15,32,33]. Interestingly, depressed suicide attempters have

an augmentation in blood perfusion in the medial temporal region, a finding that was

86

not demonstrated in healthy volunteers. It seems that depressed suicidal subjects,

since their prefrontal “search” mechanisms are dysfunctional, compensate by using

more direct strategies to have entrance to their verbal hippocampal memory systems.

In a previous study we demonstrated that the Category Fluency paradigm was more

challenging to the prefrontal cortex and cingulate gyrus than the Letter Fluency in

healthy volunteers [12] and we now extend this to depressed patients that recently

attempted suicide.

In the evaluation of our second hypothesis, a differential activation in the Category

Fluency and in the Letter Fluency task between the healthy volunteers and the

depressed suicide attempters was demonstrated. We found a blunted increase in

perfusion in the anterior cingulate in depressed suicide attempters, compared to

healthy volunteers, during the activation in both the Letter Fluency and the Category

Fluency paradigm. The role of the anterior cingulate in neuropsychological

prefrontal activation tasks in healthy volunteers, such as the Stroop Colour Word

Test [34,35] or the Wisconsin Card Sorting Test (WCST) [36] in general, and

semantic and phonological processing tasks [37] in particular, is largely documented

in literature. Functions attributed to these regions include the attention to the tasks

following motivation processes in general and formulation and initiation of a strategy

to identify the appropriate words from memory in particular [38]. Patients with

anterior cingulate lesions display diminished motivation and a loss of interest [39],

which are key symptoms in depression. Neuropsychological studies with prefrontal

probes in patients populations, such as the Stroop task [40] or the Tower of London

Test in schizophrenic patients [41], and more specifically, fluency tasks in depressed

patients [42] and in schizophrenic patients [18] demonstrated a blunted increase in

perfusion or metabolism in the anterior cingulate. Hence, the blunted activation of

the anterior cingulate cortex could be indicated as a non-specific indicator – non-

specific since present across different nosological entities and across different

prefrontal activation probes – of reduced prefrontal task performance.

In this study, comparing the depressed suicide attempters to the healthy volunteers, a

blunted perfusion was demonstrated in the left inferior prefrontal cortex (LIPC) in

87

the Category Fluency paradigm, more specifically in the left gyrus frontalis inferior.

The role of the LIPC in semantic fluency task was previously demonstrated in other

resembling semantic tasks in functional brain imaging studies [37,43]. Also in other

patient populations, prefrontal challenge tests showed an absent or blunted perfusion

increase in prefrontal cortex, such as in the Wisconsin Card Sorting Test paradigm

[44] and the Verbal Fluency Test [45] in schizophrenic patients. Not surprisingly,

we found no blunted increase in flow in the depressed suicidal subjects in the Letter

Fluency paradigm, since Letter Fluency seems not to depend on activation of the

LIPC, as was demonstrated in healthy volunteers. The reduced performance of

depressed suicide attempters on the Category Fluency can be related to a reduced

ability of the LIPC semantic executive retrieval system.

The depressed suicide attempters, investigated with the Letter Fluency paradigm,

showed a blunted perfusion in the left and right gyrus temporalis medius. The role of

the bilateral temporal cortices in a Letter Fluency in healthy volunteers was

demonstrated in a previous functional neuroimaging study in healthy volunteers [13]

and a blunted increase in perfusion was found in left temporal cortex in

schizophrenic patients [46]. These results are in favor of a hypothetical

dysconnectivity between the DLPFC and the temporal cortex in depression, in

analogy to that in schizophrenia [46].

A possible bias could lie in the fact that the healthy volunteers and the depressed

suicidal subjects were divided at random in possibly non-uniform subgroups. This is

unlikely to influence our test results since comparison of the respective subgroups

(healthy volunteers in Letter Fluency activation paradigm versus healthy volunteers

in Category Fluency activation paradigm and depressed suicide attempters in the

Letter Fluency activation paradigm versus depressed suicide attempters in the

Category Fluency paradigm) showed no differences concerning the demographic,

clinical and test variables. Some patients were receiving a stable dose of

psychotropic medication at the moment of inclusion in the study. Two patients in the

Letter Fluency group and three patients in the Category Fluency group were on stable

88

doses of antidepressants (selective serotonin reuptake inhibitors). No other

medication with central nervous effects was taken in a period of six months prior to

the testing procedure. No one of the volunteers took any psychotropic medication in

a six months period prior to the testing procedure. This raises the question of a

possible influence of these drugs on their test performance. We can not rule out their

potential confounding role, however, test results of subjects on psychotropic drug

treatment fell in the same range of the drug-free patients and our results are

consistent with the data about Verbal Fluency Test reported in drug-free patient

studies [47,48]. Moreover, no effects of anti-depressants on executive function are

described [47, 52].

Since only depressed inpatients were included, a possible selection bias could lie in

a reduced interest to participate and put forth good effort in the study. The non-

specific variables of effort and cooperation are difficult to evaluate since it’s difficult

to objectively differentiate between a willed lack of effort and an unwilled reduced

capacity to take effort. In our study, subjects were observed to complete the

neuropsychological test without remarkable problems of cooperation or fatigue.

We cannot differentiate if our findings are linked to the depressed state, to suicidality

or to both. In our opinion, this differentiation is not important since reduced

executive functioning is clinically present in both conditions. Moreover,

neuropsychological abnormalities and the corresponding functional neuroimaging

status were shown to be symptom-specific and not diagnosis-specific [49] and hence,

transnosologically present.

Conclusion

The reduced ability of depressed suicidal patients to perform the Verbal Fluency

Test, associated with a blunted increase in blood flow in the prefrontal cortex, can

have clinical significance. Our data fit the results of a neuropsychological follow-up

study that related a Verbal Fluency Test deficit in depressed patients to a reduction of

intentional behaviour and lack of initiative, rather than to a dysfunction of other

89

cognitive processes [29].

A reduced prefrontal functioning hampers the ability to mobilize effective

behavioural strategies in an unstructured situation in which individuals must resolve

conflicts between competing choices [50]. From a clinical viewpoint, this prefrontal

deficit in depressed suicidal patients can possibly can possibly serve as a functional

anatomical correlate of cognitive psychological theories that describe the relationship

between prefrontal executive dysfunctions, hopelessness and suicidal behaviour.

90

References

1. Fossati P, Amar G, Raoux N, Ergis A, Allilaire J. Executive functioning and verbal

memory in young patients with unipolar depression and schizophrenia. Psychiatry Res

1999; 89: 171-187.

2. Lafont V, Medecin I, Robert PH, Beaulieu FE, Kazes M, Danion JM, Pringuey D,

Darcourt G. Initiation and supervisory processes in schizophrenia and depression.

Schizophr Res 1998; 34: 49-57.

3. Frith CD, Dolan RJ. Higher cognitive processes. In: Frackowiak RS, Friston KJ, Frith

CD, Dolan RJ, Mazziota JC, eds. Human Brain Function. San Diego: Academic Press;

1997: 329-365.

4. Spreen O, Strauss E A compendium of neuropsychological tests. New York / Oxford:

Oxford University Press, 1997.

5. Degl'Innocenti A, Agren H, Backman L. Executive deficits in major depression. Acta

Psychiatr Scand 1998; 97: 182-188.

6. Franke P, Maier W, Hardt J, Frieboes R, Lichtermann D, Hain C. Assessment of

frontal lobe functioning in schizophrenia and unipolar major depression.

Psychopathology 1993; 26: 76-84.

7. Kolb B, Whishaw IQ. Performance of schizophrenic patients on tests sensitive to left

or right frontal, temporal, or parietal function in neurological patients. J Nerv Ment Dis

1983; 171: 435-443.

8. Miller WR. Psychological deficit in depression. Psychol Bull 1975; 82: 238-260.

9. Stuss DT, Gow CA, Hetherington CR. "No longer Gage": frontal lobe dysfunction and

emotional changes. J Consult Clin Psychol 1992; 60: 349-359.

10. Bartfai A, Winborg IM, Nordstrom P, Asberg M. Suicidal behavior and cognitive

flexibility: design and Verbal Fluency after attempted suicide. Suicide Life Threat

Behav 1990; 20: 254-256.

11. Perret E. The left frontal lobe of man and the suppression of habitual responses in

verbal categorical behaviour. Neuropsychologia 1974; 12: 323-330.

91

12. Audenaert K, Brans B, Van Laere K, Lahorte P, Versijpt J, van Heeringen C, Dierckx

R. Verbal fluency as a prefrontal activation probe: a validation study using 99mTc-

ECD brain SPECT. Eur J Nucl Med 2000; 27: 1800-1808.

13. Frith CD, Friston KJ, Liddle PF, Frackowiak RS. A PET study of word finding.




word fluency task. Neuropsychiatry Neuropyschology and Behaviour Neurology 1991;

4: 305-316.

15. Baxter LR, Schwartz JM, Phelps EA, Mazziota JC, Guze BH, Selin SE, Gerner RH,

Sumida RM. Reduction of prefrontal cortex glucose metabolism common to three

types of depression. Arch Gen Psychiatry 1989; 46: 243-250.

16. Drevets W, Videen T, Price J, Preskorn S, Carmichael S, Raichle M. A functional

anatomical study of unipolar depression. J Neurosci 1992; 12: 3628-3641.

17. Lahorte P, Vandenberghe S, Van Laere K, Audenaert K, Lemahieu I, Dierckx R.

Assessing the performance of SPM analyses of SPECT. Neuroimage 2000; 12: 757-

764.




19. Philpot MP, Banerjee S, Needham-Bennett H, Costa DC, Ell PJ. 99mTc-HMPAO

single photon emission tomography in late life depression: a pilot study of regional

cerebral blood flow at rest and during a verbal fluency task. J Affect Disord 1993; 28:

233-240.

20. American Psychiatric Association Diagnostic and Statistical Manual of Mental

Disorders: DSM-IV. Washington: American Psychiatric Association, 1994.

21. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960; 23:

56-62.

22. Oldfield R. The assessment and analysis of handedness: The Edinburgh Inventory.

92


23. Folstein MF, Folstein SE, McHugh PR. "Mini-Mental State": A practical method for

grading the cognitive state of the patients for the clinician. J Psychiatr Res 1975; 12:

189-198.

24. Nelson HE National Adult Reading Test (NART): Test manual. Windsor, UK: NFER

Nelson, 1991.

25. Shedlack KJ, Hunter R, Wyper D, McLuskie R, Fink G, Goodwin GM. The pattern of



26. Holm S, Madsen P, Sperling B, Lassen N. Use of the 99m Tc-bicisate in activation

studies by split-dose technique. J Cereb Blood Flow Metab 1994; 14 Suppl 1: S115-

120.

27. Talairach J, Tournoux P Co-planar stereotactic atlas of the Human Brain. Stuttgart:

Thieme Medical Publishers, 1988.

28. Albus M, Hubmann W, Wahlheim C, Sobizack N, Franz U, Mohr F. Contrasts in

neuropsychological test profile between patients with first-episode schizophrenia and

first-episode affective disorders. Acta Psychiatr Scand 1996; 94: 87-93.

29. Trichard C, Martinot JL, Alagille M, Masure MC, Hardy P, Ginestet D, Feline A.

Time course of prefrontal lobe dysfunction in severely depressed inpatients: a

longitudinal neuropsychological study. Psychol Med 1995; 25: 79-85.

30. Lindeboom J, Jonker C Amsterdamse Dementie-Screeningstest. Handleiding. Lisse:

Swets & Zeitlinger, 1989.

31. Yeudall LT, Fromm D, Reddon JR, Stefanyk WO. Normative data stratified by age

and sex for 12 neuropsychological tests. J Clin Psychol 1986; 42: 918-946.

32. Martinot J, Hardy P, Féline A, Huret J, Mazoyer B, Attar-Levy D, Pappata S, Syrota

A. Left prefrontal glucose hypometabolism in the depressed state: a confirmation. Am

J Psychiatry 1990; 147: 1313-1317.

33. Bench C, Friston K, Brown RG, Scott JG, Frackowiak R. The anatomy of melancholia

93

- focal abnormalities of cerrebral blood flow in major depression. Psychol Med 1992;

22: 607-615.

34. Audenaert K, Lahorte P, Brans B, Van Laere K, Goethals I, van Heeringen C, Dierckx

R. The classical Stroop Interference task as a prefrontal activation probe: a validation

study using 99mTc-ECD brain SPECT. Nuclear Med Comm 2001; 22: 145-150.

35. Bush G, Whalen P, Rosen B, Jenike M, McInerney S, Rauch S. The Counting Stroop:

an interference task specialized for functional neuroimaging-validation study with

functional MRI. Human Brain Mapping 1998; 6: 270-282.



Research: Neuroimaging 1998; section 82: 47-52.

37. Poldrack R, Wagner A, Prull M, Desmond J, Glover G, Gabrieli J. Functional

specialization for semantic and phonological processing in the left inferior prefrontal

cortex. Neuroimage 1998; 10: 15-35.

38. Warburton E, Wise R, Price C, Weiller C, Hadar U, Ramsay S, Frackowiak R. Noun and verb retrieval by normal subjects Studies with PET. Brain 1996; 119: 159-179.

39. Devinsky O, Morrell MJ, Vogt BA. Contributions of anterior cingulate to behavior.

Brain 1995; 118: 279-306.

40. Carter CS, Mintun M, Nichols T, Cohen JD. Anterior cingulate gyrus dysfunction and

selective attention deficits in schizophrenia: [15O]H2O PET study during single-trial

Stroop task performance. Am J Psychiatry 1997; 154: 1670-1675.

41. Andreasen NC, Rezai K, Alliger R, Swayze VW, Flaum M, Kirchner P, Cohen G,

O'Leary DS. Hypofrontality in neuroleptic-naive patients and in patients with chronic

schizophrenia. Assessment with xenon 133 single-photon emission computed

tomography and the Tower of London. Arch Gen Psychiatry 1992; 49: 943-958.

42. Smith KA, Morris JS, Friston KJ, Cowen PJ, Dolan RJ. Brain mechanisms associated

with depressive relapse and associated cognitive impairment following acute

tryptophan depletion. Br J Psychiatry 1999; 174: 525-529.

43. Thompson-Schill S, D'Esposito M, Aguirre G, Farah M. Role of the left inferior

prefrontal cortex in retrieval of semantic knowledge: A reevaluation. Proceedings of

94

the National Academy of Sciences of the United States of America 1997; 94: 14792-

14797.

44. Parellada E, Catafau AM, Bernardo M, Lomena F, Gonzalez-Monclus E, Setoain J.

Prefrontal dysfunction in young acute neuroleptic-naive schizophrenic patients: a

resting and activation SPECT study. Psychiatry Research: Neuroimaging 1994; 55:

131-139.

45. Curtis VA, Bullmore ET, Morris RG, Brammer MJ, Williams SC, Simmons A,

Sharma T, Murray RM, McGuire PK. Attenuated frontal activation in schizophrenia

may be task dependent. Schizophr.Res 1999; 37: 35-44.

46. Friston KJ, Frith CD. Schizophrenia: a disconnection syndrome. Clin Neurosci 1995;

3: 89-97.

47. Killian GA, Holzman PS, Davis JM, Gibbons R. Effects of psychotropic medication

on selected cognitive and perceptual measures. J Abn Psychol 1984; 1: 58-70.

48. Wolfe J, Granholm E, Butters N, Saunders E, Janowsky D. Verbal memory deficits

associated with major affective disorders: a comparison of unipolar and bipolar

patients. J Affect Disord. 1987; 13: 83-92.

49. Drevets W. Functional neuroimaging studies of depression: the anatomy of

melancholia. Ann Rev Med 1999; 49: 837-844.

50. Bechara A, Damasio H, Tranel D, Damasio A. Dissociation of Working Memory from decision making within human prefrontal cortex. J Neurosci 1998; 18: 428-437

51. Keilp J, Sackheim H, Brodsky B, Oquendo M, Malone K, Mann J. Neuropsychological

dysfunction in depressed suicide attempters. Am J Psychiatry 2001; 158: 735-741.

52. Kerr J, Fairweather D, Hindmarch I. Effects of fluoxetine on psychomotor

performance, cognitive function and sleep in depressed patients. Int Clin

Psychopharmacol 1993; 8: 341-343.

95

CHAPTER 4:

DECREASED FRONTAL SEROTONIN 5-HT2A RECEPTOR BINDING

INDEX IN DELIBERATE SELF-HARM PATIENTS.

Kurt Audenaert, Koen Van Laere, Filip Dumont, Guido Slegers, John Mertens, Cees

van Heeringen, Rudi Dierckx

(European Journal of Nuclear Medicine, 2001; 28(2): 175-182 )

Summary

Studies of serotonin metabolites in body fluids in attempted suicide patients and of

post-mortem brain tissue of suicide victims have demonstrated the involvement of

the serotonergic neurotransmission system in the pathogenesis of suicidal behaviour.

Recently developed neuroimaging techniques offer the unique possibility of

investigating in vivo the functional characteristics of this system. In this study the 5-

HT2a receptor population of patients who had recently attempted suicide was studied

by means of the highly specific radio-iodinated 5-HT2a receptor antagonist 4-amino-

N-[1-[3-(4-fluorophenoxy)propyl]-4-methyl-4-piperidinyl]-5-iodo-2-

methoxybenzamide or 123I-5-I-R91150. Nine patients who had recently (1-7 days)

attempted suicide and 12 age-matched healthy controls received an intravenous

injection of 185 MBq 123I-5-I-R91150 and were scanned with high-resolution brain

single-photon emission tomography (SPET). Stereotactic realigned images were

analysed semi-quantitatively using predefined volumes of interest. Serotonin binding

capacity was expressed as the ratio of specific to non-specific activity. The

cerebellum was used as a measure of non-specific activity. An age-dependent 5-HT2a

binding index was found, in agreement with previous literature. Deliberate self-harm

patients had a significantly reduced mean frontal binding index after correction for

96

age (P=0.002) when compared with controls. The reduction was more pronounced

among deliberate self-injury patients (DSI) (P<0.001) than among deliberate self-

poisoning patients (DSP). Frontal binding index was significantly lower in DSI

patients than in DSP suicide attempters (P<0.001). It is concluded that brain SPET of

the 5-HT2a serotonin receptor system in attempted suicide patients who are free of

drugs influencing the serotonergic system shows in vivo evidence of a decreased

frontal binding index of the 5-HT2a receptor, indicating a decrease in the number

and/or in the binding affinity of 5-HT2a receptors.

Introduction

The involvement of serotonin (5-HT), and specifically of the serotonin-2 receptor (5-

HT2), in the pathophysiology of suicide and suicidality has been the target of

considerable research in the past decade. Both indirect measurements, comprising

levels of 5-HT and its metabolite 5-HIAA in cerebrospinal fluid and in blood [1,2],

challenge tests for the serotonergic system [3], and quantification of 5-HT2 receptors

on blood platelets [4,5], and direct measurements of brain serotonin function and

receptor status in post-mortem research [6,7,8] have yielded varying results. The

interpretation of these results is limited by the questionable validity of peripheral

measures as markers of cerebral activity [9]. Post-mortem cerebral study results are

also limited by lack of sampling from multiple regions [7] and rapid alterations in

neurotransmitter concentration post-mortem [10]. Other reasons for the variability in

the study results are the use of divergent classification systems for psychiatric

diagnoses, heterogeneity with regard to the nature of suicidal behaviour, and

treatment with antidepressant drugs or neuroleptics at the time of the suicide attempt

[11].

Functional imaging techniques, such as positron emission tomography (PET) and

single-photon emission tomography (SPET), using specific 5-HT2 receptor ligands,

make it possible to evaluate in vivo the receptor binding index in patients with mood

disorders and patients following a suicide attempt. Studies in unmedicated depressed

97

patients found either no alteration or a decrease in the 5-HT2 receptor binding index

[12,13,14,15]. This is the first report on functional imaging of the 5-HT2 receptor in

patients shortly after a suicide attempt.

Preliminary research in healthy subjects has indicated that radio-iodinated 4-amino-


methoxybenzamide, or 123I-5-I-R91150, is a suitable ligand for imaging 5-HT2a

receptors in vivo. It binds reversibly and with high affinity in vitro to 5-HT2a

receptors [16]. On average, 2% of a bolus dose of 123I-5-I-R91150 is taken up by the

brain [17]. Effective blockade of 5-HT2a receptors in vivo was demonstrated in a

study of schizophrenic patients treated with risperidone or clozapine [18].

In the present study, the 5-HT2a receptor in the brain of patients who recently

attempted suicide was studied by means of this 5-HT2a receptor antagonist using

brain SPET. It was hypothesised that differences in the 5-HT2a binding index in

frontal cortex exist between suicide attempters and normal controls and furthermore

between deliberate self-poisoning and deliberate self-injury patients.

Methods

Patients

Patients were included in the study if they were aged between 18 and 60 years and if

they were admitted following a suicide attempt. Suicide attempts included deliberate

self-poisoning (DSP) and self-injury (DSI). The term DSP was used to describe the

ingestion of more than the prescribed amount of medical substances, or the ingestion

of substances never intended for human consumption, irrespective of whether harm

was intended. The term DSI was used to describe any intentional self-inflicted injury,

irrespective of the apparent purpose of the act [19]. Exclusion criteria were (1)

electroconvulsive therapy in the preceding year, (2) antidepressant treatment in the

preceding 6 months, (3) neuroleptic treatment in the preceding 6 months, (4) lithium

or carbamazepine treatment in the preceding 6 months, (5) major medical or

neurological disorder, (6) pregnancy or lactation period, (7) substance abuse, and (8)

a Mini-Mental State [20] score less than 28.

98

The study group consisted of seven men and two women. Six patients attempted

suicide by means of DSP and three by means of DSI. All patients were free of

psychotropics for at least 6 months before scanning, except for benzodiazepines

(n=4), barbiturates (n=1) or anti-epileptics (n=1), which were ingested while

attempting suicide in the DSP group.

Patients were screened for alcohol intoxication and signs of chronic alcohol abuse by

means of blood alcohol level on admission, mean erythrocyte corpuscular volume

(MCV) and gamma-glutamyl transferase (γ-GT). Two patients were positive on

alcohol screening (1.1 and 1.2 gram promille) and had normal MCV and γ-GT. All

other patients also had normal MCV and γ-GT.

Patients were clinically evaluated by two senior psychiatrists (K.A., C.V.H.) and

were assigned DSM-IV diagnoses at the time of admission. Four patients were

diagnosed as depressed and had a score on the Hamilton Depression Rating Scale

[21] of 24 or more. Two patients reported one prior suicide attempt.

Healthy volunteers

Twelve age-matched healthy volunteers, six men and six women, were recruited

from among the hospital staff. These subjects had no psychiatric or medical history,

nor a family psychiatric history. None used psychotropics or other relevant

medication or abused illegal drugs. All had a normal physical examination.

Ethical approval for the study was granted by the Local Ethics Committee. Both

suicide attempters and healthy individuals provided written informed consent to

participation in the study.

Tracer 123I-5-I-R91150 was synthesised by electrophilic substitution on the 5-position of the

methoxybenzamide group of R91150, followed by purification with high-

performance liquid chromatography. The product had a radiochemical purity of more

than 99% and was negative for bacteria and pyrogen. A specific activity of

10 Ci/µmol was obtained.

99

The tracer is a 5-HT2a antagonist with high affinity (Kd=0.11 nM) and selectivity for

5-HT2a receptors. The selectivity of the ligand for 5-HT2a receptors in relation to

other neurotransmitter receptors such as other 5-HT receptors, including 5-HT2c,

dopamine receptors, adrenoreceptors and histamine receptors, is at least a factor of

50. The tracer was displaceable with ketanserin [16, 22].

SPET scanning

Thyroid blockade was achieved by administration of a single oral dose of 100 mg

potassium iodide prior to injection. All subjects received an intravenous injection of

185 MBq 123I-5-I-R91150 in normal sitting conditions. SPET scanning was

performed at the Ghent University Hospital Division of Nuclear Medicine using a

triple-headed high-sensitivity high-resolution Toshiba gamma camera GCA-9300

with fan-beam collimation. For 123I, the resulting transaxial image resolution is

9.5 mm full-width at half-maximum.

Since sequential dynamic SPET brain scans have shown that the cortico-cerebellar

ratio reaches a plateau between 90 and 110 min, reflecting pseudo-equilibrium, and

remains stable thereafter for up to 8 h [17], acquisition was started between 110 and

140 min after tracer injection. Time activity curves of a representative healthy

volunteer (male, aged 25) are presented in Fig. 1A (with the corresponding frontal

cortex-cerebellum ratio of activity in Fig. 1B).

100

Time (hours)

3,02,52,01,51,0,50,0

Fro

nta

l:cere

bella

r ra

tio

2,2

2,0

1,8

1,6

1,4

1,2

1,0

,8

Figure 1

A Time-activity curves of a representative healthy volunteer (male, age 25), expressed as

mean activity per volume unit (counts/voxel/minute), normalized for decay, versus time

(open squares, frontal cortex; asterisks, cerebellum). B Frontal cortex-cerebellum ratio of

activity (a relative index of specific binding) for the same volunteer

A transmission scan (TCT scan) was acquired before the emission scan, using three

gadolinium-153 rod sources. This scan was used for subsequent image co-

registration.

Emission images were acquired during 40 min. The whole brain volume was

acquired within the single scanning session.

Estimation of binding index

Analysis of the scans was performed blind to patient status. Images were corrected

for scatter and attenuation. Mean images of the 5-HT2a receptor binding index for

normal volunteers and for DSP and DSI patients are depicted in Fig. 2.

Time (hours)

3,02,52,01,51,0,50,0

Mean u

pta

ke (

counts

/voxel)

70

60

50

40

30

20

10

0

101

Figure 2

A MRI template showing transverse, coronal and sagittal sections. B-D Cerebral uptake of 5-

HT2a ligand 123I-5-I-R91150 (transverse, coronal and sagittal sections) in healthy volunteers

(B, mean image), deliberate self-poisoning (DSP) suicide attempters (C, mean image) and

deliberate self-injury (DSI) suicide attempters (D, mean image)

103

After automatic image co-registration to stereotactic space (BRASS, Nuclear

Diagnostics), a predefined volume of interest (VOI) analysis was performed with 12

cortical regions. Radioactivity estimates in the cortex were assumed to represent total

ligand binding (specific + non-specific binding + free ligand) [18]. Since the

cerebellum is devoid of serotonin receptors [23] and therefore represents non-

specific activity, calculation of relative indices of specific binding index (BP) was

done by VOI normalisation to the activity per volume element in the cerebellum.

Under these pseudo-equilibrium circumstances, this binding index is directly related

to the in vivo receptor density (Bmax) and affinity (KD). Binding index was defined as

(target activity - background activity in brain)/(background activity), which was

operationally estimated as (counts/pixel in frontal cortex)/(counts/pixel in

cerebellum).

Statistical methods

The equality of age distributions between diagnostic categories was evaluated

according to the Kruskal-Wallis test. As binding index levels were normally

distributed, analysis of variance was used to compare mean levels between

categories. Linear adjustment for age was done according to analysis of co-variance.

An a priori level of α=0.05 was chosen to indicate statistical significance. Model

assumptions were checked by graphical inspection of Pearson residuals. None of the

second-order interaction terms between diagnostic group, age and gender turned out

to be significant at the 0.10 level.

Results

Effect of age and gender

The sample of 21 subjects under study consisted of 13 men and 8 women. They were

on average 30.4 years old (SD 9.2 years), with ages ranging from 19 to 48 years.

Although statistically not significant, mean ages were slightly different between the

104

three study groups, i.e. 29.0, 30.3 and 36.5 years for volunteers, patients with DSP

and patients with DSI, respectively. Since previous research has suggested that BP

levels decrease with age [24], age was taken into account as a potential confounding

variable in our analyses. Therefore age correction was performed applying linear

correction. All values were corrected to the mean age of the group, i.e. 29.4 years. In

our study, a significant correlation between age and global frontal receptor binding

index was demonstrated in the whole group. Since the study groups were quite

comparable (Fisher's Exact Test P=0.37) concerning gender and no effect of gender

on 5-HT2a binding index with this tracer has been demonstrated in our database of

healthy volunteers and by others has been found [24], no adjustments for gender

were made in further analyses. However, one study with another 5-HT2 receptor

ligand showed a higher 5-HT2 binding capacity in men [45]. Additional details in

respect of the patients are shown in Table 1.

sex age Clinical Disorder (1)

Personality disorder (1)

HAMD (3)

Type (4)

Means (3) Time (5)

(days)

M 37 Major Depressive Disorder - 25 DSP intoxication BDZ 7

F 19 Adjustment Disorder Borderline 10 DSP intoxication barbiturates

2

M 29 Major Depressive Disorder - 29 DSP intoxication phenytoin

6

M 34 Adjustment Disorder Narcissistic 11 DSP self-injection insulin

1

F 48 Major Depressive Disorder - 27 DSI wrist-cutting 7

M 19 Adjustment Disorder Narcissistic 8 DSI wrist-cutting 3

M 19 Brief Psychotic Disorder - 7 DSP intoxication BDZ 3


M 43 Adjustment Disorder Dependent 8 DSI Strangulation 1

(1) DSM, Diagnostic and Statistical Manual - Fourth Edition; (2) HAMD, Hamilton Depression Rating Scale; (3) DSP, deliberate self-poisoning; DSI, deliberate self-injury; (4) BDZ, benzodiazepines; (5) Interval between suicide attempt and image acquisition

Table 1 Demographic characteristics, clinical diagnoses, description of suicide attempt and interval until scan acquisition in patients with deliberate self-harm and normal volunteers

105

Frontal binding index

The frontal binding index of the volunteers and patients is presented as a scatterplot

in Fig. 3. Frontal BP levels were significantly lower in suicide attempters (mean

0.39, SE 0.04) than in volunteers (mean 0.68, SE 0.04) (P<0.001). A significant

difference in binding index between normal volunteers and deliberate self-harm

patients in the left (t=14.2; P=0.001) and right (t=8.9; P=0.008) frontal cortex was

found.

Figure 3

Scatterplot of the binding index of the frontal cortex in healthy volunteers, deliberate self-poisoning patients (DSP) and deliberate self-injury patients (DSI)

Figure 4

Age-adjusted mean levels of frontal 5-HT2a binding index in the various subgroups

100

120

140

160

180

Volunteers Deliberate Self

Poisoning

Deliberate Self Injury

Bin

din

g I

nd

ex (

% c

ere

be

llu

m)

Volunteers Deliberate Self Deliberate Self

Poisoning Injury

106

In Table 2 the results of the co-variance model fit are given, comparing BP levels

between the three study groups after linear adjustment for age. Both DSP (-0.14) as

DSI (-0.46) suicide attempters had significantly lower BP levels compared with the

volunteers. In this sample of 21 subjects, the model explained 73% of the variability

in binding index levels. As shown in Fig. 4, mean BP levels in the three study groups

(volunteers, DSP and DSI), adjusted for age, were 0.68, 0.53 and 0.21, respectively.

A significant difference between the three groups (F=21.0; P<0.001) with respect to

the total frontal (F=21.0; P<0.001), the left (F=17.4; P<0.001) and the right frontal

cortex (F=18.4; P<0.001) was found (Table 3). Post hoc multiple comparison

indicated a significant difference in binding index between normal volunteers and

DSI patients in the total frontal cortex (P=0.041), which turned out to be significant

only in the left (P=0.022), and not the right (P=0.137), frontal cortex. Post hoc

multiple comparisons between the volunteers and the DSI group yielded significant

differences for the total, left and right frontal cortex (all P<0.001). Post hoc multiple

comparisons between the DSP and the DSI group yielded significant differences for

the total (P<0.001), left (P=0.003) and right (P<0.001) frontal cortex binding index

between volunteers and the DSI population.

Estimated β (SE) t-value Significance

Intercept Age (years)

DSP vs. Volunteers DSI vs. Volunteers

188.06 (9.36) -0.69 (0.30)

-14.24 (6.06) -46.34 (8.17)

20.08 -2.30

-2.35 -5.67

P<0.0001 P=0.03

P=0.03 P<0.0001

Table 2 Frontal binding index levels in suicide attempts and volunteers (analysis of co-variance)

107

Due to the limitations of the SPET technique in terms of spatial resolution, we did

not report on the subregions of the frontal cortices that were significantly different

with regard to binding index between the three populations. Since we analysed them

initially, the binding index of the subregions is presented in Table 3.

Populations Statistics

Attempted Suicide Patients T-test ANOVA

Volun-teers

DSH DSP DSI T Sig F Sig Multiple Comparison

L frontal cortex

total 0.69

0.41 0.52 0.19 14.2 0.001 17.4 0.000 VOL > DSP> DSI

L orbitofrontal cortex 0.71

0.55 0.62 0.43 1.7 0.216 1.3.0 0.298

L dorsolat prefrontal cortex 0.69

0.14 0.53 0.14 12.7 0.002 18.7 0.000 VOL > DSP> DSI

L frontal cortex remaining 0.66

0.36 0.46 0.17 17.2 0.001 16.1 0.000 VOL > DSP> DSI

R frontal cortex total 0.67

0.42 0.56 0.13 8.9 0.008 18.4 0.000 VOL = DSP> DSI

R orbitofrontal cortex 0.70

0.54 0.62 0.36 2.0 0.170 2.1 0.147

R dorsolat prefrontal cortex 0.68

0.41 0.57 0.08 8.5 0.009 19.9 0.000 VOL = DSP> DSI

R frontal cortex remaining 0.64

0.39 0.54 0.10 8.0 0.011 17.5 0.000 VOL = DSP> DSI

L + R frontal cortex 0.68

0.39 0.53 0.13 12.3 0.002 21.0 0.000 VOL > DSP> DSI

Table 3 Comparison of binding indices between volunteers and deliberate self-harm patients (DSH) and between volunteers, deliberate self-poisoning (DSP) and deliberate self-injury (DSI) patients. Post hoc multiple comparisons (Bonferroni corrected) are presented and indicate significant differences between the subgroups

108

Discussion

This is the first in vivo study of 5-HT2a cerebral receptors in the prefrontal cortex of

deliberate self-harm patients. When compared with normal controls, a significantly

reduced 5-HT2a binding index in frontal cortex was found. The reduction was

significant in both the left and the right frontal cortex. A significantly larger

reduction in the binding index was found in DSI patients than in DSP patients.

Pairwise comparison between the three populations yielded significant differences

for all comparisons, with the exception of the right frontal cortex binding index

between the volunteers and the DSP population. We have no reasonable explanation

for the latter finding, but expect that the difference between the normal volunteers

and the DSP patients will become significant when the number of patients increases.

Although only four of the nine patients were diagnosed as depressed, our results are,

at least partially, in keeping with the findings of other functional imaging studies in

unmedicated depressed patients. Using the highly specific 5-HT2a PET radioligands 18F-altanserin and 18F-setoperone, a significant decrease in the receptor binding index

in frontal cortex [12, 14] and an unaltered receptor binding index were demonstrated,

respectively [15]. Using the less specific 5-HT2 SPET ligand 123I-ketanserin in

depressed patients, D'Haenen et al. found an increase in the left-to-right prefrontal

cortex binding index, unlike the findings of our study [25]. However, it has to be

noted that 123I-ketanserin has a low signal-to-noise ratio in comparison with 123I-I-

R91150.

Results from in vitro and indirect 5-HT2 measurement techniques are inconclusive.

Post-mortem measurements of 5-HT2 and 5-HT2a receptor density, using various

ligands with autoradiography, in depressed or non-depressed suicide victims have

yielded discrepant results. Most studies have demonstrated an increase in 5-HT2 or 5-

HT2a receptor binding, more specifically in the frontal cortex [6,26,27] of depressed

patients who committed suicide but also in the prefrontal cortex of schizophrenic

patients [28]. Other studies have found no significant differences in 5-HT2 or 5-HT2a

receptor density in the prefrontal areas between normal controls and suicide victims,

irrespective of psychiatric diagnosis [29,30], or between controls and depressed

patients [7,31,32,33]. There have been reports of significant decreases in Bmax and

KD of the 5-HT2a receptor, most markedly in the prefrontal cortex of depressed but

109

antidepressant-free suicide victims who died violently [34], and of the 5-HT2

receptor in the prefrontal cortex of suicide victims [8].

Indirect estimation of the BP of brain 5-HT2 receptors can be done by platelet 5-HT2

receptor measurement [5]. Elevated platelet 5-HT2 receptor binding has commonly

been found in depression, suicidal states and depression with recent suicide attempt,

and has been attributed to a significant increase in receptor density, with no

significant change in affinity (KD) [4, 35]. Other studies, however, have not found

changes in the concentration of 5-HT2 receptors [36].

In the current study, a significant negative association between age and receptor

binding index in frontal cortex was demonstrated in the whole group and in the group

of suicide attempters. The magnitude of the decrease (6.9%±3.0% per age decade) is

in agreement with previously published studies [23].

A possible confounder of the study results lies in the fact that a physical trauma, due

to the recent suicide attempt, could have had an impact on the binding index of the

tracer due to a decrease in cerebral blood flow. We cannot rule out this possibility;

however, our results were fairly homogeneous irrespective of the method used in the

suicide attempt. From a theoretical perspective, the results could be artifactual if the

group of normal volunteers were to have had a slower metabolic rate for the tracer

than the deliberate self-harm group. However, there are no indications in the

literature to support this possible confounder.

In evaluating the data, one must also pay attention to the medication and alcohol that

were ingested prior to or on the occasion of the suicide attempt. Indeed, a further

possible confounder for the study could reside in an impact of consumption of large

amounts of alcohol on the clearance of the ligand, possibly by induction of liver

enzymes. Moreover, patients who are entrapped in a crisis often seek relief in

alcohol. In our study sample, two patients were positive on alcohol screening (>0.8

mg/100 ml) when admitted to the hospital. Blood tests on mean erythrocyte

corpuscular volume and γ-GT in all patients were normal. Patients did not use any

alcohol during their stay in the hospital in the period between admission and

scanning. None of the patients had a history of alcohol abuse. Hence, it is unlikely

that results were confounded by an impact of abuse of large amounts of alcohol on

110

clearance of the ligand.

Concerning psychotropic medication, it must be noted that some of the patients

attempted suicide by means of DSP using psychotropics, and hence the possible

effects of benzodiazepines, barbiturates and anti-epileptics on the results must be

considered. Benzodiazepines have been reported to increase 5-HT2 receptor numbers

in animal studies [37], but an absence of any effect on 5-HT2 receptor numbers has

also been reported [38]. An effect of benzodiazepine withdrawal could not be ruled

out, although patients showed no clinical withdrawal symptoms in the period

following the intoxication. In this study and in a 18F-setoperone study in depressed

patients [14], the 5-HT2a binding index fell within the same range in patients who

received benzodiazepines and patients who were free of benzodiazepines. Hence, it is

unlikely that a decrease in binding index, as observed in our study, might be induced

by the ingestion of benzodiazepines. There are no reports of effects of barbiturates or

phenytoin on 5-HT2a receptor binding characteristics.

Patients were only enrolled in our study when there was evidence that they had not

used neuroleptics or antidepressants for at least 6 months prior to their suicide

attempt. This evidence was provided by interviews with the patients and contacts

with their general practitioner. Systematic toxicological screening for

butyrophenones, phenothiazines, tricyclic antidepressants, benzodiazepines and

barbiturates was performed in the DSP patients. No screening was done for atypical

neuroleptics or for the newer antidepressants. As the use of psychotropics could not

be ruled out with certainty, and as some studies [7,39], though not all [40,41], have

shown that chronic treatment with antidepressant drugs may cause down-regulation

of the 5-HT2a postsynaptic receptors, our results could be due to chronic use of

tricyclic antidepressants. However, D'Haenen et al. reported no significant difference

in uptake values with 123I-ketanserin between patients who had not received any

antidepressant drug for at least 3 weeks and those who had taken antidepressants up

to 7 days before imaging [25]. The use of selective serotonin re-uptake inhibitors

(SSRIs) has been reported to increase the number [37,42,43] and the binding index

[13] of 5-HT2a receptors. Thus, it is unlikely that the reduction in binding index could

be attributed to the use of SSRIs. Moreover, the non-uniform reduction of the

binding index in different cortical regions is another indication that the reduction is

111

not likely to be attributable to medication.

Although a causal relation between a reduction in 5-HT2a binding index and

suicidality has never been demonstrated, the fact that SSRIs increase the number of

5-HT2a receptors suggests that the results of this study could explain the

pathophysiological basis of the findings of a controlled clinical trial of paroxetine

that showed a reduction in the repetition rate of deliberate self-harm in non-

depressed patients [44].

Conclusion

Our study provides evidence of a decreased receptor binding index in drug-free

deliberate self-harm patients. This decrease was most marked in patients who

attempted suicide by DSI. Further cross-sectional studies should replicate these

findings in larger populations. Longitudinal pharmacological intervention studies

should evaluate the 5-HT2a binding index in relation to clinical parameters of

suicidality, and effects of treatment on repetition of suicidal behaviour.

112

References

1. Nordstrom P, Asberg M. Suicide risk and serotonin. Int Clin Psychopharmacol 1992; 6 (Suppl 6):12-21.

2. Mann JJ, Malone KM, Sweeney JA, Brown RP, Linnoila M, Stanley B, Stanley M. Attempted suicide characteristics and cerebrospinal fluid amine metabolites in depressed patients. Neuropsychopharmacology 1996; 15:576-586.

3. Coccaro EF, Siever LJ, Klar HM, Maurer G, Cochrane K, Cooper TB, Mohs RC, Davis KL. Serotonergic studies in patients with affective and personality disorders: correlates with suicidal and impulsive aggressive behavior. Arch Gen Psychiatry 1989; 46:587-599.

4. Biegon A, Weizman A, Karp L, Ram A, Tiano S, Wolff M. Serotonin 5-HT2 receptor binding on blood platelets: a peripheral marker for depression? Life Sci 1987; 41:2485-2492.

5. Arora RC, Meltzer HY. Increased serotonin2 (5-HT2) receptor binding as measured by 3H-lysergic acid diethylamide (3H-LSD) in the blood platelets of depressed patients. Life Sci 1989; 44:725-734.

6. Mann JJ, Stanley M, McBride PA, McEwen BS. Increased serotonin 2 and beta-adrenergic receptor binding in the frontal cortex of suicide victims. Arch Gen Psychiatry 1986; 43:954-959.

7. Lowther S, De Parmentier F, Crompton MR, Katona CL, Horton RW. Brain 5HT2 receptors in suidice victims: violence of death, depression and effects of antidepressant treatment. Brain Res 1994; 642:281-289.

8. Gross-Isseroff R, Salama D, Israeli M, Biegon A. Autoradiographic analysis of age-dependent changes in serotonin 5HT2 receptor of the human brain postmortem. Brain Res 1990; 507:223-227.

9. Gjerris A. Do concentrations of neurotransmitters in lumbar CSF reflect cerebral dysfunction in depression? Acta Psychiatr Scand 1988; 345:21-24.

10. Palmer AM, Lowe SL, Francis PT, Bowen DM. Are post-mortem biochemical studies of human brain worthwhile? Biochem Soc Trans 1988; 16:472-475.

11. Stockmeier CA. Neurobiology of serotonin in depression and suicide. In: Mann JJ, Stoff DM, eds. The psychobiology of suicide. New York: The New York Academy of Sciences, 1997.

12. Biver F, Wikler D, Lotstra F, Damhaut P, Goldman S, Mendlewicz J. Serotonin 5-HT2 receptor imaging in major depression: focal changes in orbito-insular cortex. Br J Psychiatry 1997; 171:444-448.

13. Massou JM, Trichard C, Attar-Levy D, Feline A, Corruble E, Beaufils B, Martinot JL. Frontal 5-HT2A receptors studied in depressive patients during chronic treatment by selective serotonin reuptake inhibitors. Psychopharmacology (Berl) 1997; 133:99-101.

113

14. Attar-Lévy D, Martinot J, Blin J, Dao-Castellana M, Crouzel C, Mazoyer B, Poirier M, Bourdel M, Aymard N, Syrota A, Féline A. The cortical serotonin-2 receptors studied with positron-emission tomography and 18F-setoperone during depressive illness and antidepressant treatment with clomipramine. Biol Psychiatry 1999; 45:180-186.

15. Meyer JH, Kapur S, Houle S, DaSilva J, Owczarek B, Brown GM, Wilson AA, Kennedy S. Prefrontal cortex (5-HT2 receptors in depression: an [18F]setoperone PET imaging study. Am J Psychiatry 1999; 156:1029-1034.

16. Mertens J, Terriere D, Sipido V, Gommeren W, Janssen P, Leysen J. Radiosynthesis of a new radioiodinated ligand for serotonin-5HT2-receptors, a promising tracer for gamma-emission tomography. J Lab Compd Radiopharm 1994; 34:795-806.

17. Busatto G, Pilowsky L, Costa D, Mertens J, Terriere D, Ell P, Mulligan R, Travis M, Leysen J, Gacinovic S, Waddington W, Lingford-Hughes A, Kerwin R. Initial evaluation of 123I-5-I-R91150, a selective 5-HT2A ligand for single-photon emission tomography, in healthy subjects. Eur J Nucl Med 1997; 24:119-124.

18. Travis M, Busatto G, Pilowsky L, Mulligan R, Acton P, Gacinovic S, Mertens J, Terriere D, Costa D, Ell P, Kerwin R. 5-HT2a receptor blockade in patients with schizophrenia treated with risperidone and clozapine. A SPET study using the novel 5-HT2a ligand 123I-5-I-R-91150. Br J Psychiatry 1998; 173:236-241.

19. Hawton K, Catalan J. Attempted suicide: a practical guide to its nature and management. Oxford: Oxford Medical Publications, 1987.

20. Folstein MF, Folstein SE, McHugh PR. "Mini-Mental State": a practical method for grading the cognitive state of the patients for the clinician. J Psychiatr Res 1975; 12:189-198.

21. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960; 23:56-62.

22. Abi-Dargham A, Zea Ponce Y, Terriere D, Al Tikriti M, Baldwin R, Hoffer P, Charney D, Leysen J, Laruelle M, Mertens J, Innis R. Preclinical evaluation of [I-123] R93274 as a SPECT radiotracer for imaging 5-HT2a receptors. Eur J Pharmacol 1997; 321:285-293.

23. Pazos A, Probst A, Palacios J. Serotonin receptors in the human brain-IV. Autoradiographic mapping of serotonin-2 receptors. Neuroscience 1987; 21:123-139.

24. Baeken C, D'Haenen H, Flamen P, Mertens J, Terriere D, Chavatte K, Boumon R, Bossuyt A. 123I-5-I-R91150, a new single-photon emission tomography ligand for 5-HT2a receptors: influence of age and gender in healthy subjects. Eur J Nucl Med 1998; 25:1617-1622.

25. D'Haenen H, Bossuyt A, Mertens J, Bossuyt-Piron C, Gijsemans M, Kaufman L. SPECT imaging of serotonin 2 receptors in depression. Psychiatry Res Neuroimaging 1992; 45:227-237.

26. Arora RC, Meltzer HY. Serotonergic measures in the brains of suicide victims: 5-HT2 binding sites in the frontal cortex of suicide victims and control subjects. Am J Psychiatry 1989; 146:730-736.

114

27. Hrdina PD, Demeter TB, Vu P, Sotonyi P, Palkovits M. 5HT uptake sites and 5HT2 receptors in brain of antidepressant-free suicide victims/depressives: increases in 5-HT2 sites in cortex and amygdala. Brain Res 1993; 614:37-44.

28. Laruelle M, Abi-Dargham A, Casanova MF, Toti R, Weinberger DR, Kleinman JE. Selective abnormalities of prefrontal serotonergic receptors in schizophrenia. Arch Gen Psychiatry 1993; 50:810-818.

29. Owen F, Cross AJ, Crow TJ, Deakin JF, Ferrier IN, Lofthouse R, Poulter M. Brain 5HT2 receptors and suicide. Lancet 1983; 2:1256.

30. Owen F, Chambers DR, Cooper SJ, Crow TJ, Johnson JA, Lofthouse R, Poulter M. Serotonergic mechanisms in brains of suicide victims. Brain Res 1986; 362:185-188.

31. Stockmeier CA, Dilley GE, Shapiro LA, Overholser JC, Thompson PA, Meltzer HY. Serotonin 1A and 2A receptors in prefrontal cortex and hippocampus of suicide victims with major depression. Neuropsychopharmacology 1996; 16:162-173.

32. Cheetham SC, Crompton MR, Katona CL, Horton RW. Brain 5-HT2 receptor binding sites in depressed suicide victims. Brain Res 1988; 443:272-280.

33. Arranz B, Eriksson A, Mellerup E, Plenge P, Marcusson J. Brain 5HT1A, 5HT1D, and 5HT2 receptors in suicide victims. Biol Psychiatry 1994; 35:457-463.

34. Rosel P, Arranz B, Vallejo J, Oros M, Crespo JM, Menchon JM, Navarro MA. Variations in [3H]imipramine and 5HT2A but not [3H]paroxetine binding sites in suicide brains. Psychiatry Res Neuroimaging 1998; 82:161-170.

35. Pandey GN, Pandey SC, Janicak PG, Marks RC, Davis JM. Platelet serotonin-2 receptor binding sites in depression and suicide. Biol Psychiatry 1990; 28:215-222.

36. Cowen P, Charig E, Fraser S, Elliott J. Platelet 5HT receptor binding during depressive illness and tricyclic anti-depressant treatment. J Affect Disord 1987; 13:45-50.

37. Green AR, Johnson P, Mountford JA, Nimgaonkar VL. Some anticonvulsant drugs alter monoamine medicated behaviour in mice in ways similar to electroconvulsive shock: implications for antidepressant therapy. Br J Pharmacol 1985; 84:337-346.

38. Leysen JE, Van Gompel P, de Chaffoy de Courcelles D, Niemegeers CJ. Opposite regulation of serotonin-S2 and dopamine-D2 receptors in rat brain following chronic receptor blockade. J Recept Res 1987; 7:223-229.

39. Peroutka SJ, Snyder SH. Long-term antidepressant treatment decreases spiroperidol-labelled serotonin receptor binding. Science 1980; 210:88-90.

40. Mann E, Enna SJ. Neurochemical and behavioral correlates of antidepressant drug action. Life Sci 1982; 30:1653-1661.

41. Schoups AA, Dillen L, Claeys M, et al. Characterization of serotonin receptors and lack of effect of antidepressant therapy on monoamine functions in various regions of the rabbit brain. Eur J Clin Pharmacol 1986; 126:259-271.

42. Wamsley JK, Byerley WF, McCabe RT, et al. Receptor alterations associated with serotonergic agents - an autoradiographic analysis. J Clin Psychiatry 1987; 48 (Suppl 3):19-25.

115

43. Wagner HR, Reches A, Yablonskaya E, Fahn S. Clonazepam-induced up-regulation of serotonin1 and serotonin2 binding sites in rat frontal cortex. In: Fahn S, Marsden D, Van Woert M, eds. Myoclonus. New York: Raven Press; 1986:645-651.

44. Verkes R, Van der Mast R, Hengeveld M, Tuyl J, Zwinderman A, Van Kempen G. Reduction by paroxetine of suicidal behavior in patients with repeated suicide attempts but no major depression. Am J Psychiatry 1998; 155:543-547.

45. Biver F, Lotstra F, Monclus M, Wikler D, Damhaut P, Mendlewicz J, Goldman S. Sex difference in 5HT2 receptor in living human brain. Neuroscience-Lett 1996; 204: 25-28.

117

CHAPTER 5:

PREFRONTAL 5-HT2A RECEPTOR BINDING INDEX, HOPELESSNESS

AND PERSONALITY CHARACTERISTICS IN ATTEMPTED SUICIDE.

Kees van Heeringen, Kurt Audenaert, Koen Van Laere, Filip Dumont, Guido

Slegers, John Mertens, Rudi Dierckx

(Journal of Affective Disorders, Accepted pending corrections)

Summary

Depression, hopelessness, impaired problem solving capacities and deficient

serotonergic functions have been identified as major causes of suicidal behaviour. In

general, the relation between biological markers of attempted suicide and

psychological functions has been investigated by using indirect peripheral markers

of e.g. the serotonergic system. Recently, functional neuroimaging techniques with

radioligands allow direct in vivo assessment of the neurobiological status of the

central nervous system.

We studied the binding index of serotonin-2a (5-HT2a) receptors in the frontal cortex

of attempted suicide patients (n= 9) and normal controls (n= 13) using 123I-5-I-

R91150, a highly selective 5-HT2a receptor ligand. Moreover, we measured

personality characteristics (using Cloninger’s Temperament and Character Inventory)

and levels of hopelessness (using Beck’s Hopelessness Scale), and studied the

association between 5-HT2a receptor binding index, hopelessness and these

personality dimensions.

When compared to normal controls, attempted suicide patients had a significantly

lower binding potential of frontal 5-HT2a receptors, a higher level of hopelessness, a

higher score on the temperament dimension harm avoidance and lower scores on the

118

character dimensions self-directedness and cooperativeness. A significant correlation

was found between harm avoidance, hopelessness and binding index in the deliberate

self harm population.

The number of patients and ingestion of psychotropic drugs may influence the results

of the study.

Reduced central serotonergic function, hopelessness and harm avoidance are

interrelated phenomena which may determine the occurrence of attempted suicide.

Introduction

Major advances have been made in the study of suicidal behaviour in the past

decades, especially in psychiatric, psychological and biological domains of research.

In the psychiatric domain, a major contribution to our knowledge has been made by

psychological autopsy studies of individuals who committed suicide. More than 2000

suicides have been studied by means of this method, and, although the retrospective

nature of this approach limits their validity, the consistency of the results of these

studies in children, adolescents and adults is remarkable. These studies have shown

that completed suicide is particularly associated with major depression,

schizophrenia, alcohol dependence and borderline personality disorder [1]. However,

longitudinal studies have shown an increased risk of suicide in all psychiatric

disorders [2]. Taken together, these findings indicate that an increased risk of suicide

may occur across the boundaries of specific diagnostic categories. Further research

has shown that trait-like characteristics including impulsivity and a disturbed

regulation of anxiety or aggression may represent risk factors for suicidal behaviour

that occur across diagnostic boundaries [3,4].

Research in the psychological domain has shown that suicidal behaviour is

associated with hopelessness [5] and deficient problem solving strategies [6].

Moreover, evidence is accumulating that suicidal behaviour may be associated with

certain personality characteristics [7]. Recently we have demonstrated that the

occurrence of attempted suicide might be related to the personality characteristic

“reward dependence” [8], a personality dimension supposed to reflect biases in the

119

sensitivity in social communication [9].

In the biological domain the emphasis has been on the involvement of the

serotonergic system in the development of suicidal behaviour. Although it has

become clear that disturbances in the metabolism or functioning of many

neurotransmitters may play a role in the pathophysiology of suicidal behaviour,

including the noradrenergic system [8], the involvement of the serotonergic system

constitutes one of the most replicated findings in biological psychiatry. Since the

early work of Asberg and colleagues in the seventies [10], the serotonergic system

has been studied intensively. The involvement of serotonin (5-HT), and specifically

of the serotonin-2 receptor (5-HT2), in the pathophysiology of suicide and suicidality

has been the target of considerable research. Both indirect measurements, comprising

levels of 5-HT and its metabolite 5-HIAA in cerebrospinal fluid and in blood [11,12],

challenge tests for the serotonergic system [13], quantification of 5-HT2 receptors on

blood platelets [14,15], and direct measurements of brain serotonin function and

receptor status in post-mortem research [16,17,18] have yielded varying results. The

interpretation of these results is limited due tot the questionable validity of peripheral

measures as markers of cerebral activity [19]. Post-mortem cerebral study results are

also limited by lack of sampling from multiple regions [18] and rapid alterations in

neurotransmitter concentration postmortem [20]. Other reasons for the variability in

the study results are the use of divergent classification systems for psychiatric

diagnoses, heterogeneity with regard to nature of suicidal behaviour, and treatment

with antidepressant drugs or neuroleptics at the time of the suicide attempt [21].

Thus it has become clear that psychological and personality-related characteristics

and biological factors including functional disorders in the serotonin

neurotransmission system are involved in the development of suicidal behaviour.

These psychological and biological factors can be considered as trait-like

characteristics underlying the vulnerability for showing suicidal behaviour when

confronted with particular stressors [22]. Only a few studies have addressed the

potential relation between these psychological and biological characteristics, mainly

showing an association between impulsivity and serotonergic dysfunction. A major

limitation in the interpretation of the results of the scarce studies on the effect of the

association between psychological and biological characteristics on the occurrence of

120

suicidal behaviour is that the applied biological markers are peripheral in nature and

hence, assess neurotransmitter functioning in the central nervous system in only an

indirect way.

Recently we have in vivo demonstrated the involvement of the brain serotonergic

system in suicidal behaviour by means of a SPET 5-HT2a receptor study using a

highly selective ligand [23]. When compared to normal controls, patients with a

recent history of suicide attempts showed a significant decrease in 5-HT2a binding

index in the prefrontal cortex, the decrease being significantly more marked among

patients who used violent methods to attempt suicide than among those who

attempted suicide by means of self-poisoning.

This paper reports on a study of the association between 5-HT2a functioning in the

prefrontal cortex as assessed by means of a SPET receptor ligand procedure,

psychological characteristics including hopelessness, and personality dimensions.

Methods and Materials

Subjects

Twenty-one individuals participated in the study including 9 patients and 13

volunteers as normal controls. The patient group consisted of individuals who were

admitted to the Dept. of Psychiatry of the University Hospital Gent following a

suicide attempt. Attempted suicide was defined as the deliberate ingestion of more

than the prescribed amount of medical substances, or the ingestion of substances

never intended for human consumption, or any deliberate self-inflicted injury,

irrespective of the apparent purpose of the act [24]. Exclusion criteria were: age

younger than 18 years, electroconvulsive treatment in the preceding year, treatment

with antidepressants, neuroleptics, lithium or carbamazepine in the preceding six

months, major medical or neurological disorder, substance abuse disorder, pregnancy

or lactation period, and Mini-Mental State Examination [25] score less than 28.

The patient group consisted of seven males and two females. All patients were free

of psychotropic drugs for at least six months before the study, except for

121

benzodiazepines (n=4), barbiturates (n=1) or antiepileptics (n=1) which were

ingested at the occasion of the suicide attempt.

Patients were screened for alcohol intoxication and signs of chronic alcohol abuse by

means of blood alcohol level on admission, mean erythrocyte corpuscular volume

(MCV) and gamma-glutamyl transferase (γ-GT). Two patients were positive on

alcohol screening (1,1 and 1,2 gram promille) and had normal MCV and y-GT. All

other patients had normal MCV and y-GT .

The control group consisted of seven males and six females who were recruited

among the hospital staff and screened for personal and familial psychiatric and

medical history. None used psychotropic or other relevant medication or abused

illegal drugs. Physical examination was normal.

This study was approved by the local Ethical Committee. All participants gave

written informed consent.

Assessment of psychiatric and psychological characteristics

The patients were assigned a psychiatric diagnosis according to DSM-IV criteria by

the first two authors who are senior psychiatrists and were assessed with the

Hamilton Depression Rating Scale [26]. Psychological characteristics of patients and

healthy volunteers were assessed with Beck’s Hopelessness Scale [27] and the

Temperament and Character Inventory (TCI) in order to assess personality

dimensions, using the Dutch version of Cloninger’s 240-item (TCI) [28,29]. The

reliability, intercorrelations and factor structure of the translation are comparable to

the US version in Dutch [30]and Flemish [31] general population samples.

Assessment of serotonin-2a receptor binding index in frontal cortex

Binding index of 5-HT2a receptors was assessed via SPECT scanning using the

highly selective 5-HT2a antagonist 123I-5-I-R91150. The tracer was synthetised by

electrophilic substitution on the 5-position of the methoxybenzamide group of

R91150, followed by purification with high-performance liquid chromatography.

The product had a radiochemical purity of more than 99 % and was negative for

122

bacteria and pyrogen tests. A specific activity of 10 Ci/µmol was obtained. The

tracer is a 5-HT2a antagonist with high affinity (Kd = 0.11 nM) and selectivity for 5-

HT2a receptors. The selectivity of the ligand for 5-HT2a receptors with regard to


dopamine receptors, adrenoreceptors and histamine receptors is at least a factor of

50. The tracer was displaceable with ketanserin [32, 33].

Before SPECT scanning, thyroid blockade was achieved by administration of a

single oral dose of 100 mg potassium iodide prior to injection. All subjects received

an intravenous injection of 185 MBq 123I-5-I-R91150 in normal sitting conditions.

SPECT scanning was performed using a triple-headed high-sensitivity high-

resolution Toshiba gamma camera (GCA-9300) with fan-beam collimation. For 123I,

the resulting transaxial image resolution is 9.5 mm full-width at half-maximum

(FWHM). Since sequential dynamic SPECT brain scans have shown that the

cortico-cerebellar ratio reaches a plateau between 90 and 110 minutes reflecting

pseudo-equilibrium and remains stable thereafter for up to 8 hours [34], acquisition

was started between 110 and 140 minutes after tracer injection. A time-specific-

activity curve of a representative healthy volunteer (male, age 25) is published

elsewhere [23]. A transmission scan (TCT-scan) was acquired before the emission

scan, using three 153-Gd rod sources. This scan was used for subsequent image co-

registration. Emission images were acquired during 40 minutes. The whole brain

volume was acquired within the single scanning session.

Analysis of the scans was performed blind to patient status. Images were corrected

for scatter and attenuation. After automatic image co-registration to stereotactic

space (BRASS, Nuclear Diagnostics), a predefined volume-of-interest (VOI)

analysis was performed with 12 cortical regions. Radioactivity estimates in the

cortex were assumed to represent total ligand binding (specific + non-specific

binding + free ligand) [35]. Since the cerebellum is void of serotonin receptors [36]

and therefore represents non-specific activity, calculation of relative indices of

specific binding index (BP) was done by VOI normalisation to the activity per

volume element in the cerebellum. Under these pseudo-equilibrium circumstances,

this binding index is directly related to the in vivo receptor density (Bmax) and affinity

(KD). Binding Index was defined as (target activity – background activity in brain) /

123

(background activity) which was operationally estimated as (counts/pixel in frontal

cortex) / (counts/pixel in cerebellum).

Statistical methods

The equality of age-distributions between diagnostic categories was evaluated

according to the Kruskal-Wallis test. As binding index levels were normally

distributed, analysis of variance was used to compare mean levels between

categories. Linear adjustment for age was done according to analysis of covariance.

An a priori level of α=0.05 was chosen to indicate statistical significance. Model

assumptions were checked by graphical inspection of Pearson residuals. None of the

second order interaction terms between diagnostic group, age and gender turned out

to be significant at the 0.10 level.

Results

The mean age of the participants was 30.4 years (range: 19 to 47; SD: 9.2), being

32.4 (SD: 11.4) years among attempted suicide patients and 28.9 (SD: 7.6) years

among controls. The difference in mean ages between patients and controls was not

significant ( MWU 49.5; P= 0.56). Although statistically not significant, mean ages

were slightly different, and since previous research has suggested that binding index

levels decrease with age [37], this was taken into account as a potential confounding

variable in our analyses. Therefore age correction was performed applying linear

correction. All values were corrected to the mean age of the group, i.e. 30.4 years.

Since the study groups were not significantly different regarding distribution of

gender (Fisher Exact’s Test P=0.25), and no effect of gender on 5-HT2a binding

index with this tracer has been demonstrated in our database of healthy volunteers

and by others has been found [24], no adjustments for gender were made in further

analyses. However, one study with another 5-HT2 receptor ligand showed a higher

5-HT2 binding capacity in men [54].

124

With regard to psychiatric disorders the clinical diagnoses that were assigned are

listed in Table 1 Methods of attempted suicide included deliberate self-poisoning (n

= 6) and deliberate self injury (n = 3).

sex age Clinical Disorder

Personality disorder

HAMD

(4)

Type (1) Means (2) Time (3)


F 19 Adjustment Disorder Borderline 10 DSP intoxication barbiturates

2

M 29 Major Depressive Disorder - 29 DSP intoxication phenytoin

6

M 34 Adjustment Disorder Narcissistic 11 DSP self-injection insulin

1

F 48 Major Depressive Disorder - 27 DSI wrist-cutting 7

M 19 Adjustment Disorder Narcissistic 8 DSI wrist-cutting 3

M 19 Brief Psychotic Disorder - 7 DSP intoxication BDZ 3


M 43 Adjustment Disorder Dependent 8 DSI Strangulation 1

(1) DSP: Deliberate Self Poisoning; Deliberate Self Injury

(2) BDZ: benzodiazepines

(3) Time between suicide attempt and image acquisition (days)

(4) Hamilton Depression Rating Scale

Table 1: Demographical characteristics, clinical diagnosis, description of suicide attempt, and scan acquisition results of patients with deliberate self-harm and normal volunteers. Frontal binding index is given in relative to cerebellar activity. Time = interval between suicide attempt and scan acquisition.

Table 2 shows the means and standard deviations for binding potential, scores on

personality dimensions and levels of hopelessness for attempted suicide patients and

normal controls. Attempted suicide patients showed a significantly lower prefrontal

cortex binding potential of the 5-HT2a receptor ligand and a significantly higher score

on the personality dimensions harm avoidance and self-transcendence. Patients

scored significantly lower on the character dimensions self-directedness and

cooperativeness.

125

Attempted suicide Healthy

Patients (n=9) volunteers (n=13) t-statistics

mean (SD) mean (SD)

Binding potential 140.7 (22.2) 168.0 (13.6) 3.59 **

Novelty Seeking 21.3 (9.1) 19.5 (4.9) - 0.53

Harm Avoidance 22.6 (4.4) 11.0 (3.7) - 6.38 **

Reward Dependence 14.1 (3.6) 16.9 (3.5) 1.69

Persistence 3.8 (1.8) 4.6 (1.9) - 1.18

Self-Directedness 18.1 (7.3) 32.6 (9.2) 3.93 **

Cooperativeness 28.8 (5.8) 34.5 (2.4) 3.19 **

Self-Transcendence 14.0 (8.4) 7.9 (3.7) - 2.32 *

Hopelessness 10.3 (3.1) 2.3 (1.3) - 3.58 **

Table 2

Comparison between attempted suicide patients and normal controls for prefrontal 5-HT2a receptor binding index, hopelessness and personality characteristics

In order to study any potential associations between prefrontal serotonin functioning

(as measured by means of the binding potential) and assessed psychological

characteristics among DSH patients, Spearman’s r correlation coefficients were

calculated (Table 3). Thus a significant negative correlation was found between

binding potential and harm avoidance scores and a significant positive correlation

between binding potential and scores on self-directedness and cooperativeness. A

strongly significant positive correlation could be shown between hopelessness and

harm avoidance while hopelessness correlated significantly and negatively with

cooperativeness and self-transcendence.

126

Binding potential Hopelessness

TCI scores Novelty Seeking - 0.07 - 0.54 Harm Avoidance - 0.72 * 0.72 *

Reward Dependence 0.03 - 0.01 Persistence - 0.38 0.17

Self-Directedness - 0.04 0.29 Cooperativeness 0.33 - 0.10

Self-Transcendence - 0.13 - 0.71 * Hopelessness - 0.70 *

* p < 0.05

Table 3 Correlations between prefrontal 5-HT22a receptor binding index, hopelessness and personality characteristics (Spearman’s r coefficients)

Discussion

The main findings from this study on biological and psychological risk factors for

suicidal behaviour and their potential association can be summarized as follows.

First, in addition to the previously reported decreased binding potential of prefrontal

5-HT2a receptors when compared to normal controls [38], attempted suicide patients

show comparatively higher levels of hopelessness, higher scores on the personality

dimensions harm avoidance and self-transcendence, and lower scores on the

dimensions self-directedness and cooperativeness. Secondly, the binding potential of

prefrontal 5-HT2a receptors is negatively correlated with harm avoidance and

hopelessness scores and positively correlated with scores on the personality

dimensions self-directedness and cooperativeness. Thirdly, levels of hopelessness are

positively correlated with the score on harm avoidance and correlates negatively with

scores on cooperativeness and self-transcendence.

127

Before discussing these findings in the context of the treatment and prevention of

suicidal behaviour some methodological shortcomings should be addressed. These

shortcomings concern, first, the small size (potentially leading to a type II error) and

the composition of the patient sample. When compared to attempted suicide patients

in general there is a relative preponderance of males and patients using violent

methods to attempt suicide in the current study, which may be responsible for a

relative low mean binding potential of 5-HT2a receptors in this patient group. Since

previous reports showed no effect of gender on the 5-HT2a binding potential [37], no

adjustment for gender was made in further analyses. An age-dependent decline in

frontal 5-HT2a receptor binding potential has been reported [37]. However, the mean

age of our study group was slightly lower than that of the normal volunteers so that

age cannot be expected to account for the decreased binding potential.

Secondly, the results can be biased by the fact that a physical trauma, due to the

recent suicide attempt, may have had an impact on the binding index of the tracer

due to a decrease in cerebral blood flow. We cannot rule out this possibility, but the

results were fairly homogeneous, irrespective of the method used to attempt suicide.

Moreover, the results can be due to a slower metabolic rate for the tracer among the

volunteers than among the DSH patients. However, no indications for differences in

metabolic rate can be found in the literature.

In interpreting the data, attention should also be paid to a potential effect of the

medication and alcohol that were ingested prior to or at the occasion of the suicide

attempt. Indeed, the consumption of large amounts of alcohol may have an impact on

the clearance of the ligand, possibly through the induction of liver enzymes.

Moreover, patients that are entrapped in a crisis often seek relief in alcohol. In our

study sample, two patients were positive on alcohol screening (<0.8 pro mille) when

admitted to the hospital. Blood tests on Mean Erythrocyte Corpuscular Volume and

gamma-GT in all patients were normal. Patients did not use any alcohol during their

stay in the hospital in the period between the admission and the scanning. None of

the patients had a history of alcohol abuse. Thus the results of this study are not

likely to be confounded by the abuse of large amounts of alcohol and its impact on

clearance of the ligand is unlikely to confound the results.

128

Concerning psychotropic medication, it must be noted that some of the patients

attempted suicide ingesting psychotropics, and hence, the possible effects of

benzodiazepines (BDZ), barbiturates and anti-epileptics on the results must be

considered. Benzodiazepines were reported to increase 5-HT2 receptor numbers in

animal studies [39], but a lack of effect on 5-HT2 receptor numbers has also been

reported [40]. An effect of benzodiazepine withdrawal could not been ruled out,

although patients showed no clinical withdrawal symptoms in the period following

the intoxication. In this study and in a 18F-setoperone study in depressed patients [41]

patients who received BDZ and patients who were free of BDZ fell in the same range

of 5-HT2a binding index. Hence, it is unlikely that the decrease in binding index as

observed in our study was induced by the ingestion of benzodiazepines. There are no

reports of effects of barbiturates or phenytoin on 5-HT2a receptor-binding

characteristics.

Patients were enrolled in our study only when there was evidence that they had not

used neuroleptics or anti-depressants at least six months prior to their suicide

attempt. This evidence was provided by interviews with the patients, and contacts

with their general practitioner. Toxicological screening for butyrophenones,

phenothiazines, tricyclic antidepressants, benzodiazepines and barbiturates was

performed systematically in the DSP patients. They were not screened

toxicologically for atypical neuroleptics and for the newer anti-depressants. Thus, as

the use of psychotropics could not be ruled out with certainty, and some studies [18,

42], but not all [43,44] have shown that chronic treatment with anti-depressant drugs

may possibly cause down-regulation of the 5-HT2a postsynaptic receptors, our

results could be due to chronic use of tricyclic antidepressants. However, D’Haenen

et al. reported no significant difference in uptake values with 123I-ketanserin between

patients who did not receive any antidepressant drug for at least 3 weeks and those

who had taken antidepressants up to 7 days before imaging [45]. Most studies, but

not all [55], reported that selective serotonin reuptake inhibitors increase the number

[39,46,47] and the binding index [48] of 5-HT2a receptors. Thus, it is unlikely that

the reduction in binding index could be attributed to the use of SSRI’s. Moreover,

the non-uniform reduction of the binding index in different cortical regions is another

indication that the reduction probably cannot be attributed to the use of medication.

129

Although this is the first report of an association between a reduction in 5-HT2a

binding index and suicidal behaviour and given the fact that SSRIs increase the

number of 5-HT2a receptors, the results of this study could explain the

pathophysiological basis of those of a controlled clinical paroxetine trial showing a

reduction in the repetition rate of deliberate self harm in non-depressed patients [49].

Taking these shortcomings into account, the question of the meaning of the current

findings with regard to the understanding, treatment and prevention of suicidal

behaviour arises. A potential approach to understanding these findings is by

considering the function of the 5-HT2a system in the prefrontal cortex. Deakin (1996)

has suggested that 5-HT2a projections from the brainstem dorsal raphe nucleus, in

conjunction with the dopaminergic system, modulate expressive functions of the

prefrontal cortex thus facilitating approach or avoidance behaviour when confronted

with adverse stimuli. The anticipation of future events is biased by disturbed 5-HT2a

modulation, giving rise to distorted cognitions about the future and thus to

hopelessness. This motor-expressive system constitutes the basolateral circuit (or

'social brain') together with a sensory-receptive system which is thought to be

modulated by 5-HT1 projections from the median raphe nucleus to the temporal

lobes, in conjunction with the noradrenergic system [50]. This sensory-receptive

system is thought to modulate the sensory-perceptual processing of (social) stimuli,

and thus to define the resilience, adaptation or tolerance to stress. Evidence is

accumulating that this resilience system may break down in case of psychosocial

adversity due to raised cortisol concentrations thus resulting in depressive ideation,

feelings of subordinance and low self-esteem.

With regard to the motor-expressive part of the basolateral circuit our current

findings in attempted suicide patients indeed show a cluster of phenomena including

increased avoidance behaviour, hopelessness and disturbed 5-HT2a functioning. We

have recently studied aspects of the sensory-receptive system in attempted suicide

patients, and found suicidal behaviour to be associated with increased cortisol

secretion (as measured by means of 24-hour urinary cortisol secretion), low levels of

plasma MHPG (a metabolite of noradrenalin) and low scores on the personality

dimension reward dependence [8]. Reward dependence scores thereby correlated

negatively with cortisol levels. Functional characteristics of the 5-HT1 system were

130

not assessed directly in this study but both an association between reward

dependence and 5-HT1 functioning [9] and a role of the noradrenergic system in the

modulation of the sensory-receptive part of the basolateral circuit [50] have been

suggested.

Taken together, these findings suggest the following neurobiological model to

understand suicidal behaviour. At the sensory-perceptual level, a reduced resilience

to psychosocial stressors is constituted by low reward dependence and reduced

noradrenergic functioning. The confrontation with particular psychosocial stressors

consequently results in increased cortisol secretion (which may reduce 5-HT1

functioning and thus lead to), subordinance, depressive ideation and low self-esteem.

As reward dependence scores reflect biases in the decoding of social signals,

stressors are particularly of a social nature in that they threaten the (perceived)

integration in the social system in which individuals live. The crucial role of such

events, leading to feelings of subordinance and low self-esteem, in the development

of suicidal behaviour has independently been recognized by cognitive psychologists

[5].

At the motor-expressive level, the occurrence of suicidal behaviour is associated with

a cluster of interdependent phenomena including increased harm avoidance, reduced

5-HT2a functioning in the prefrontal cortex and hopelessness.

These findings need to be replicated and a number of elements of this model need to

be studied in more detail. For instance, the precise role of the 5-HT2a system in the

prefrontal cortex is not yet clear. Post-mortem studies have previously shown an

increased binding potential of 5-HT2 receptors in the prefrontal cortex in suicide

victims [16] which may, however, be due to methodological shortcomings of the

post-mortem study approach including the use of less specific radioligands. Our

current finding of a reduced 5-HT2a receptor binding potential may therefore indeed

reflect a hypofunction of the 5-HT2a system in association with suicidal behaviour.

However, Deakin (1996) has also suggested an excessive function of the 5-HT2a

system in the development of hopelessness. As the currently presented findings are

the first ever published in vivo measurements of serotonergic activity in the central

nervous system in patients showing suicidal behaviour, replication is clearly needed

[50].

131

The findings from the present study suggest a role of the prefrontal 5-HT2a system in

emotional (i.e. feelings of hopelessness) and behavioural (i.e. behavioural inhibition)

characteristics which may, however, be secondary to cognitive aspects. In view of

the demonstrated deficiencies in autobiographical memory in attempted suicide

patients resulting in reduced problem-solving skills [5], the localisation of working

memory in the prefrontal cortex [51], and the role of the serotonergic system in the

process of memory and learning [52], we suggest that the crucial dysfunction in the

motor-expressive part of the basolateral circuit in association with suicidal behaviour

consists of a serotonergically mediated disturbance in memory functions which

underlies a deficiency in the retrieval of appropriate problem-solving strategies.

Further study is needed to determine whether the decreased binding potential of

prefrontal 5-HT2a receptors in attempted suicide patients is trait- or state-dependent,

and can be influenced by serotonergic drugs as was shown in depressed patients [48].

With regard to the psychosocial and psychopharmacological treatment of attempted

suicide patients which has currently been shown to lack efficacy in preventing

recurrence [53], this could mean that treatment should be started with serotonergic

drugs followed by psychotherapy in a second stage, as these drugs may restore the

biological substrate for the learning of e.g. new problem-solving skills.

132

References

1. Hawton K, van Heeringen C. The International Handbook of Suicide and Attempted

Suicide. Chicester: John Wiley, 2000.

2. Harris E, Barraclough B. Excess mortality of mental disorder. Br J Psychiatry 1998; 173: 11-53.

3. Apter A, Van Praag H. Anxiety, impulsivity and depressed mood in relation to violent and suicidal behaviour. Acta Psychiatr Scand 1993; 87: 1-5.

4. Van Praag H. Serotonin-related, anxiet/aggression driven, stressor-precipitated depression. A psychobiological hypothesis. Eur Psychiatry 1996; 11: 57-67.

5. Williams J, Pollock L. The psychology of suicidal behaviour. In: van Heeringen C, eds. Understanding suicidal behaviour: the suicidal process approach to research and treatment. Chicester: John Wiley; 2000: 79-95.

6. Pollock L, Williams J. Problem solving and suicidal behaviour. Suicide and Life-Threatening Behaviour 1999; 28: 375-388.

7. Apter A, Ofek H. Personality constellations in suicidal behaviour. In: van Heeringen C, eds. Understanding suicidal behaviour: the Suicidal Process Approach to Research and Treatment. Chicester: John Wiley; 2001 (In press).

8. van Heeringen C, Audenaert K, Van de Wiele L, Verstraete A. Cortisol in violent suicidal behaviour: association with personality and monoaminergic activity. J Affect Dis 2000; 60: 181-189.

9. Cloninger C. The genetics and psychobiology of the seven-factor model of personality. In: Silk K, eds. Biology of Personality Disorders. Washington: American Psychiatry Press; 1998: 66-93.

10. Asberg M, Traskman L, Thoren P. 5-HIAA in the cerebrospinal fluid: a biochemical suicide predictor? Arch Gen Psychiatry 1976; 33: 1193-1197.

11. Mann JJ, Malone KM, Sweeney JA, Brown RP, Linnoila M, Stanley B, Stanley M. Attempted suicide characteristics and cerebrospinal fluid amine metabolites in depressed patients. Neuropsychopharmacology 1996; 15: 576-586.

12. Nordstrom P, Asberg M. Suicide risk and serotonin. Int Clin Psychopharmacol 1992; 6[Suppl 6]: 12-21.

13. Coccaro EF, Siever LJ, Klar HM, Maurer G, Cochrane K, Cooper TB, Mohs RC, Davis KL. Serotonergic studies in patients with affective and personality disorders: correlates with suicidal and impulsive aggressive behavior. Arch Gen Psychiatry 1989; 46: 587-599.

14. Arora RC, Meltzer HY. Increased serotonin2 (5-HT2) receptor binding as measured by 3H-lysergic acid diethylamide (3H-LSD) in the blood platelets of depressed patients. Life Sci 1989; 44: 725-734.

133

15. Biegon A, Weizman A, Karp L, Ram A, Tiano S, Wolff M. Serotonin 5-HT2 receptor binding on blood platelets: a peripheral marker for depression? Life Sci 1987; 41: 2485-2492.

16. Mann JJ, Stanley M, McBride PA, McEwen BS. Increased serotonin 2 and beta-adrenergic receptor binding in the frontal cortex of suicide victims. Arch Gen Psychiatry 1986; 43: 954-959.

17. Gross-Isseroff R, Salama D, Israeli M, Biegon A. Autoradiographic analysis of age-dependent changes in serotonin 5HT2 receptor of the human brain postmortem. Brain Res 1990; 507: 223-227.

18. Lowther S, De Parmentier F, Crompton MR, Katona CL, Horton RW. Brain 5HT2 receptors in suidice victims: violence of death, depression and effects of antidepressant treatment. Brain Res 1994; 642: 281-289.

19. Gjerris A. Do concentrations of neurotransmitters in lumbar CSF reflect cerebral dysfunction in depression? Acta Psychiatr Scand 1988; 345: 21-24.

20. Palmer AM, Lowe SL, Francis PT, Bowen DM. Are post-mortem biochemical studies of human brain worthwhile. Bioch Soc Trans 1988; 16: 472-475.

21. Stockmeier CA. Neurobiology of serotonin in depression and suicide. In: Mann JJ, Stoff DM, eds. The psychobiology of suicide. New York: The New York Academy of Sciences; 1997: 220-232.

22. Van Praag H. About the biological interface between psychotraumatic experiences and affective regulation. In: van Heeringen C, ed. Understanding suicidal behaviour: the suicidal process approach to research and treatment. Chicester: John Wiley; (in press)

23. Audenaert K, Van Laere K, Dumont F, Slegers G, Mertens J, van Heeringen C, Dierckx R. Decreased frontal serotonin 5-HT2a receptor binding index in deliberate self harm patients. Eur J Nucl Med 2000; 27: 1800-1808.

24. Hawton K, Catalan J Attempted suicide: a practical guide to its nature and management. Oxford: Oxford Medical Publications, 1987.

25. Folstein MF, Folstein SE, McHugh PR. 'Mini-mental State'. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12: 189-198.

26. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960; 23: 56-62.

27. Beck A, Weissmann A, Lester D, Trexler L. The measurement of pessimism: the Hopelessness Scale. J Cons Clin Psychol 1974; 42: 861-865.

28. Cloninger C, Svrakic D, Pryzbeck T. A psychobiological model of temperament and character. Arch Gen Psychiatry 1993; 30: 975-990.

29. Cloninger C, Pryzbeck T, Svrakic D The Temperament and Character Inventory (TCI): a Guide to its Development and Use. St Louis: Center for Psychobiology of Personality, 1994.

134

30. Duijssens I, Goekoop J, Spinhoven P, Eurelings-Bontekoe E. De Temperament en karakter vragenlijst. Ned Tijdschr Psychol 1997; 52: 199-202.

31. Meerschaert T, Audenaert K, van Heeringen C. Temperament en karakter in Vlaanderen: een onderzoek naar de distributie van persoonlijkheidsdimensies (Temperament and character in Flanders: a study of the distribution of personality dimensions), submitted to Tijdschr Psychiatrie

32. Abi-Dargham A, Zea Ponce Y, Terriere D, Al Tikriti M, Baldwin R, Hoffer P, Charney D, Leysen J, Laruelle M, Mertens J, Innis R. Preclinical evaluation of [I-123] R93274 as a SPECT radiotracer for imaging 5-HT2a receptors. Eur J Pharmacol 1997; 321: 285-293.

33. Mertens J, Terriere D, Sipido V, Van Gommeren W, Janssen P, Leysen J. Radiosynthesis of a new radio-iodinated ligand for serotonin-5HT(2)-receptors, a promising tracer for gamma-emission tomography. J Lab Comp Radiopharm 1994; 34: 795-806.

34. Busatto G, Pilowsky L, Costa D, Mertens J, Terriere D, Ell P, Mulligan R, Travis M, Leysen J, Gacinovic S, Waddington W, Lingford-Hughes A, Kerwin R. Initial evaluation of 123I-5-I-R91150, a selective 5-HT2A ligand for single-photon emission tomography, in healthy subjects. Eur J Nucl Med 1997; 24: 119-124.

35. Travis M, Busatto G, Pilowsky L, Mulligan R, Acton P, Gacinovic S, Mertens J, Terriere D, Costa D, Ell P, Kerwin R. 5-HT2a receptor blockade in patients with schizophrenia treated with risperidone and clozapine. A SPET study using the novel 5-HT2a ligand 123I-5-I-R-91150. Br J Psychiatry 1998; 173: 236-241.

36. Pazos A, Probst A, Palacios J. Serotonin receptors in the human brain-IV. Autoradiographic mapping of serotonin-2 receptors. Neuroscience 1987; 21: 123-139.

37. Baeken C, D'Haenen H, Flamen P, Mertens J, Terriere D, Chavatte K, Boumon R, Bossuyt A. 123I-5-I-R91150, a new single-photon emission tomography ligand for 5-HT2a receptors: influence of age and gender in healthy subjects. Eur J Nucl Med 1998; 25: 1617-1622.

38. Lahorte P, Vandenberghe S, Van Laere K, Audenaert K, Lemahieu I, Dierckx R. Assessing the performance of SPM analyses of SPECT. Neuroimage 2000; 12: 757-764.

39. Green AR, Johnson P, Mountford JA, Nimgaonkar VL. Some anticonvulsant drugs alter monoamine medicated behaviour in mece in ways similar to electroconvulsive showk: implications for antidepressant therapy. Br J Pharmacol 1985; 84: 337-346.

40. Leysen JE, Van Gompel P, de Chaffoy de Courcelles D, Niemegeers CJ. Opposite regulation of serotonin-S2 and dopamine-D2 receptors in rat brain following chronic receptor blockade. J Recept Res 1987; 7: 223-229.

41. Attar-Lévy D, Martinot J, Blin J, Dao-Castellana M, Crouzel C, Mazoyer B, Poirier M, Bourdel M, Aymard N, Syrota A, Féline A. The cortical serotonin-2 receptors studied with positron-emission tomography and 18F-setorperone during depressive illness and antidepressant treatment with clomipramine. Biol Psychiatry 1999; 45: 180-186.

135

42. Peroutka SJ, Snyder SH. Long-term antidepressant treatment decreases spiroperidol-labelled serotonin receptor binding. Science 1980; 210: 88-90.

43. Mann E, Enna SJ. Neurochemical and behavioral correlates of antidepressant drug action. Life Sci 1982; 30: 1653-1661.

44. Schoups AA, Dillen L, Claeys M, et al. Characterization of serotonin receptors and lack of effect of antidepressant therapy on monoamine functions in various regions of the rabbit brain. Eur J Clin Pharmacol 1986; 126: 259-271.

45. D'Haenen H, Bossuyt A, Mertens J, Bossuyt-Piron C, Gijsemans M, Kaufman L. SPECT imaging of serotonin 2 receptors in depression. Psychiatry Res Neuroimaging 1992; 45: 227-237.

46. Wagner HR, Reches A, Yablonskaya E, Fahn S. Clonazepam-induced up-regulation of serotonin1 and serotonin2 binding sites in rat frontal cortex. In: Fahn S, Marsden D, Van Woert M, eds. Myoclonus. New York: Raven Press; 1986: 645-651.

47. Wamsley JK, Byerley WF, McCabe RT, et al. Receptor alterations associated with serotonergic agents - An autoradiographic analysis. J Clin Psychiatry 1987; 48(suppl 3): 19-25.

48. Massou JM, Trichard C, Attar-Levy D, Feline A, Corruble E, Beaufils B, Martinot JL. Frontal 5-HT2A receptors studied in depressive patients during chronic treatment by selective serotonin reuptake inhibitors. Psychopharmacology (Berl) 1997; 133: 99-101.

49. Verkes R, Van der Mast R, Hengeveld M, Tuyl J, Zwinderman A, Van Kempen G. Reduction by paroxetine of suicidal behavior in patients with repeated suicide attempts but no major depression. American Journal of Psychiatry 1998; 155: 543-547.

50. Deakin J. 5-HT, antidepressant drugs and the psychosocial origins of depression. J Psychopharmacol 1996; 10: 31-38.

51. Fuster J The prefrontal cortex: anatomy, physiology and neuropsychology of the frontal lobe. Philadelphia: Lippincott-Raven, 1997.

52. Madden I Neurobiology of Learning, emotion and affect? New York: Raven Press, 1986.

53. Hawton K, Arensman E, Townsend E, Bremner S, Feldman E, Goldney R, Gunnell D, Hazell P, van Heeringen C, House A, Owens D, Sakinofsy I, Traskman-Bendz L. Deliberate self harm: systematic review of the efficacy of psychosocial and pharmacological interventions in preventing repetition. Br Med J 1998; 317: 441-447.

54. Biver F, Lotstra F, Monclus M, Wikler D, Damhaut P, Mendlewicz J, Goldman S. Sex difference in 5HT2 receptor in living human brain. Neuroscience-Lett 1996; 204: 25-28.

55. Meyer J, Kapur S, Eisfeld B, Brown G, Houle S, Da Silva J, Wilson A, Rafi Tari S, Mayberg H, Kennedy S. The effect of paroxetin on 5HT2a receptors in depression: a 18F-setoperone PET imaging study. Am J Psychiatry 2001; 158: 75-85.

137

CHAPTER 6:

DECREASED 5-HT2A RECEPTOR BINDING IN PATIENTS WITH

ANOREXIA NERVOSA.

Kurt Audenaert, Koen Van Laere, Filip Dumont, Myriam Vervaet, Ingeborg

Goethals, Guido Slegers, John Mertens, Cees van Heeringen, Rudi Dierckx

(European Journal of Nuclear Medicine, submitted)

Summary

Indirect estimations of brain neurotransmitters in patients with Anorexia Nervosa

(AN) and low weight demonstrated a reduction in brain serotonin (5-HT) turnover

in general and hypothesised a dysfunctional 5-HT2a receptor system in particular.

It was our aim to investigate the central 5-HT2a receptor binding index using

SPET brain imaging.

In this study the 5-HT2a receptor of low weight patients with AN was studied by

means of the highly specific radio-iodinated 5-HT2a receptor antagonist 4-amino-


methoxybenzamide or 123I-5-I-R91150. Fifteen patients with the clinical

diagnosis of an AN and 11 age-matched healthy controls received an intravenous

injection of 185 MBq 123I-5-I-R91150 and were scanned with high-resolution

brain Single Photon Emission Computed Tomography (SPECT).

Patients with eating disorders, compared to healthy volunteers, have a significantly

reduced 5-HT2a binding index in the left frontal cortex, in the left and right parietal

cortex and the left and right occipital cortex. There was a significant left-right

asymmetry in the frontal cortex (left<right).

These results are in accordance with diminished metabolism and perfusion of

138

frontal and parietal cortex in recent neuroimaging studies and imply localized

disturbed serotonergic function. The data are discussed in the light of possible

confounding factors that are related to the low weight AN status. A regional, and

not a global cortical reduction in 5-HT2a binding index, is not likely to be caused

by a general reduction in serotonergic function due to the possible confounding

factors. Directions about further research are given.

Introduction

The essential features of Anorexia Nervosa (AN) are that the individual refuses to

maintain a minimally normal body weight and is intensely afraid of gaining weight

[1]. Weight loss can be accomplished primarily through excessive physical exercise

and through voluntarily reduction in total food intake, sometimes accompanied by

purging behaviour, i.e. self-induced vomiting or the misuse of laxatives or diuretics.

Besides this, a disturbance in the perception of body shape and weight is an essential

neuropsychological feature of AN [1].

Psychological as well as biological mechanisms appear to play a key role in the

pathogenesis of Anorexia Nervosa. Neuropsychological investigations have

indicated cognitive deficits in frontal cortex and in parietal cortex [2,3]. Concerning

parietal dysfunctions, more specifically distortion of body image is described [4,5].

Functional brain imaging with PET and SPET has shown a reduced parietal and

frontal cortex metabolism [6,7] or cerebral perfusion [8] in patients with AN.

In past decades the involvement of serotonin as a neurotransmitter in Anorexia

Nervosa was largely indicated via indirect estimations of brain serotonergic function.

Impaired serotonin turnover and function was demonstrated through plasma

measurements of tryptophan, the precursor of 5-HT [9] or 5-hydroxyindole acetic

acid (5-HIAA), the metabolite of 5-HT, in cerebrospinal fluid (CSF) of low weight

AN patients [10]. Interestingly, long time weight restored AN patients have elevated

CSF 5-HIAA concentrations [10]. Another indication for a reduced serotonergic

function in low weight AN patients was demonstrated through blunted physiological

responses to the administration of selective pharmacological agonists (5-HTP, m-

139

CPP, fenfluramine) in low-weight AN patients [11,12,13,14,15]. Again, after weight

restoration, these findings normalized [13].

The involvement of the 5-HT2a receptor in the pathophysiology of AN was

demonstrated indirectly via blood platelet studies, both through an enhanced

mobilization of intracellular platelet calcium content, mediated via 5-HT2 receptors

[16] as through enhanced platelet serotonin 5-HT2a binding, measured in vitro with 3H-LSD [17]. Some recent molecular genetic studies in eating disorder patients [18,

19], but not all [20], have demonstrated an increased frequency of one of the alleles

on the promotor region of the 5-HT2a gene. Also twin and family studies suggest

that there may be a genetic vulnerability to AN [21], and the hypothesis has been put

forward that this vulnerability may be expressed in the central serotonergic system

[21, 22].

Functional imaging techniques, such as Positron Emission Tomography (PET) and

Single- Photon Emission Computed Tomography (SPECT), using specific 5-HT2

receptor ligands, make it possible to evaluate in vivo receptor binding in patients with

Anorexia Nervosa. Preliminary research in healthy subjects has indicated that 123I-5-

I-R91150 or 4-amino-N-[1-[3-(4-fluorophenoxy)propyl]-4-methyl-4-piperidinyl]-5-

iodo-2-methoxybenzamide is a suitable ligand for imaging 5-HT2a receptors in vivo.

It binds reversibly and with high-affinity in vitro to 5-HT2a receptors [23]. On

average 2% of a bolus dose of 123I-5-I-R91150 is taken up by the brain [24].

Effective blockade of 5-HT2a receptors in vivo was demonstrated in a study of

schizophrenic patients treated with risperidone or clozapine [25].

The aim of this study was to evaluate the 5-HT2a binding index in patients with

Anorexia Nervosa.

Methods

Patients

Patients were included in the study if they were aged between 16 and 30 years and

if they were diagnosed as having Anorexia Nervosa, according to the diagnostic

140

criteria of the Diagnostic and Statistical Manual of Mental Disorders Fourth

Edition [1]. Exclusion criteria were 1) a co-morbid psychiatric, major medical or

neurological disorder, 2) antidepressant, neuroleptic or electroconvulsive therapy

in the preceding year, 3) substance abuse 4) pregnancy or lactation period , and 5)

a Mini-Mental State Examination [26] score less than 28.

Healthy volunteers

Eleven age-matched healthy volunteers, 4 men and 7 women, were recruited

among the hospital staff. These subjects had no psychiatric or medical history, nor

a family psychiatric history. None used psychotropics or other relevant

medication or abused illegal drugs. All had a normal physical examination.

Ethical approval for the study was granted by the Local Ethics Committee. Both

patients and healthy individuals provided written informed consent to take part in

this study.

Tracer

123I-5-I-R91150 was synthetised by electrophilic substitution on the 5-position of the

methoxybenzamide group of R91150, followed by purification with high-

performance liquid chromatography. The product had a radiochemical purity of

more than 99 % and was negative for bacteria and pyrogen tests. A specific activity

of 10 Ci/µmol was obtained.

The tracer is a 5-HT2a antagonist with high affinity (Kd = 0.11 nM) and selectivity for

5-HT2a receptors. The selectivity of the ligand for 5-HT2a receptors with regard to


dopamine receptors, adrenoreceptors and histamine receptors is at least a factor of 50

[23, 27].

141

SPECT scanning

Thyroid blockade was achieved by administration of a single oral dose of 100 mg

potassium iodide prior to injection. All subjects received an intravenous injection

of 185 MBq 123I-5-I-R91150 in normal sitting conditions. SPECT scanning was

performed using a triple-headed high-resolution Toshiba gamma camera GCA-

9300 with fan-beam collimation. For 123I, the resulting transaxial image resolution

is 9.5 mm full-width at half-maximum (FWHM).

Since sequential dynamic SPECT brain scans have shown that the cortico-

cerebellar ratio reaches a plateau between 90 and 110 minutes reflecting pseudo-

equilibrium and remains stable thereafter for up to 8 hours [24], acquisition was

started between 110 and 140 minutes after tracer injection, according to previous

protocols [28]

A transmission scan (TCT-scan) was acquired before the emission scan, using

three 153-Gd rod sources. This scan was used for subsequent image coregistration

to stereotactic coordinates. Emission images were acquired during 40 minutes.

The whole brain volume was acquired within the single scanning session. Images

were reconstructed using filtered back-projection and corrected for scatter and

non-uniform attenuation [29].

Estimation of binding index

Analysis of the scans was performed blind to patient status. After automatic image

coregistration to stereotactic space using the transmission image (BRASS,

Nuclear Diagnostics) [30], a predefined volume-of-interest (VOI) analysis was

performed with 12 cortical regions. Radioactivity estimates in the cortex were

assumed to represent total ligand binding (specific + non-specific binding + free

ligand) [24]. Since the cerebellum is void of serotonin receptors [31] and therefore

represents non-specific activity, calculation of relative indices of specific binding

index (BI) was done by VOI normalisation to the activity per volume element in

the cerebellum. Under these pseudo-equilibrium circumstances, this binding index

is directly related to the in vivo receptor density (Bmax) and affinity (KD). Binding

142

Index was defined as (target activity – background activity in brain) / (background

activity) which was operationally estimated as (counts/pixel in frontal cortex) /

(counts/pixel in cerebellum).

Statistical methods

The equality of age- and BMI-distributions between diagnostic categories was

evaluated according to the Mann Whitney-U test and the equality of gender-

distributions with the Fischer Exact's test. As binding indexes were normally

distributed (Kolmogorov-Smirnov testing), t statistics were used to compare mean

levels between categories. Correlation analyses were used to examine any

relationships between binding index, BMI and disease duration.

Results

Demographic and physical variables

The total sample of 26 subjects were on average 23.8 years old (SD 4.1), with ages

ranging from 16 to 30 years. Mean ages were not significantly different (MWU=

52.5; p= 0.12) between the two study groups, i.e. 22.5 years (SD 2.5) for AN patients

and 25.6 years (SD 4.7) for healthy volunteers. There was a significant difference in

gender (Fisher’s Exact Test p= 0.022) between the normal volunteers (4 males; 7

females) and the patients with AN (no males; 15 females).

The BMI in the AN patients was 14.9 (SD1.6; range 11.9-16.9) which was

significantly different (MWU= 0; p< 0.0001) from the BMI of HV (mean 22.3; SD

1.4; range 20.5-24.5). The number of years of presence of disease was 4.3 years (SD

4.48; range 1-14 years). All patients were postmenarcheal and amenorrheic.

Healthy Anorexia Statistics

143

Volunteers Nervosa

Age 25.6 (2.5) 22.5 (4.7) MWU = 52.5 p= 0.12

Gender 4 M / 7 F 0 M / 15 F χ² = 6.45 FE: p= 0.02

BMI 22.3 (1.4) 14.9 (1.6) MWU = 0.00 p< 0.001

Disease duration (yrs) 3.7 (3.5)

Table 1

Demographic variables Anorexia Nervosa patients compared to healthy volunteers.

Binding index

Patients with AN, compared to HV, had a significantly reduced 5-HT2a binding

potential in the left frontal cortex, in the left and right parietal cortex and the left and

right occipital cortex (see Table 2). Additional adjustment for age did not alter the

results regarding the magnitude as well as the statistical significance of the observed

difference between both groups.

144

Healthy volunteers

(N=11)

Anorexia Nervosa

(N=15) Binding index

Mean (SD) Mean (SD)

T-statistic*

Frontal cortex

Left + Right

Left

Right

0.94

0.94

0.95

(0.14)

(0.15)

(0.13)

0.85

0.81

0.88

(0.13)

(0.13)

(0.13)

1.83

2.27 *

1.26

Parietal cortex

Left + Right

Left

Right

1.02

1.03

1.01

(0.18)

(0.21)

(0.18)

0.86

0.88

0.85

(0.15)

(0.17)

(0.14)

2.40*

2.08*

2.50*

Temporal cortex

Left + Right

Left

Right

0.91

0.93

0.88

(0.13)

(0.14)

(0.12)

0.91

0.91

0.90

(0.13)

(0.13)

(0.15)

0.04

0.49

-0.40

Occipital cortex

Left + Right

Left

Right

1.03

1.04

1.03

(0.17)

(0.20)

(0.15)

0.91

0.91

0.90

(0.11)

(0.12)

(0.11)

2.15*

2.10*

2.32*

* p< 0.05

Table 2:

Means and standard deviation of regional cortical binding index in patients with Anorexia Nervosa compared to normal volunteers.

Individual values of left frontal and biparietal BI were plotted in Figure 1. Since

there was a significant difference in gender-distribution between the HV and the AN,

individual values of males in the HV are indicated with an arrow.

145

Figure 2: Example of a 25 year old Anorexia Nervosa patient (right) compared to the anatomically standardized normal age-matched control population used for this study (left).

147

Figure 1: Plot of individual binding index values in bilateral parietal cortex and left frontal cortex of healthy volunteers and Anorexia Nervosa patients. Values of male healthy volunteers are indicated with an arrow.

There was a significant difference in the left-right ratio in the frontal cortex of

patients with eating disorders, compared to healthy volunteers (see Table 3 ).

Healthy Volunteers Anorexia Nervosa

Left/Right regional cortical ratio

Mean (S.D.)

Mean (S.D.)

T-statistic

t

Frontal cortex 1.01 (0.03) 1.04 (0.02) - 3.31**

Parietal cortex 0.99 (0.06) 0.99 (0.06) 0.07

Temporal cortex 0.97 (0.03) 1.00 (0.04) -1.60

Occipital cortex 1.00 (0.04) 1.00 (0.05) - 1.18

** p< 0.01

Table 3:

Means and standard deviations of regional cortical left-right ratios of binding index in patients with Anorexia Nervosa compared to normal volunteers

148

Correlation analyses did not reveal significant relationships between regional cortical

binding indexes and BMI and years of disease duration.

Discussion

In this in vivo study of cortical 5-HT2a receptors in patients with Anorexia Nervosa,

compared to healthy volunteers, a significantly reduced 5-HT2a binding index in the

left frontal, bilateral parietal and occipital cortex was demonstrated. Concerning the

left-right ratios of the BI, a significant difference (left < right) was demonstrated in

the frontal cortex.

When evaluated in the light of indirect studies assessing serotonergic function and 5-

HT2a binding status in low weight AN patients, the reduction in regional cortical

binding index are in keeping with the findings of neuro-endocrine challenge tests and

blood platelet studies that demonstrated a possible 5-HT2 involvement in the

pathogenesis of AN [14,15,16,17].

The reduction in binding index was not present in all cortical regions but was

restricted to the left frontal, the bilateral parietal and occipital cortex. This may be in

accordance with functional imaging studies, assessing cerebral blood flow or

metabolism, in patients with AN that have shown a regional hypometabolism in

frontal and parietal cortex [6,7,8]. The occipital cortex was not evaluated in these

reports. Neuropsychological studies have identified a disturbed body image

perception in anorectic patients [4], a cognitive function that is attributed to parietal

cortex. Others found also deficits in attention [3] and problem solving abilities [2,3],

which are mediated by the frontal cortex, and deficits in digit symbol test [32],

visuospatial construction [2,33] and mental arithmetic [34], which are related to

parietal functions.

However, it remains unclear whether this reduction in BI is a cause or a

consequence of AN and whether this reduction is trait- or state-dependent. Many of

149

the biological findings in low weight AN patients normalized or were even inversely

disturbed after long term weight restoration [10,13,35], and were state dependent and

possibly caused by the low weight or neuroendocrine status.

A reduction in 5-HT2a binding index can be caused by reduced estrogen

concentrations since it was demonstrated that estrogen increases 5-HT2a receptor

expression [36]. We did not evaluate estrogen concentrations in the patients in our

study but, since all girls were in postmenarchel amenorrheic state, it can be expected

that estrogen concentration would have been low and be directly or indirectly, the

driving force behind the reduction in 5-HT2a expression.

Another important confounder lies in the reduced food intake, especially concerning

protein intake, since the essential amino acid tryptophan is the precursor of serotonin.

Alternatively, as a third possible confounder, a reduced BI can occur as a

compensatory response to chronic over-release of 5-HT. Physical hyperactivity,

which is common among AN patients, may cause lipolysis of inter-muscular lipid,

resulting in release of free fatty acids. These fatty acids displace tryptophan from

albumin, leading to an increase in free tryptophan and in 5-HT turnover in the brain

[37,38,39]. All the aforementioned variables, estrogen, protein reduction and

physical hyperactivity, can be responsible for the reduction in 5-HT2a BI in our

study. However, one would then expect a global reduction in 5-HT2a BI which is

contrary to the regional reduction in our study, with a clear sparing of the left and

right temporal cortex.

Our findings can also be confounded by the mood status since depression is a

common co-morbid disorder in AN patients. This possibility cannot be ruled out,

however no functional imaging study of the 5-HT2a receptor in depressed [40,41]

or attempted suicide patients [28] showed a reduction in parietal BI, a finding that

was significantly present in our study. D’Haenen et al. demonstrated a frontal

asymmetry in 5-HT2a binding in depressed patients [43]. Finally, as a potential

methodological limitation of our study, the difference in gender distribution can

be a possible confounder. However, it is unlikely since gender-differences in BI

were not reported with this tracer [42], in correspondence with our own

(unpublished) data, and since for this study male BI values fell in the range of the

female healthy volunteers, as is shown in Figure 2.

150

Conclusion

A specific regional reduction in binding index of the 5-HT2a receptors is

demonstrated in low-weight Anorexia Nervosa patients, compared to healthy

volunteers. State-dependent variables of AN can be causal of the BI reduction. In

order to address this question, future work should aim to replicate the findings of

this study, in the same diagnostic group but also in other eating disorder

subgroups. A follow-up study of AN patients, with a re-evaluation of the 5-HT2a

binding index at long term weight restoration could help in differentiating if this

status is state-dependent or a trait variable.

151

References

1. American Psychiatric Association Diagnostic and Statistical Manual of Mental. Washington DC: American Psychiatric Association, 1994.

2. Szmukler G, Andrewes D, Kingston K, Chen L, Stargatt R, Stanley R. Neuropsychological impairment in anorexia nervosa: before and after refeeding. J Clin Exp Neuropsychol 1992; 14: 347-352.

3. Lauer C, Gorzewski B, Gerlinghoff M, Backmund H, Zihl J. Neuropsychological assessments before and after treatment in patients with anorexia nervosa and bulimia nervosa. J Psychiatr Res 1999; 33: 129-138.

4. Horne R, Van-Vactor J, Emerson S. Disturbed body image in patients with eating disorders. Am J Psychiatry 1991; 148: 211-215.

5. Franzen U, Florin I, Schneider S, Meier M. Distorted body image in bulimic women. J Psychosomatic Res 1988; 32: 445-450.

6. Delvenne V, Lotstra F, Goldman D, Biver F, De Maertelaer V, Appelboom-Fondu J, Schoutens A, Bidaut L, Luxen A, Mendelwicz J. Brain hypometabolism of glucose in anorexia nervosa: a PET scan study. Biol Psychiatry 1995; 37: 161-169.

7. Delvenne V, Goldman S, De Maertelaer V, Lotstra F. Brain glucose metabolism in eating disorders assessed by positron emission tomography. Int J Eating Dis 1999; 25: 29-37.

8. Nozoe S, Naruo T, Yonekura R, Nakabeppu Y, Soejima Y, Nagai N, Nakajo M, Tanaka H. Comparison of regional cerebral blood flow in patients with eating disorders. Brain Res Bull 1995; 36: 251-255.

9. Coppen AJ, Gupta RK, Eccleston EG, Wood KM, Wakeling A. Plasma tryptophan in anorexia nervosa. Lancet 1976; 1: 961.

10. Kaye W, Gwirtsman H, George D, Ebert M. Altered serotonin activity in anorexia nervosa after long term weight restoration: Does elevated cerebrospinal fluid 5-hydroxyindolacetic acied level correlate with rigid and obsessive behaviour? Arch Gen Psychiatry 1991; 48: 556-562.

11. Goodwin G, Shapiro M, Bennie J, Dick H, Carroll S, Fink G. The neuroendocrine response and psychological effects of infusion of L-tryptophan in anorexia nervosa. Psychol Med 1989; 19: 857-864.

12. Brewerton T, Jimerson D. Studies of serotonin function in anorexia nervosa. Psychiatry Res 1996; 62: 31-42.

13. O'Dwyer ALJ, Russell G. Serotonin activity in anorexia nervosa after long-term weight restoration: response to D-fenfluramine challenge. Psychol Med 1996; 26: 353-359.

14. Monteleone P, Brambilla F, Bortolotti F, La Rocca A, Maj M. Prolactin response to d-fenfluramine is blunted in people with anorexia nervosa. Br J Psychiatry 1998; 172: 438-442.

15. Ward A, Brown N, Lightman S, Campbell I, Treasure J. Neuroendocrine, appetitive and

152

behavioural responses to d-fenfluramine in women recovered from anorexia nervosa. Br J Psychiatry 1998; 172: 351-358.

16. Berk M, Kessa K, Szabo CP, Butkow N. The augmented platelet intracellular calcium response to serotonin in anorexia nervosa but not bulimia may be due to subsyndromal depression. Int J Eat Disord. 1997; 22: 57-63.

17. Spigset O, Andersen T, Hagg S, Mjondal T. Enhanced platelet serotonin 5-HT2A receptor binding in anorexia nervosa and bulimia nervosa. Eur Neuropsychopharmacol. 1999; 9: 469-473.

18. Collier D, Arranz M, Li T, Mupita D, Brown N, Treasure J. Association between the 5-HT2A gene promotor polymorphism and anorexia nervosa. Lancet 1997; 350: 412.

19. Sorbi S, Nacmias B, Tedde A, Ricca V, Mezzani B, Rotella C. 5-HT2A promoter polymorphism in anorexia nervosa. Lancet 1998; 351: 1785.

20. Hinney A, Ziegler A, Nothen M, Remschmidt H, Hebebrand J. 5-HT2A receptor gene polymorphism, anorexia nervosa and obesity. Lancet 1997; 350: 1324-1325.

21. Treasure J, Campbell I. The case for biology in the aetiology of anorexia nervosa [editorial]. Psychol.Med. 1994; 24: 3-8.

22. Study Group on Anorexia Nervosa. Anorexia nervosa: directions for future research. Int J Eat.Disord. 1995; 17: 235-241.

23. Mertens J, Terriere D, Sipido V, Gommeren W, Janssen P, Leysen J. Radiosynthesis of a new radioiodinated ligand for serotonin-5HT2-receptors, a promising tracer for gamma-emission tomography. J Labelled Compounds Radiopharmaceuticals 1994; 34: 795-806.

24. Busatto G, Pilowsky L, Costa D, Mertens J, Terriere D, Ell P, Mulligan R, Travis M, Leysen J, Gacinovic S, Waddington W, Lingford-Hughes A, Kerwin R. Initial evaluation of 123I-5-I-R91150, a selective 5-HT2A ligand for single-photon emission tomography, in healthy subjects. Eur J Nucl Med 1997; 24: 119-124.

25. Travis M, Busatto G, Pilowsky L, Mulligan R, Acton P, Gacinovic S, Mertens J, Terriere D, Costa D, Ell P, Kerwin R. 5-HT2a receptor blockade in patients with schizophrenia treated with risperidone and clozapine. A SPET study using the novel 5-HT2a ligand 123I-5-I-R-91150. Br J Psychiatry 1998; 173: 236-241.

26. Folstein MF, Folstein SE, McHugh PR. 'Mini-mental State'. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12: 189-198.

27. Abi-Dargham A, Zea Ponce Y, Terriere D, Al Tikriti M, Baldwin R, Hoffer P, Charney D, Leysen J, Laruelle M, Mertens J, Innis R. Preclinical evaluation of [I-123] R93274 as a SPECT radiotracer for imaging 5-HT2a receptors. Eur J Pharmacol 1997; 321: 285-293.

28. Audenaert K, Van Laere K, Dumont F, Slegers G, Mertens J, van Heeringen C, Dierckx R. Decreased frontal serotonin 5-HT2a receptor binding index in deliberate self harm patients. Eur J Nucl Med 2000; 27: 1800-1808.

29. Van Laere K, Koole M, Kauppinen T, Monsieurs M, Bouwens L, Dierckx R. Non-uniform transmission in brain SPET using 201-Tl, 99m-Tc and 153-Gd static line sources: antropomorphic dosimetry studies and brain quantification. J Nucl Med 2000;

153

41: 2051-2062.

30. Van Laere K, Koole M, D'Asseler Y, Versijpt J, Dumont F, Audenaert K, Dierckx R. .Automated stereotactic standardization of brain receptor data using single-photon transmission images. J Nucl Med 2001, 42(2), 361-375.

31. Pazos A, Probst A, Palacios J. Serotonin receptors in the human brain-IV. Autoradiographic mapping of serotonin-2 receptors. Neuroscience 1987; 21: 123-139.

32. Palazidou E, Robinson P, Lishman W. Neuroradiological and neuropsychological assessment in anorexia nervosa. Psychol Med 1990; 20: 521-527.

33. Mathias J, Kent P. Neuropsychological consequences of extreme weight loss and dietary restriction in patients with anorexia nervosa. J Clin Exp Neuropsychol 1998; 20: 548-564.

34. Hamsher K, Halmi K, Benton A. Prediction of outcome in anorexia nervosa from neuropsychological status. Psychiatry Res 1981; 4: 79-88.

35. Delvenne V, Goldman D, De Maertelaer V, Simon Y, Luxen A, Lotstra F. Brain hypometabolism of glucose in anorexia nervosa: normalization after weight gain. Biol Psychiatry 1996; 40: 761-768.

36. Connan F, Treasure J. Stress, eating and neurobiology. In: Hoek H, Treasure J, Katzman M, eds. Neurobiology in the treatment of eating disorders. Chicester: John Wiley & Sons; 1998: 211-234.

37. Blomstrand E, Cesling F, Newsholme E. Changes in plasma concentrations of aromatic and branched chain amino acides during sustained exercise in man and their probable role in fatigue. Acta Psychiatr Scand 1988; 133: 115-122.

38. Fisher G, Hollman W, De Meirleir K. Exercise changes in plasma tryptophan fraction and relationship with prolactin. Int J Sports Med 1991; 12: 487-489.

39. Chaouloff F, Kennett G, Serrurier B, Merino D, Curzon G. Amino acid analysis demonstrates that increased plasma free tryptophan causes increases of brain tryptophan during exercise in the rat. J Neurochem 1986; 46: 1647-1650.

40. Biver F, Wikler D, Lotstra F, Damhaut P, Goldman S, Mendlewicz J. Serotonin 5-HT2 receptor imaging in major depression: focal changes in orbito-insular cortex. Br J Psychiatry 1997; 171: 444-448.

41. Massou JM, Trichard C, Attar-Levy D, Feline A, Corruble E, Beaufils B, Martinot JL. Frontal 5-HT2A receptors studied in depressive patients during chronic treatment by selective serotonin reuptake inhibitors. Psychopharmacology (Berl) 1997; 133: 99-101.

42. Baeken C, D'Haenen H, Flamen P, Mertens J, Terriere D, Chavatte K, Boumon R, Bossuyt A. 123I-5-I-R91150, a new single-photon emission tomography ligand for 5-HT2a receptors: influence of age and gender in healthy subjects. Eur J Nucl Med 1998; 25: 1617-1622.

43. D'Haenen H, Bossuyt A, Mertens J, Bossuyt-Piron C, Gijsemans M, Kaufman L. SPECT imaging of serotonin 2 receptors in depression. Psychiatry Res Neuroimaging 1992; 45:227-237.

155

EPILOGUE

This chapter represents a critical review of the studies described in this thesis. The

two major research lines, the neuropsychological activation studies (chapter 1-3) and

the serotonin-2a studies (chapter 4-6), are summarized and brought into a more

unifying but more hypothetical construct. This has the advantage that it can function

as a framework for further investigations but has the disadvantage that it is, again, a

further reduction of a very complex reality. Based upon the research findings,

potential clinical applications are formulated and implications for further research are

discussed.

General overview of the thesis

As medical clinicians we are trained to take anamneses from patients, to investigate

them with additional diagnostic instruments, to make up a differential diagnosis, to

diagnose and to start a therapy. In the field of psychiatry we are confronted with a

huge overlap between nosological entities, as is discussed in the Introduction of the

thesis. If we listen to the critiques of Van Praag on the nosological categorical

diagnosing process and we follow his suggestion to look at (neuro)psychological

disturbances that are present transnosologically, i.e. across the boundaries of

nosological entities, we can look at psychiatric diseases from a different perspective.

From a neuropsychological point of view, a vast amount of patients, across

diagnostic boundaries, suffer from dysexecutive behavioural disturbances, such as

reduced planning capacities and mental inflexibility, or from behavioural inhibition,

leading to impulsive conduct. These behavioural and cognitive disturbances can be

related to a dysfunction in circuits involving prefrontal cortex structures. Functional

brain imaging, using SPECT and a neuropsychological activation paradigm, can

offer an investigational tool. We demonstrated the feasibility of the Stroop Test

(chapter 1) and the Verbal Fluency Test (chapter 2) in healthy volunteers. The

advantage and novelty of the technique lies in the fact that the modality is feasible in

156

a one hour protocol, offers neuropsychological testing conditions and can be

evaluated with Statistical Parametrical Mapping (SPM). This technique was applied

in depressive suicidal patients (chapter 3) and demonstrated a blunted increase in

perfusion in prefrontal cortex structures in this population, compared to healthy

volunteers.

Again, as clinicians, we try to understand the biological underpinnings of psychiatric

diseases, primarily to be able to put our therapeutic armament to better use. Over and

over again, we are puzzled by the fact that for some patients, a certain psychotropic

drug benefits in the recovery of the disease, while patients, belonging to the same

diagnostic category, have no benefit at all from the same psychotropic drug. Also,

it’s remarkable that drugs initially designed to treat patients in one diagnostic

category, e.g. anti-depressants, are now successfully applied in diseases that belong

to distinct categories, such as anxiety disorders, eating disorders or impulsivity

disorders.

Van Praag suggests a possible answer to these puzzling observations, as is discussed

in the introduction of this thesis. From that point of view, we partially directed our

research to groups of patients with comparable cognitive or behavioural problems.

Both patients who attempt suicide and patients with eating disorders, in this case

patients with anorexia nervosa, are described to be impulsive by nature. From a

neuroanatomical point of view, earlier research demonstrated the involvement of

prefrontal cortex structures in the pathogenesis of the condition. Indeed, reduced

frontal perfusion or metabolism and neuropsychological impairment pointing at a

frontal cortex dysfunction was demonstrated. From a biochemical point of view, a

vast amount of literature demonstrated the involvement of the serotonergic system in

impulsive suicide attempters and in patients with eating disorders.

We demonstrated a reduced serotonin-2a binding index in the frontal cortex of

suicide-attempters (chapter 4 and 5) and patients with anorexia nervosa (chapter 6).

Interestingly, in the suicide-attempt patients, the binding index is significantly more

reduced in the self-injury group compared to the self-poisoning group. In the

anorexia nervosa patients, the binding index is also reduced in the parietal cortex, a

finding which is in line with the neuropsychological disturbance of a disturbed body

image and with PET glucose metabolism findings. We also demonstrated a

157

correlation with Harm Avoidance, a temperament factor (Temperament and

Character Inventory, Cloninger) which is likely to be stable throughout life.

Summary of findings

In the first chapter the classical Stroop interference task, a prefrontal

neuropsychological activation probe in SPECT, was validated in healthy volunteers.

The testing procedure following the single-day split dose activation paradigm and the

application of Statistical Parametric Mapping (SPM) to determine voxelwise

significant changes, is explained. Activation regions were demonstrated in the

prefrontal cortex and in the anterior cingulate, which is in concordance with previous

PET and fMRI findings during Stroop-like interference tasks. The advantages, the

patient-friendly procedure and the near-neuropsychological standardized test

conditions, and the major drawbacks of this SPECT technique, the lower spatial

resolution and the stringent limit on the number of possible conditions, are discussed.

In the second chapter the Verbal Fluency task, both in its letter fluency and category

fluency modality, as a prefrontal activation probe in SPECT was validated in healthy

volunteers. From a methodological point of view, the SPM results were compared

and found superior to a Voxel-of-interest (VOI) approach. The letter fluency and the

category fluency activation paradigm had a differential brain activation pattern in the

prefrontal cortex and showed both an activation in the anterior cingulate cortex.

These findings were in concordance with previous PET and fMRI findings during

tasks that require semantic and phonological processing.

In the third chapter the Verbal Fluency task, as a prefrontal SPECT activation probe,

was tested in depressed suicide-attempt patients and results were compared to

healthy volunteers. Depressed suicide-attempters showed a blunted increase in

perfusion changes in the prefrontal cortex that paralleled the hampered test results.

Methodological restrictions concerning group uniformity, medication bias and

subjective effort of the participants are discussed. The findings of a blunted increase

in prefrontal blood perfusion changes as a possible biological underpinning of

reduced drive and loss of initiative in depressed suicide- attempters are evaluated.

158

In the fourth chapter the brain serotonin-2a binding index in patients who very

recently attempted suicide, was estimated through the highly specific radio-iodinated

5-HT2a receptor antagonist 123I-5-I-R91150 and compared to healthy volunteers.

Methodology of image acquisition and image processing, including a time-activity in

a healthy volunteer, is discussed. Deliberately self-harming patients had a

significantly reduced mean frontal binding index after correction of age when

compared with controls. The reduction was more pronounced among deliberately

self-injury patients than among deliberately self-poisoning patients (DSP). It is

concluded that brain SPET of the 5-HT2a serotonin receptor system in suicide-

attempt patients who are free of drugs influencing the serotonergic system shows in

vivo evidence of a decreased frontal binding index of the 5-HT2a receptor, indicating

a decrease in the number and/or in the binding affinity of 5-HT2a receptors.

In the fifth chapter, we investigated the prefrontal 5-HT2a receptor-binding index,

hopelessness, using Beck’s Hopelessness Scale, and personality characteristics, using

Cloninger’s Temperament and Character Inventory, in the suicide-attempt patients

that were described in the fourth chapter. When compared to normal controls,

suicide-attempt patients had a significantly lower binding potential of frontal 5-HT2a

receptors, a higher level of hopelessness, a higher score on the temperament

dimension of harm avoidance and lower scores on the character dimensions of self-

directedness and cooperativeness. A significant correlation was found between harm

avoidance, hopelessness and binding index in the deliberately self harm population.

Limitations of the study were discussed and the meaning of the current findings was

evaluated with regard to the understanding, treatment and prevention of suicidal

behaviour.

In the sixth chapter the brain serotonin-2a binding index in patients with anorexia

nervosa was estimated through 123I-5-I-R91150 and compared to healthy volunteers.

Patients with eating disorders, compared to healthy volunteers, have a significantly

reduced 5-HT2a binding index in the left frontal cortex, in the left and right parietal

cortex and the left and right occipital cortex. There was a significant left-right

asymmetry in the frontal cortex (left<right). These results are in accordance with

diminished metabolism and perfusion of frontal and parietal cortex in recent neuro-

imaging studies and imply localized disturbed serotonergic function. The data are

159

discussed in the light of possible confounding factors that are related to the low-

weight Anorexia Nervosa status.

Review of the results: the issue of coincidence by chance, causality, reversed

causality and state or trait features

Neuropsychological activation studies

Concerning the prefrontal neuroactivation studies, we demonstrated a blunted

increase in perfusion in the prefrontal cortex after a prefrontal challenge test in

depressed patients that very recently attempted suicide (see chapter 3).

This finding is not likely to be a coincidence by chance since several lines of

research demonstrated a dysfunctional prefrontal cortex in depressed patients and

suicide attempters. Neuropsychological testing revealed deficits in this population

pointing at dysexecutive symptoms (see Chapter 3, Results) and impaired

behavioural control, symptoms related to prefrontal cortex dysfunction and

functional imaging studies showed a reduced prefrontal perfusion and metabolism in

patients with depression or urge to suicide (see Chapter 3, Introduction), but also in

patient populations with impulsivity [1]or dysexecutive symptoms [2] as core

features.

The next question is about causality. No study could ever demonstrate a direct

causality between hypofrontality, here demonstrated through a blunted increase

following prefrontal activation, and the occurrence of a suicide attempt. However,

cognitive psychology provided us with indirect evidence of disturbed prefrontal

functioning as a major causative contributor in a suicide attempt, since it was

demonstrated that hopelessness and the inability to see solutions for problems,

which were related to a prefrontal dysfunction, are major contributors in the suicidal

process [3].

There are no arguments for direct reversed causality since there is no evidence that

160

an impulsive act or dysexecutive behaviour provokes hypofrontality or impaired

prefrontal neuropsychological function, and hence, a blunted increase in a

neuroactivation paradigm.

A fourth question is about the state or trait occurrence of hypofrontality. Some

studies demonstrated that patients, who had a disorder with dysexecutive symptoms

or impulsive behaviour, show, at least partially, normalized neuropsychological

testing [4] and frontal perfusion and metabolism [5] after recovery of their

disturbance. This indicates at least a state component. However, other studies

indicate also a trait component since Raine et al (1994) demonstrated that an

incarcerated population of homicide offenders showed bilateral diminution of

glucose metabolism in frontal cortex compared to healthy controls [6]. Others

demonstrated an inverse relationship between a life history of impulsive behaviour

and the regional cerebral glucose metabolism in the frontal, right temporal cortex and

the cingulate cortex [1]. The occurrence of a state and trait role of hypofrontality

suits the hypothesis that subjects, with a certain personality and temperamental

profile, defined as a set of automatic emotional responses to experience that are

moderately heritable and stable throughout life [7], can have hypofrontality as

biological underpinning (“trait component”). These patients are then vulnerable to

have overt prefrontal symptoms, such as impulsivity and reduced planning

capabilities, in certain stressful conditions (“state component”).

Serotonin-2a studies

Concerning the serotonin-2a studies, a reduction in binding index in the prefrontal

cortex in suicide-attempt patients and in patients with anorexia nervosa. In the

suicide-attempt patients the binding index was negatively correlated to Hopelessness

and to the personality factor of Harm Avoidance.

Again firstly, it is unlikely that these findings are the results of a coincidence by

chance. A disturbance of the serotonergic neurotransmission in these two conditions

was demonstrated through indirect measurements of serotonin metabolites in blood

and cerebrospinal fluid and through pharmacological challenge tests (see

Introduction chapter 4 and 5). Moreover the contribution of the serotonin-2a

161

receptor was demonstrated through post-mortem studies in completed suicides (see

Introduction chapter 4) and indirectly through blood platelet serotonin-2a studies (see

Introduction chapter 4 and 5). The fact that the reductions were predominantly in the

prefrontal cortex can be explained by the fact that there is evidence of robust

serotonergic innervation in prefrontal cortex regions [8] and that high concentrations

of 5-HT2a and 5-HT1a receptors have been identified in the human prefrontal cortex

[9]. There is also evidence from preclinical research that regionally specific effects

of serotonin activity influence behaviour. Primate studies demonstrated that the

number of 5-HT2a receptors in the frontal cortex were inversely related to aggressive

behaviour, but not elsewhere in the brain. In contrast, 5-HT2a receptor numbers in

the prefrontal and temporal cortex were directly correlated with prosocial behaviour

[10]. A high density of 5-HT-2a receptors in frontal cortex is thought to promote

cooperative grooming behaviour, whereas decreased level of serotonin, reflected by

low 5-HT2a density, might promote aggressivity. But also in human studies,

fenfluramine challenge tests showed elevation in prefrontal cortical metabolism in

healthy controls [11], and in contrast, depressed patients demonstrated no significant

changes in glucose metabolism in response to fenfluramine compared to placebo

[12]. The same finding was demonstrated in impulsive aggressive patients where no

significant differences were seen in baseline scans between normal subjects and

patients. But, impulsive patients had blunted responses in glucose metabolism to

fenfluramine in the orbital frontal cortex, in the ventral medial frontal cortex and in

the cingulated cortex as compared to normal subjects [13].

Concerning the question about causality, a causal relation between serotonin and

impulsivity [14], mood lowering [15], relapse in eating disorder behaviour [16] and

the urge to suicide [17] was demonstrated through tryptophan depletion. Also

inverse correlations between 5-HT or basal CSF 5-HIAA concentrations and

inwardly-directed aggressive (ie suicidal) behaviour [18], inverse relationship

between a life history of aggressive behaviour and basal CSF 5-HIAA in patients

with personality disorders, indicated a causal relationship between serotonin and

impulsivity [19]and inverse correlation between aggression assessments and CSF-5-

HIAA in children and adolescents with behavioural disruptive disorder or OCD [20].

Concerning a causal relation between the binding index of serotonin-2a receptors and

Field Code ChangedUnknown

162

the urge to suicide or impulsivity, only indirect arguments indicate a possible causal

relationship between the number of serotonin-2a receptors and suicidality. Selective

serotonin re-uptake inhibitors were found to increase the number of 5-HT2a receptors

[21] and SSRIs showed a reduction in the repetition rate of deliberately self-harming

in non-depressed patients [22].

Can there be reversed causality? Theoretically, a suicide attempt could lead to a

stress induced cortisol surge, that can lower the brain serotonin concentration and

that can further lead to postsynaptic serotonin-2a receptor changes that are measured

as a reduced serotonin-2a binding index. However, we have not found any

straightforward arguments supporting this hypothetical construct. Interestingly, a

case study of one patient that attempted suicide, showed a low peripheral serotonin

measurement just before the attempt and a dramatic 70-fold blood serotonin increase

just after the attempt, which finally reached the reference interval of healthy subjects.

Possibly, this huge raise in serotonin could have provoked a rapid down-regulation of

post-synaptic receptors, including the serotonin-2a receptor [23]. This is unlikely to

have happened in our study since the reduced serotonin-2a binding was

predominantly in the prefrontal cortex, with the sparing of some cortical regions.

The next issue, the trait or state nature of the reduction in binding index is important,

both from clinical and pathogenetic point of view. Concerning serotonin in

impulsivity, suicidality and eating disorders, there are arguments for both a state and

trait involvement. Reduced serotonergic function as a trait factor was demonstrated

through estimations of CSF 5-HIAA in subjects with a life-time episode of hetero-

aggression [19]and impulsive suicidality [24] and through neuro-endocrine challenge

tests in suicide attempters and outward-directed impulsive subjects. Elevated

concentrations of 5-hydroxyindoleacetic acid in the cerebrospinal fluid after recovery

in eating disorder patients suggest that altered serotonin activity in AN and BN is a

trait-related characteristic [25]. However, Brewerton suggests that the increased

serotonergic function after recovery may be phase dependent, arising from a rebound

up-regulation of the system [26]. On the other hand, the aforementioned diet

tryptophan depletion experiments in impulsivity [14], eating disorder behaviour [16]

and suicidality [17] demonstrate a state related role for serotonin in these disorders.

Both CSF 5-HIAA [27] and fenfluramine challenge tests [28] in anorexia nervosa

163

subjects showed a state-related reduction in serotonergic function that is reversed

once biological and subjective recovery occur.

Concerning the state or trait characteristics of serotonin-2a binding index, there is

one increasingly important research line that supports the trait and genetic hypothesis

that serotonin-2a receptors are involved in impulsive, suicidal and eating disorder

behaviour. Recently, a 5-HT2A receptor promotor polymorphism was reported to be

significantly increased in patients with anorexia nervosa [29, 30], although others

could not confirm this finding[31]. Similarly, concerning suicidality, some authors

[32] did, but others [33] did not find the increased 5-HT2A receptor promotor

polymorphism.


In all chapters, methodological issues that are particularly related to the topic are

discussed. In this paragraph, more general methodological considerations are dealt

with. Concerning the neuropsychological activation studies, the SPECT technique in

neuroactivation studies, the role of SPM and VOI analyses and the single-day split-

dose activation paradigm will be discussed. Concerning the serotonin-2a studies,

most methodological considerations are extensively discussed in the chapters 4-6. In

this paragraph, comment on the low number of patient subjects in the suicide attempt

study and on the additional psychological tests will be given.

Neuropsychological activation studies

Concerning the neuropsychological activation studies, most methodological

problems are connected to the nature of the SPECT technique itself and to the

physical properties of the tracer.

Compared to the other functional neuro-imaging modalities, fMRI offers the

possibility of almost unlimited repeatability of studies owing to the absence of any

radiation burden, in combination with a spatial resolution unmatched by any other

functional imaging technique. PET, due to the short half-life of the perfusion tracers,

164

allows the performance of a limited number of repetitions of both baseline and

activation tasks, and its optimal spatial resolution is still superior to that of SPET.

But inherent to the use of fMRI and PET is the almost real-time functional imaging

of the brain in combination with a rigid experimental set-up. Indeed, subjects are


position. In contrast, brain SPET allows to freeze a mental state with a duration of 2-

3 min due to the short accumulation period for HMPAO and ECD. This opens up the

possibility of applying the tracer outside the nuclear medicine department, or at least

away from the arousing conditions linked to the camera. Hence, the tracer can be

injected intravenously while the subject is comfortably sitting at a table in the

research room of the neuropsychologist and is performing the neuropsychological

test. Evidently, this condition approaches the classical neuropsychological test

conditions much more closely than the afore mentioned experimental conditions

linked to fMRI or PET. If the researcher has the access to all modalities, PET, fMRI

and SPECT, he must carefully evaluate his research goal and balance the afore-

mentioned pros and cons of the different functional imaging modalities.

Other methodological problems lie in the activation paradigm itself. In these studies,

images of the resting and activation condition can be compared in different ways.

Up to now, most reports use the ROI/VOI approach. This technique is time-

consuming and prone to operator-bias. At first, one can circumvent this problem by

applying a template on which the images can be fitted and automated VOI analysis

can be performed, as was demonstrated by our group [34]. Alternatively, one can

use SPM to statistically compare the two conditions. However, the SPM technique

was originally designed for image comparison in PET and fMRI and is optimised for

comparison of large image sets, since this is commonly generated with PET or fMRI,

due to its multiple condition modality. The validation of SPM in SPECT activation

studies was carried out by our group [35]. We further evaluated our results, generated

by the SPM method, by comparing them with the ROI/VOI method (see chapter 2).

Of course, one of the most critical items in applying SPM in SPECT lies in the low

number of images, due to the limited repeatability, that can be generated in SPECT.

This can be countered by including both a large population of patients and healthy

controls (N=30) in the study. We are aware of the fact that we only included ten

165

patients and ten healthy controls in the Stroop, letter fluency and category fluency

paradigm. However, Lahorte et al. demonstrated that a group size can be lowered to

ten subjects when activation levels of 10% are expected [35]. Statistical Parametric

Mapping analysis in SPECT, like in PET and fMRI, is the technique of choice to

study differences between conditions for groups of subjects. The technique is hardly

applicable for individual patient evaluation since our results are not transferable to

individual patient evaluations. The ROI/VOI technique, on the other hand, can be

used to evaluate individual patients as was tentatively demonstrated in a

neuropsychological activation study with the Stroop test in patients with severe head

trauma [36].


paradigm, the main motivation being our search for a subject-friendly experimental

design. Compared with a 2-day protocol, our 1-day design is experienced as much

more comfortable by out-patients. A second advantage of the 1-day paradigm is the

fact that the physiological variability in brain perfusion on different days or due to

spontaneous activations and changes in mental activity is minimized. This can be

especially important for patients with mood disorder or schizophrenia.

It needs to be stressed that, at this moment, the interpretation of neuropsychological

activation studies in psychiatric patients is only of pathophysiological, and not

diagnostic value, since these procedures need further validation.

Serotonin-2a studies

In the fourth and fifth chapter we report on the investigation in a population of

patients that very recently attempted suicide, irrespective of their syndromal

diagnosis. We cannot overestimate the fact that only nine patients with a very recent

suicide attempt were described in the study. Subsequently, this group was even

further split in a group that used non-violent methods (N=6) and violent methods

(N=3). However we found statistically different binding indices between healthy

volunteers, patients that used a non-violent and patients that used a violent method.

Now, we have included 25 patients in the study and feel more comfortable since the

166

initial results could be confirmed. Due to the limited number of patients included in

the study, a possible effect of selection bias must be taken into account.

In the research protocol, we only included systematically an evaluation of

personality structure through the TCI, and the Beck Hopelessness Scale, besides

the clinical investigational tools as described in the studies. We did not

systematically plan a battery of neuropsychological testing and a series of

psychological questionnaires, such as the Barrett Impulsiveness Scale and the

Spielberger State and Trait Aggression Inventory. Post hoc, in order to test more

complex psychobiological hypotheses, we now include psychological

questionnaires (Barrett Impulsiveness Score, Spielberger Trait and State Anxiety)

and a neuropsychological testing battery (d2 attention test, Stroop Colour Word

Test, Controlled Oral Word Association Test, Tower of London, Wisconsin Card

Sorting Test, Trail Making Test, Auditory Verbal Learning Test., Visual Design

Learning Test) to our biological research. We believe that the disposal of these

additive results can refine the conclusions that are drawn from these experiments.

To our belief, the major drawback of these studies is that it cannot, due to its cross-

sectional nature, differentiate between state or trait dependency. Follow-up studies

must be planned with different intervention strategies in order to rule out the state or

trait status of the finding. This could help in optimising interventional tools and

could offer an important pathophysiological insight in the pathogenesis of the disease

or symptom.

I t needs to be stressed, at this moment, that serotonin-2a binding index studies have

only relevance in studying pathophysiology in populations of psychiatric patients,

and have very limited value in interpratation of individual patients.

Clinical implications and recommendations for further study

Clinical implications of the neuropsychological activation studies

From a clinical point of view, groups of patients who have a blunted increase in

167

frontal perfusion are likely prone to have behavioural and cognitive symptoms that

are related to prefrontal dysfunctions. It is clinically important to identify and

delineate these patient populations in order to direct specific preventive and

therapeutic measures to these groups For instance, if planning capacity or mental

flexibility is impaired, one can expect that patients have difficulties to find solutions

to (even daily life) problems. This can lead to a feeling of behavioural impotence

and brings subjects to the perception of being unable to change the situation, not now

and not in the near future. This leads to a feeling of entrapment and hopelessness. If

this is combined with impaired inhibition, behavioural supervision is reduced and

patients are more vulnerable to choose for a suicide attempt as a radical but defective

strategy in problem solving. But also from a therapeutic point of view, one must take

into account the impaired prefrontal functions. The therapist must adjust

psychotherapeutic strategies to the prefrontal malfunction since attention functions

and learning capacities are reduced, and hence, classical psychotherapeutic

interventions are not resorting the same effect compared to subjects with unaffected

prefrontal functions.

Recommendations for future neuropsychological activation studies

Concerning the neuropsychological activation studies, this work primarily aimed

at testing the feasibility of the one day split dose neuropsychological activation

paradigm with prefrontal tests. It was only tested in a group of patients with a

depressive episode and a recent suicide attempt. For the moment the technique has

been tested in other patient populations (Stroop test in brain damaged patients, K.

Audenaert), with more experimental paradigms in healthy volunteers (divided

attention paradigm, M Sc thesis G. Vermeir, University Ghent), with other

established prefrontal tests in neurological patient populations (Wisconsin Card

Sorting Test in Parkinson’s disease patients and elderly healthy volunteers, PhD

thesis B. Pickut, Middelheim Hospital, Antwerp), in adult psychiatric patient

populations (Tower of London test in depressed suicidal patients and healthy

volunteers, M Sc thesis H. Pyck, University Ghent), and in child psychiatric

populations (Stroop test in adolescent delinquents, R. Vermeire, Middelheim

Hospital, Antwerp). These experiments confirmed our initial findings on the

168

feasibility of this technique. For the evaluation of prefrontal function on patient-

group level, we believe that this technique has now established feasibility and can

be extended to other neuropsychological prefrontal activation paradigms and to

other patient groups. The evaluation of prefrontal functions with

neuropsychological activation paradigms in patients with personality disorders,

such as borderline or antisocial PD, or in patients with disorders of impulsivity,

such as patients with periodic explosive disorder, pyromania, kleptomania and in

patients with eating disorders has not been previously investigated and published.

Up to now, we carried out cross-sectional paradigms. Further studies should plan

longitudinal experiments in which patients are scanned again after remission of

their symptoms and disease. This could offer valuable information on the trait or

state nature of our findings and can help the clinician to select patient groups that

need close supervision and secondary prevention concerning their relapse

vulnerability. Moreover, longitudinal studies can offer an interesting tool in

evaluating the impact of therapy, be it psychotherapeutic and/or pharmacological

in nature, on the restoration of prefrontal function and perfusion in the prefrontal

cortex.

Clinical implications of the serotonin-2a studies

Concerning the serotonin-2a studies, the clinical importance of these studies lies, at

first, in the identification of groups of patients with deficient serotonin-2a post-

synaptic receptor binding. Ongoing studies in patients with impulsive personality

disorders and in patients with bulimia nervosa seem to confirm our findings in

disorders with “impulsivity” as a transnosological symptom. Up to now, the value of

an estimation of the serotonin-2a binding index in individual patients has not been

assessed. Practically, and for the time being, a significant reduction in binding index

in an individual patient, compared to a normal and age-matched database can be

interpreted as an extra clue in the decision-making process for risk assessment and

therapeutic intervention. The presence of a reduced serotonin-2a binding index can,

for the moment and in the absence of follow-up interventional studies, add to the

clinical decision to select serotonergic anti-depressants, e.g. SSRI’s, in the therapy of

169

the patient, irrespective of the presence of a depressive or anxiety disorder (see

Discussion chapter 4). Further studies, as described in the next paragraph, can help

in further elucidating the value of an individual study result.

Recommendations for future serotonin-2a studies

We only reported on two groups of patients, a group of patients with depression and

recent suicide attempt and a group of patients with anorexia nervosa. The results of

these studies, although quite comparable, could certainly not be extended to other

populations with similar behavioral or cognitive dysfunctions. The reason that we, at

first, have chosen the afore-mentioned populations lies in the fact that our Hospital is

a referral site for patients with eating disorders (M. Vervaet) and patients with

suicidal behaviour (C. van Heeringen). Therefore, firstly, our initial findings should

be replicated in larger and extended populations. For the moment, the study of

patients that attempted suicide is extended to 24 subjects and seems to confirm our

initial findings [37]. The group of anorexia nervosa patients was broadened to

bulimia nervosa patients and preliminary results indicate differences in serotonin-2a

binding index between AN and BN. This finding needs confirmation in larger series.

Further research on impulsive behaviour patients should further be extended to

patients with personality disorders with outward and/or inward directed aggression,

with obsessive-compulsive disorder and to patients with intermittent explosive

disorder, pyromania, trichotillomania, kleptomania etc., but also to neurological

disorders with impulsivity as a core symptom.

Longitudinal studies must be planned in order to evaluate the serotonin-2a binding

index after therapeutic interventions that can be psychotherapeutic and/or

pharmacological in nature. This can offer more insight in the pathogenesis of

diseases and again, help the clinician to select patient groups that need close

supervision and secondary prevention concerning their relapse vulnerability.

Interestingly, our group extended the use of this tracer to other species with

impulsive aggression as a behavioural problem. The technique was found to be

feasible in healthy Shepherd dogs (n=10) (K. Peremans, personal communication)

and is now extended to impulsive aggressive dogs. Preliminary research

170

demonstrated an altered binding index on visual inspection (K. Peremans, personal

communication), suggesting that the altered serotonin-2a binding index is present

across the boundaries of species.

Another urgent research issue lies in the investigation of the applicability of study

results to the individual subject. A large, age-dependent normal database should be

made up in order to compare individual patient results to a mean group result of

healthy volunteers that are comparable in age. Also further efforts need to be

undertaken to label the R91150 molecule with 18-fluorine to make a PET tracer

available. This would offer possibilities for absolute quantification and present a

better resolution in comparison to the iodine-labelled SPECT tracer.

Since the investigation of the serotonin-2a receptor was found to be an elegant

technique, future research should be directed to developping other radiotracers that

are highly selective to other serotonin-receptor subtypes. Especially, the availability

of a highly selective serotonin-1a tracer in biological psychiatric research could shed

a new light on the biological underpinnings of psychiatric diseases and behavioural

or cognitive symptoms. Current research (Vandecapelle & Slegers, Department

Radiopharmacy, University Ghent) is promising concerning the development of a

selective serotonin-1a tracer.

171

References

1. Goyer P, Andreason P, Semple W, Clayton A, King A, Compton-Toth B, Schulz S. Positron-emission tomography and personality disorders. Neuropsychopharmacology 1994; 10: 21-28.

2. Fontaine A, Azouvi P, Remy P, Bussel B, Samson Y. Functional anatomy of neuropsychological deficits after severe traumatic brain injury. Neurology 1999; 53: 1963-1968.

3. Williams M. Cry of pain: understanding suicide and self-harm. London: Penguin, 1997

4. Trichard C, Martinot JL, Alagille M, Masure MC, Hardy P, Ginestet D, Feline A. Time course of prefrontal lobe dysfunction in severely depressed inpatients: a longitudinal neuropsychological study. Psychol Med 1995; 25: 79-85.

5. Bench C, Frackowiak R, Dolan R. Changes in regional cerebral blood flow on recovery from depression. Psychol Med 1995; 25: 247-261.

6. Raine A, Buchsbaum M, Stanley J, Lottenberg S, Abel L, Stoddard J. Selective reductions in prefrontal glucose-metabolism in murderers. Biol Psychiatry 1994; 36: 365-373.

7. Cloninger R.. The genetics and psychobiology of the seven-factor model of personality. In: Silk K, eds. Biology of Personality Disorders. Washington: American Psychiatry Press; 1998: 66-93.

8. New A, Novotny S, Buchsbaum M, Siever L. Neuro-imaging in impulsive-aggressive personality disorders. In: Maes M, Coccaro E, eds. Neurobiology and clinical views on aggression and impulsivity. Chicester: John Wiley & Sons; 1998: 81-93.

9. Arango V, Ernsberger P, Marzuk PM. Autoradiographic demonstration of increased serotonin 5HT2 and beta-adrenergic receptor binding sites in the brain of suicide victims. Arch Gen Psychiatry 1990; 47: 1038-1044.

10. Raleigh M, Brammer G. Individual differenced in serotonin-2 receptors and social behavior in monkeys. Society for Neuroscience Abstracts 1993; 19: 592.

11. Mann J, Malone K, Diehl D, Perel J, Nichols T, Mintun M. Positron emission tomographic imaging of serotonin activation effects on prefrontal cortex in healthy volunteers. J Cereb Blood Flow Metab 1996; 16: 418-426.

12. Mann J, Malone K, Diehl J, Perel J, Cooper T, Mintun M. Demonstration in vivo of reduced serotonin responsivity in the brain of untreated depressed patients. Am J Psychiatry 1996; 153: 174-182.

13. Siever L, Buchsbaum M, New A, Spiegel-Cohen J, Wei T, Hazlett E, Sevin E. PET studies and fenfluramine in impulsive patients. Neuropsychopharmacology 1999; 20: 413-423.

172

14. LeMarquand D, Benkelfat C, Pihl R, Palmour R, Young S. Behavioral disinhibition induced by tryptophan depletion in nonalcoholic young men with multigenerational family histories of paternal alcoholism. Am J Psychiatry 1999; 156: 1771-1779.

15. Smith K, Fairburn C, Cowen P. Relapse of depression after rapid depletion of tryptophan. Lancet 1997; 349: 915-919.

16. Smith K, Fairburn C, Cowen P. Symptomatic relapse in bulimia nervosa following acute tryptophan depletion. Arch Gen Psychiatry 1999; 56: 171-176.

17. Hughes J, Dunne F, Young A. Effects of acute tryptophan depletion on mood and suicidal ideation in bipolar patients symptomatically stable on lithium. Br J Psychiatry 2000; 177: 447-451.

18. Asberg M, Traskman L, Thoren P. 5-HIAA in the cerebrospinal fluid: a biochemical suicide predictor? Arch Gen Psychiatry 1976; 33: 1193-1197.

19. Brown G, Goodwin F, Ballenger J, Goyer P, Major L. Aggression in humans correlates with cerebrospinal-fluid amine metabolites. Psychiatry Res 1979; 1: 131-139.

20. Kruesi MJ, Rapoport JL, Hamburger S, Hibbs E, Potter WZ, Lenane M, Brown GL. Cerebrospinal fluid monoamine metabolites, aggression, and impulsivity in disruptive disorders of children and adolescents. Arch Gen Psychiatry 1990; 47: 419-426.

21. Massou JM, Trichard C, Attar-Levy D, Feline A, Corruble E, Beaufils B, Martinot JL. Frontal 5-HT2A receptors studied in depressive patients during chronic treatment by selective serotonin re-uptake inhibitors. Psychopharmacology (Berl) 1997; 133: 99-101.

22. Verkes R, Van der Mast R, Hengeveld M, Tuyl J, Zwinderman A, Van Kempen G. Reduction by paroxetine of suicidal behavior in patients with repeated suicide attempts but no major depression. American Journal of Psychiatry 1998; 155: 543-547.

23. Bräunig P, Pollentier F, Rao M. Increased serotonin after a suicide attempt: coincidence or catharsis? Biol Psychiatry 1988; 24: 725-727.

24. Linnoila M, Virkkunen M. Aggression, suicidality, and serotonin. J Clin Psychiatry 1992; 53: 46-51.

25. Kaye W, Gendall K, Strober M. Serotonin neuronal function and selective serotonin re-uptake inhibitor treatment in anorexia and bulimia nervosa. Biol Psychiatry 1998; 44: 825-838.

26. Brewerton TD. Toward a unified theory of serotonin dysregulation in eating and related disorders. Psychoneuroendocrinology 1995; 20: 561-590.

27. Kaye W, Gwirtsman H, George D, Ebert M. Altered serotonin activity in anorexia nervosa after long term weight restoration: Does elevated cerebrospinal fluid 5-hydroxyindolacetic acid level correlate with rigid and obsessive behaviour? Arch Gen Psychiatry 1991; 48: 556-562.

28. Brewerton T, Jimerson D. Studies of serotonin function in anorexia nervosa. Psychiatry Res 1996; 62: 31-42.

29. Collier D, Arranz M, Li T, Mupita D, Brown N, Treasure J. Association between the 5-HT2A gene promotor polymorphism and anorexia nervosa. Lancet 1997; 350: 412.

173

30. Ohara K, Nagai M, Tsukamoto T, Tani K, Suzuki Y, Ohara K. 5-HT2A receptor gene promoter polymorphism--1438G/A and mood disorders. Neuroreport 1998; 20: 1139-1141.

31. Campbell D, Sundaramurthy D, Markham A, Pieri L. Lack of association between 5-HT2A gene promoter polymorphism and susceptibility to anorexia nervosa. Lancet 1998; 351: 499.

32. Du L, Bakish D, Lapierre Y, Ravindran A, Hrdina P. Association of polymorphism of serotonin 2A receptor gene with suicidal ideation in major depressive disorder. Am J Med Genet 2000; 96: 56-60.

33. Bondy B, Kuznik J, Baghai T, Schule C, Zwanzger P, Minov C, de Jonge S, Rupprecht R, Meyer H, Engel R, Eisenmenger W, Ackenheil M. Lack of association of serotonin-2A receptor gene polymorphism (T102C) with suicidal ideation and suicide. Am J Med Genet 2000; 96: 831-835.

34. Van Laere K, Koole M, D'Asseler Y, Versijpt J, Dumont F, Audenaert K, Dierckx R. Automated stereotactic standardization of brain receptor data using single-photon transmission images. J Nucl Med 2001; 42: 361-375.

35. Lahorte P, Vandenberghe S, Van Laere K, Audenaert K, Lemahieu I, Dierckx R. Assessing the performance of SPM analyses of SPECT. Neuroimage 2000; 12: 757-764.

36. Audenaert K, Van Laere K, Brans B, Lahorte P, Koole M, Goethals I, van Heeringen C, Dierckx R. A feasibility study of the use of 99m-Tc ECD brain perfusion SPECT following a split dose paradigm in a neuropsychological activation task. Eight Biennal Congress of the South African Society of Nuclear Medicine, 1998, Cape Town.

37. Audenaert K, van Heeringen C, Van Laere K, Dumont F, Slegers G, Dierckx R. SPECT investigation with suicidal patients and controls. Max Planck Institut für Psychiatrie "Neurobiologie suizidalen Verhaltens", 2001, Günzburg.

175

Samenvatting van de thesis

Binnen het biologisch-psychiatrisch onderzoek is er een recente tendens om afstand

te nemen van het categorisch diagnostisch model en patiënten te evalueren over de

grenzen van nosologische entiteiten. De toepassing van deze transnosologische

benadering in het functioneel beeldvormend onderzoek in de psychiatrische

diagnostiek wordt beschreven in de inleiding van het proefschrift.

Een eerste luik van het proefschrift handelt over de evaluatie van prefrontale

neuropsychologische functies met SPECT. In de eerste twee hoofdstukken werd

respectievelijk de bruikbaarheid van prefrontale testen als neuropsychologische

activatietaken in een single day split-dose SPECT activatie paradigma in gezonde

vrijwilligers beschreven. In een eerste hoofdstuk werd de Stroop test geëvalueerd als

activatie- paradigma in gezonde vrijwilligers. Deze test evalueert de mate waarin een

subject zich kan beschermen tegen interferentie en inhibitie kan uitoefenen over zijn

gedrag. Er werden activatieregio' s aangetoond in de prefrontale cortex en in het

corpus cingulum anterior. In een tweede hoofdstuk wordt het gebruik van de Verbal

Fluency Test (COWAT; FAS) als neuropsychologische activatietaak beschreven bij

gezonde vrijwilligers. Deze test evalueert de initiatie van zoekstrategieën als

executieve functie. De subtests van de Verbal Fluency Test, de Letter Fluency en de

Category Fluency Test, hadden verschillende patronen qua prefrontale activatie.

Vanuit methodologisch oogpunt worden de resultaten van Statistical Parametric

Mapping (SPM) vergeleken met een conventionele Volumes-of-Interest (VOl)

benadering. Een derde hoofdstuk beschrijft de Verbal Fluency Test in de evaluatie

van depressieve suïcidale patiënten in vergelijking met normale vrijwilligers. De

depressieve suïcidale patiënten hadden een afgevlakte toename in de prefrontale

perfusie tijdens het uitvoeren van de activatietaak.

Een tweede luik van het proefschrift handelt over de evaluatie van de serotonine-2a

bindingsindex met de hoogselectieve radiojood gemerkte 5-HT2a receptor antagonist

1231-5-I-R91150 in psychiatrische patiëntenpopulaties. In het vierde hoofdstuk

wordt de 5- HT2a bindingsindex bepaald in patiënten met een zeer recente

zelfmoordpoging. Deze patiënten hadden een significant lagere bindingsindex in de

176

prefrontale schors in vergelijking met normale vrijwilligers. Deze lagere

bindingsindex was het meest uitgesproken in patiënten met zelfverwonding in

vergelijking met patiënten met auto-intoxicatie. In het vijfde hoofdstuk werden de

prefrontale 5-HT2a bindingsindex, de mate van hopeloosheid (Beck Hopelessness

Scale) en persoonlijkheidskarakteristieken (Cloninger's Temperament and Character

Inventory) in de zelfmoordpogers, die beschreven werden in het vierde hoofdstuk,

geëvalueerd. In vergelijking met normale vrijwilligers hadden zelfmoordpogers een

lagere prefrontale 5-HT2a bindingsindex, een hogere score inzake hopeloosheid, een

hogere score in Harm Avoidance als temperamentdimensie en lager scores op de

karakterdimensies Self Directedness en Cooperativeness. Er werd een significante

correlatie aangetoond tussen Harm Avoidance, hopeloosheid en de 5-HT2a

bindingsindex in zelfmoordpogers. In het zesde hoofdstuk werd de 5- HT2a

bindingsindex in patiënten met anorexia nervosa onderzocht. Anorexia nervosa

patiënten hadden een significant verminderde bindingsindex in de linker frontale

cortex, in de linker en rechter pariëtale cortex en in de rechter occipitale cortex.

In de epiloog worden de begrippen causaliteit, inverse causaliteit en het "state" of

"trait" karakter van de onderzoeksresultaten besproken en worden voorstellen naar

verder onderzoek in deze materie geformuleerd. Hierin werd ook een kritische

beschouwing van de methodologie beschreven.

177

Résumé de thèse

Récemment, dans la recherche de psychiatrie biologique, il y a une tendance de

prendre distance du modèle diagnostique de catégorisation selon DSM et d'évaluer

les patients en passant les frontières d'une classification nosologique. L'application

de cette approche transnosologique dans la recherche de l'imagerie fonctionnelle en

diagnostique en psychiatrie est Ie sujet de 1 'introduction de cette thèse.

Le premier volet de cette thèse concerne l'évaluation de fonctions

neuropsychologiques préfrontales. Les deux premiers chapitres évaluent la valeur des

testes préfrontales comme tâches d'activation dans un single day split-dose

paradigme dans une population de volontaires. Dans le premier chapitre le Stroop

Test est décrit. Ce test mesure l'exercice d'inhibition du comportement dans un

individu. Des régions d ' activation sont démontrées dans le cortex préfrontale et dans

le corpus cingulum anterior. Dans le deuxième chapitre Ie Verbal Fluency Test

(COWAT; FAS) comme tâche d'activation est évalué. Ce teste mesure l'initiation de

stratégie de cherche comme fonction exécutive. Les subtestes du Verbal Fluency

Test, Ie Letter Fluency et le Category Fluency Test, montrent des régions d'activation

préfrontales différentes. Du point de vue méthodologique les résultats de Statistical

Parametric Mapping (SPM) sont comparés avec une approche conventionnelle

d'analyse de Volumes-of-Interest (VOI). Un troisième chapitre décrit le Verbal

Fluency Test dans une évaluation des sujets dépressifs et suicidaires. Ces sujets

avaient une augmentation de perfusion préfrontale tempérée pendant l'exercice de

cette tâche d'activation.

Le deuxième volet de cette thèse concerne l'évaluation d'index de liaison du

récepteur serotonine-2a avec un antagoniste du récepteur de sérotonine-2a iodé

(1231-5-I-R9l150) dans des sujets souffrant d'une maladie psychiatrique. Dans le

quatrième chapitre, l'index de liaison de 5-HT2a est mesuré dans des sujets avec une

tentative de suicide très récente. Ces sujets avaient un index de liaison dans le cortex

préfrontal qui était diminué de façon significative en comparaison avec des sujets

normales. Dans le cinquième chapitre le rapport entre l'index de liaison 5- HT2a, le

niveau de désespérément (Beck Hopelessness Scale) et les characteristiques de

personalité (Cloninger's Temperament and Character Inventory) dans une population

178

de personnes avec une tentative de suicide est évalué. En comparaison avec des

volontaires normales, les patients avaient un index de liaison 5-HT2a dans le cortex

préfrontal qui était diminué, une score augmentée de Harm Avoidance et score

diminuée de Self Directedness et Cooperativeness. Une corrélation de façon

significative entre Harm Avoidance, désespérément et l'index de liaison 5-HT2a était

démontrée. Dans le sixième chapitre, l'index de liaison 5-HT2a dans les sujets qui

souffrent d'anorexia nervosa est évalué. Ces sujets avaient un index de liaison

diminué dans le cortex frontal gauche, le cortex pariétal gauche et droite et dans le

cortex occipital droite.

Dans l'épilogue, les concepts de causalité, inversion de causalité et le state-trait

caractère des résultats sont discutés et des perspectives d' avenir sont proposées.

179

Summary of the thesis

Recently, there is a tendency in biological psychiatry to evaluate patients across the borders of nosological entities and distance from the categorical diagnostic model.

The application of this transnosological approach in functional imaging research in

I psychiatry is described in the introduction of the thesis. A first part of the thesis

concerns the evaluation of prefrontal neuropsychological functions. The first two

chapters of the thesis discuss the feasibility of prefrontal tests as neuropsychological

activation tasks in a single day split dose SPECT activation paradigm in healthy

volunteers. In the first chapter the classical Stroop interference task as a prefrontal

neuropsychological activation probe in SPECT, was validated in healthy volunteers.

Activation regions were demonstrated in the prefrontal cortex and in the anterior

cingulate, which is in concordance with previous PET and fMRI findings during

Stroop-like interference tasks. In the second chapter the Verbal Fluency task, both in

its letter fluency and category fluency modality, as a prefrontal activation probe in

SPECT was validated in healthy volunteers. From a methodological point of view,

the SPM results were compared and found superior to a Voxel-of-interest (VOl)

approach. The letter fluency and the category fluency activation paradigm had a

differential brain activation pattern in the prefrontal cortex and showed both an

activation in the anterior cingulate cortex. In the third chapter the Verbal Fluency

task, as a prefrontal SPECT activation probe, was tested in depressed suicide-attempt

patients and results were compared to healthy volunteers. Depressed suicide

attempters showed a blunted increase in perfusion changes in the prefrontal cortex

that paralleled the hampered test results.

A second part of the thesis concerns the evaluation of the serotonin-2a binding index,

with the highly specific radiolabelled 5-HT2a receptor antagonist (1231-5-I-R9ll50)

in psychiatric patients. In the fourth chapter the brain serotonin-2a binding index in

patients who very recently attempted suicide, was estimated through the highly

specific radio-iodinated 5-HT2a receptor antagonist 1231-5-I-R91150 and compared

to healthy volunteers. Deliberately self-harming patients had a significantly reduced

mean frontal binding index after correction of age when compared with controls. The

reduction was more pronounced among deliberately self-injury patients than among

180

deliberately self-poisoning patients (DSP). In the fifth chapter, we investigated the

prefrontal 5-HT2a receptor-binding index, hopelessness, using Beck's Hopelessness

Scale, and personality characteristics, using Cloninger's Temperament and Character

Inventory , in suicide-attempt patients. When compared to normal controls, suicide-

attempt patients had a significantly lower binding index offtontal5-HT2a receptors, a

higher level of hopelessness, a higher score on the temperament dimension of harm

avoidance and lower scores on the character dimensions of self-directedness and

cooperativeness. A significant correlation was found between harm avoidance,

hopelessness and binding index in the deliberately self harm population. In the sixth

chapter the brain serotonin-2a binding index in patients with anorexia nervosa was

estimated through 123I-5-I-R91150 and compared to healthy volunteers. Patients

with Anorexia Nervosa, compared to healthy volunteers, have a significantly reduced

5-HT2a binding index in the left frontal cortex, in the left and right parietal cortex

and the left and right occipital cortex.

The epilogue of the thesis treats the items of causality, inverse causality and the state

or trait nature of the results and formulates recommendations for future research

concerning this matter.

Decreased frontal serotonin 5-HT 2a receptor binding index in deliberate self-harm patients

Documents

Transcript of Decreased frontal serotonin 5-HT 2a receptor binding index in deliberate self-harm patients