Reduced Parietal P300 Amplitude is Associated with anIncreased Risk for a First Psychotic EpisodeMirjam J. van Tricht, Dorien H. Nieman, Johannes H.T.M. Koelman, Johan N. van der Meer, Lo J. Bour,Lieuwe de Haan, and Don H. Linszen

Background: P300 abnormalities indicate changes in information processing and are one of the most reliable biological markers ofschizophrenia. We sought to investigate whether abnormalities in P300 (P3) or other event-related potentials are also present in subjects atultra high risk (UHR) for developing psychosis and whether they are helpful in predicting transition to psychosis.

Methods: The N1, N2, N2b, P2, and P3 amplitudes were assessed in 61 UHR subjects, of whom 18 subjects (30%) made a transition topsychosis over a 3-year follow-up period (UHR � T: age 20.4 years) and 43 (70%) did not (UHR � NT: age 19.3 years), and 28 age- andintelligence-matched healthy control subjects (age 20.0 years). Psychopathology was also assessed.

Results: The UHR � T subjects showed smaller parietal P3 amplitudes, compared with control subjects and UHR � NT subjects. Moreover,the N2b was higher in control subjects compared with both UHR groups. We found no differences in N1 or P2 components between thegroups, and our UHR subjects did not exhibit bilateral P3 asymmetry. Reduced P3 amplitudes were the best predictor for subsequentpsychosis in the UHR group. The P3 reduction was related to increased social anhedonia and withdrawal and a lower global assessment ofsocial functioning and social personal adjustment.

Conclusions: The UHR � T subjects showed reduced parietal P3 amplitudes. In addition, a reduced P3 amplitude was the best predictor forsubsequent psychosis. If replicated, these findings might contribute to a more accurate prediction of a first psychotic episode. Furthermore,

reduced social functioning might be related to information processing deficits in UHR subjects.

omaiPaih(m

Uptsaf

hfpai(aptm

prUte(

Key Words: P300, psychopathology, psychosis prediction, socialfunctioning, ultra high risk subjects

Schizophrenia and other psychotic disorders are usuallypreceded by a prodromal phase, in which a change inpremorbid functioning occurs (1,2). Many attempts have

been made to identify factors in this prodromal phase that predicttransition to psychosis. The term ultra high risk (UHR) indicatesa syndrome that is regarded as a risk factor for subsequentpsychosis within the next year (3,4). UHR subjects are thought tobe at high risk for a first psychotic episode on the basis of one ofthe following clinical symptoms: 1) genetic risk in combinationwith reduced functioning; 2) Attenuated Positive Symptoms, and3) Brief Limited Intermittent Psychotic Symptoms (BLIPS). How-ever, psychosis is not inevitable. Despite an acceptable predic-tive validity of these UHR criteria (5–7), there are still many falsepositives. Therefore, a need to identify signs within this group ofan impending first psychotic episode still exists. Studies in UHRsubjects have already revealed abnormalities on neuroimaging(8–10) and neuropsychological (11,12) examinations before theonset of a first psychotic episode, indicating a pre-existingvulnerability to psychosis.

Event related potentials (ERP) have been frequently used inthe search for biological markers of schizophrenia and psychosis(13–16). The P3 has been proposed to be one of the most reliablebiological markers of schizophrenia (13,17). The P3 is a scalp-recorded late ERP, occurring approximately 300 msec after anattended unusual or task-relevant stimulus. Two subcomponents

From the Departments of Psychiatry (MJvT, DHN, LdH, DHL) and Neurologyand Clinical Neurophysiology (MJvT, JHTMK, JNvdM, LJB), AcademicMedical Center, University of Amsterdam, Amsterdam, the Netherlands.

Address correspondence to Mirjam J. van Tricht, M.Sc., Academic MedicalCenter, Department of Neurology D2-136, PO Box 22660, 1100 DD Am-sterdam, the Netherlands; E-mail: m.j.vantricht@amc.uva.nl.

sReceived Nov 8, 2009; revised Apr 14, 2010; accepted Apr 14, 2010.

0006-3223/$36.00doi:10.1016/j.biopsych.2010.04.022

f the P3 can be distinguished (18). The P3a potential, which isaximal at frontal electrode positions, is thought to reflect

utomatic initial processing when a novel or distracting stimuluss detected. This component is followed by a parietal maximal3b, which is a proportional reflection of subsequent effortfulttentional resources, associated with updating task-relevantnformation in memory (19–21). Both P3a and P3b abnormalitiesave been reported in schizophrenia patients and their relatives19,22–24), albeit the P3b component has received significantlyore attention.Recent studies have demonstrated smaller P3 amplitudes in

HR subjects (15,25–27). These studies have consistently re-orted P3 amplitude reductions at midline scalp sites. In addi-ion, one of these studies reported abnormalities in P3 left–rightymmetry in UHR subjects (25). Other studies investigating P3symmetry in UHR subjects, however, have not replicated thisinding (15,26,27).

Abnormalities in N1 (28), P2 (29,30), and N2 (31) componentsave also been reported in schizophrenia patients, althoughindings are contradictory (32,33). The N1 is presumed to reflectrimary and secondary auditory functions as well as alertnessnd motivation (28), whereas the N2 indexes cognitive processesnvolved in the categorization of deviant stimuli (34). The N2bi.e., N2 difference as reflected by target-minus-frequent N2mplitude) is also thought to be affected in schizophreniaatients (34). To our knowledge, no reports are yet available onhe predictive value of these ERP components for the develop-ent of a first psychotic episode.The present study focused on the predictive value of ERP

arameters for transition to psychosis. Our longitudinal studyevealed whether ERP abnormalities were also present in ourHR group and whether baseline differences existed between

he UHR groups with and without transition to psychosis. WexpectedUHR subjectswhomade a transition to psychosis at followupUHR � T) to show smaller ERP amplitudes compared with UHR

ubjects without transition (UHR � NT) and healthy control

BIOL PSYCHIATRY 2010;68:642–648© 2010 Society of Biological Psychiatry

St

cSshHNei

C

gdpscgTit

vafn

M

ocwahtasrcftg

; UHR

M.J. van Tricht et al. BIOL PSYCHIATRY 2010;68:642–648 643

subjects. Following the literature, we hypothesized that theamplitude reductions were most pronounced for the P3 compo-nent. The relation between ERP parameters and UHR symptomswas evaluated on an exploratory basis.

Methods and Materials

ParticipantsUHR Group. Sixty-one subjects (19 women) with a UHR for

developing psychosis were included (Table 1). The UHR subjectswere referred to the Academic Medical Center (AMC), Amster-dam, the Netherlands, by professionals from mental healthservices for a second opinion with the question of whether apsychotic development was taking place. The subjects wereexamined within the Dutch Prediction of Psychosis Study of theAMC. The inclusion criteria for the UHR group were: agebetween 15 and 35 years, and belong to one or more of thefollowing three groups (as described in Yung et al. [3] andMcGorry et al. [4]):

Genetic Risk in Combination with Reduced Functioning:subjects who have a first-degree relative with a psychoticdisorder or who themselves have a schizotypal personalitydisorder and who have experienced a significant decrease infunctioning during the past year (i.e., 30% reduction ofGlobal Assessment of Functioning-score for at least 1month).

Attenuated Positive Symptoms: subjects who have experi-enced subthreshold, attenuated positive psychotic symp-toms, defined by at least one of the following symptoms,appearing several times/week for at least 1 week within thelast 3 months: unusual thought content/delusional ideas,suspiciousness/persecutory ideas, grandiosity, perceptualabnormalities/hallucinations, disorganized communication,and odd behavior/appearance.

BLIPS: subjects who have experienced episodes of frankpsychotic symptoms. BLIPS were defined by hallucinations,delusions, or formal thought disorders occurring within thelast 3 months and resolving spontaneously within 1 week.

ome of the UHR subjects were included on the basis of more

Table 1. Sociodemographic Characteristics of UHR and Control Subjects at

Mean Age (SD)Gender (male/female)Medication Usage (%)

AntipsychoticsAntidepressantsOtherNone

NART IQ Score (SD)Inclusion Symptoms, n (%)Attenuated SymptomsBLIPSBLIPS and Attenuated SymptomsGenetic Risk and Reduced FunctioningGenetic Risk, Reduced Functioning and Attenuated Symptoms

UHR, ultra high risk; UHR � T, UHR subjects with transition to psychosisReading Test (41); BLIPS, Brief Limited Intermittent Psychotic Symptoms.

han one criterion. The exclusion criteria were: previous psy- s

hotic episode for more than 1 week, as assessed with thetructured Clinical Interview for Diagnosis, Sections B and C (35);ymptoms due to substance abuse, as assessed with the Compre-ensive International Diagnostic Interview, Sections J and L; Worldeath Organization (36) IQ below 85, as assessed with the Dutchational Adult Reading Test (37); vision disorders; endocrine dis-ase; and known neuropsychological impairment (e.g., closed headnjury). All participants gave written consent.

ontrol Group

Twenty-eight participants (15 women) served as a controlroup for ERP performance. Exclusion criteria were: visualisorders, endocrine disease, psychiatric illness present or in theast, familial history of psychiatric illness (evaluated for first- andecond-degree relatives), known neurological impairment (e.g.,losed head injury), and IQ below 85. They were matched as aroup on age and estimated premorbid IQ of the UHR subjects.he premorbid IQ was estimated with the National Adult Read-

ng Test. The control subjects received a fee for their participa-ion of €40.

The investigation was carried out in accordance with the latestersion of the Declaration of Helsinki. The study design waspproved by the Medical Ethical Committee of the AMC. In-ormed consent of all the participants was obtained after theature of the procedures had been fully explained.

aterialsERP Recording. ERPs were assessed with an active auditory-

ddball paradigm. The subjects were seated in a comfortablehair with eyes open, in a dimly lit, quiet room. Target stimuliith a frequency of 2000 Hz and standard, nontarget stimuli withfrequency of 1000 Hz were presented binaurally through

eadphones at an intensity of 50 dB above hearing threshold. Aotal of 300 tones were presented, of which 80% were nontargetsnd 20% were targets, with a duration of 100 msec in a randomequence. The subjects were instructed to count the targets andespond to them with a button press. The total number ofounted targets was asked at the end of each session. Toamiliarize the subjects with the task, three practice trials witharget and nontarget stimuli were presented. The stimuli wereenerated with an interstimulus interval of 1480 ms (i.e., a

line

� T (n � 18) UHR � NT (n � 43)Control Subjects

(n � 28)

20.4 (4.0) 19.3 (3.7) 20.0 (3.3)13/5 29/14 13/15

7 (39) 9 (21) —2 (11) 7 (16) —1 (6) 5 (12) —8 (44) 22 (51) 28 (100)

04.4 (8.4) 101.1 (9.5) 106.9 (10.4)12 (67) 37 (86)

3 (17) 0 (0)1 (6) 4 (9)1 (6) 1 (2) —1 (6) 1 (2)

� NT, UHR subjects without transition to psychosis; NART, National Adult

Base

UHR

1

timulation frequency of .67 Hz).

www.sobp.org/journal

Sdpitdsoosbat(y1

tf

M

ccsaaem

P

fsffDpfdp

insd

S

tDwrdBetpEgot

644 BIOL PSYCHIATRY 2010;68:642–648 M.J. van Tricht et al.

Twenty-one silver–silver chloride disk electrodes were at-tached to electrode sites according to the international 10–20system, with a reference electrode on linked mastoids and aground electrode on the forehead. Additionally, four electrodeswere attached at the outer canthi of both eyes and above andbelow the left eye for the registration of eye movements andblinks. The interelectrode impedances were always �5 k� ineach subject investigated. The electroencephalograph was re-corded with a band-pass filter of .04–300 Hz, with a samplingrate of 1000 Hz. Digitized data for each subject were stored in adatabase for subsequent offline analysis with Brainvision Ana-lyzer (Brainproducts; http://www.brainproducts.com). Afterbaseline correction, the signals were digitally filtered with alow-pass filter of 30 Hz and a high-pass filter of .10 Hz (24dB/oct) and were epoched at 50 msec before stimulus and 450msec after stimulus. The maximum allowed absolute differencebetween two values in one segment was 200 �V, and themaximum allowed voltage step was 50 �V. Segments in whichthese values were exceeded were removed. Both vertical andhorizontal eye-movements were detected and removed witheye-movement detection measures developed by Gratton et al.(38). Epochs were averaged separately for nontarget and targettones. If the number of artifact free trials was below 26, therecording was excluded from further analyses.

Peak amplitudes were semiautomatically detected and calcu-lated relative to pre-stimulus baseline of 50 msec (39). The N1and P2 were measured from averages elicited by nontarget tones.The N1 amplitudes were detected as the most negative pointbetween 75 and 125 msec after stimulus (28,40–43). The P2amplitudes were detected as the most positive point after the N1,with a latency range of 150–220 msec. The N2 and P3 compo-nents were calculated as waveforms generated by target tones.The N2 was scored within a timeframe of 180–320 msec afterstimulus, whereas the P3 was defined as a the largest positivevalue between 250 and 450 msec after stimulus. The N2bdifference component was calculated by subtracting the mostnegative point after the nontarget stimulus from the most nega-tive point after the target stimulus within the N2 time frame. TheN1 and N2 components were assessed at central midline (Cz)scalp site, whereas P2 and P3 components were assessed atparietal (Pz) scalp site. Peak amplitudes and peak latencies werecalculated with a computer algorithm. All peaks were visuallyinspected. If necessary, adjustments were made hereafter.

Laterality

To investigate possible laterality differences between the twogroups, P3 waves were also assessed at T3 and T4 scalppositions. Following previous studies (44,45) lateral analyseswere conducted on normalized peak data. Each lateral voltagewas normalized by dividing the lateral voltage by the meangroup Cz amplitude. Asymmetry differences between the groupswere investigated with ratios (amplitude T3/amplitude T4) anddifference scores (amplitude T4 � amplitude T3).

Task Performance

Reaction times and the number of correct hits were calculatedin all groups. Trials with no button press were excluded.

InstrumentsAll UHR subjects were assessed with the Structured Interview

for Prodromal Syndromes (SIPS) (46), the Positive and Negative

Syndrome Scale (PANSS) (47), and the Premorbid Assessment s

www.sobp.org/journal

cale (PAS) (48). The SIPS is a comprehensive diagnostic toolesigned specifically for the assessment of the whole spectrum ofrodromal signs and symptoms. The scale is composed of 19

tems (5 positive, 6 negative, 4 disorganization, 4 general symp-oms), each of which is given a score of zero to six according toefined criteria. A score between 3 and 5 on the positiveymptoms indicates attenuated psychotic symptoms, and a scoref 6 indicates a psychotic state or BLIPS. The PANSS is composedf 30 items, each of which is given a score of one to seven. Subcores for positive, negative, and disorganization symptoms cane calculated (49,50). The PAS is a 28-item rating scale thatssesses sociability and withdrawal, peer relationships, adapta-ion to school, and scholastic performance in four life stageschildhood, 11 years and younger; early adolescence, 12–15ears; late adolescence, 16–18 years; and adulthood, older than8 years) as well as social aspects of life after age 15.

Control subjects were also interviewed with the SIPS. Whenheir scores surpassed the UHR threshold, they were excludedrom the study.

edicationFor the analyses, all UHR subjects were divided into four

ategories: 1) antipsychotic medication (patients with antipsy-hotic medication and antidepressants and/or other medication,uch as benzodiazepines, were also assigned to this category); 2)ntidepressants or antidepressants with medication other thanntipsychotic medication; 3) other, for instance, benzodiaz-pines, methylphenidate, and/or lithium carbonate; and 4) noedication (Table 1).

rocedureIn this longitudinal cohort study, subjects were followed up

or 3 years. A transition to psychosis was operationalized as acore of 4 or more on PANSS for hallucinations, delusions, orormal thought disorders for longer than 1 week. To establish aormal DSM-IV diagnosis, the Structured Clinical Interview foriagnosis was administered to all subjects after transition tosychosis. After transition to psychosis, the subjects received theollowing diagnoses: schizophrenia (n � 12), schizophreniformisorder (n � 3), schizoaffective disorder (n � 2), and briefsychotic disorder (n � 1).

In all participants, ERPs were assessed in a session of approx-mately 20 min, within 1 week after the psychiatric question-aires were administered. All participants were instructed not tomoke, use drugs, or drink coffee or other caffeinated drinks theay before the ERP recording.

tatistical AnalysisThe ERP amplitudes of the three groups were compared by

wo-tailed t tests, after separate multivariate analyses of variance.ifferences between the UHR groups in psychiatric symptomsere evaluated with Mann–Whitney U test. A multivariate algo-

ithm that optimizes prediction of conversion to psychosis waserived from a Cox proportional hazards model, with theackward Stepwise method. This model was used to estimate theffects of the covariates on survival time. In this model, transitiono psychosis was entered as the status variable, time to onset ofsychosis or to follow-up was entered as the time variable, andRP and clinical variables that differed significantly between theroups were entered as covariates. Given the small group sizes,nly subscales of the psychiatric questionnaires were included inhe Cox Model. In addition to the hazard ratio, sensitivity and

pecificity were calculated for the covariates that contributed

smb(tT(

wamt[fjp

(yad1tFc

E

(m

s

Tbd.scst.Upce.lpnUb

tS

P

td

P

f(u

P

dsUssfta(

t

M.J. van Tricht et al. BIOL PSYCHIATRY 2010;68:642–648 645

significantly to the prediction of the event. Survival curves wereobtained with Kaplan Meier survival analyses. Correlations be-tween ERP parameters and psychiatric symptoms were examinedwith nonparametric correlation tests (Spearman’s �). Differencesbetween groups in medication use and gender were analyzedwith �2 tests. For all tests, p values of � .05 were accepted asignificant. We did not correct the level of significance forultiple comparisons to reduce the probability of Type I error,ecause in these exploratory correlation analyses a Type II errorfailure to detect an effect when it actually exists) is more serioushan Type I error (considering an effect to be real when it is not).he data were analyzed with a statistical computer programSPSS 17.0; SPSS, Chicago, Illinois).

Results

Demographic and Clinical CharacteristicsAs shown in Table 1, 61 UHR subjects and 28 control subjects

ere included at baseline. Of the UHR subjects, 18 (30%) madetransition to psychosis over a 3-year follow-up period. Theean interval between inclusion and transition to psychosis in

he UHR � T group was 9.4 months (range � 2–25 months, meanSD] � 7.2). Three of the UHR subjects (5%) were lost toollow-up. These subjects did not differ significantly from sub-ects with follow-up information in terms of demographic or ERParameters at baseline.

There were no differences between groups with respect to ageF � .65, p � .53). The mean age of the UHR � T subjects was 20.4ears, whereas UHR � NT and healthy control subjects were 19.3nd 20.0 years of age, respectively. In addition, the groups did notiffer in terms of gender (�2 � 4.2, p � .12) or premorbid IQ (F �.74, p � .11). There were no differences between UHR groups inhe amount of received treatment at baseline (F � .98, p � .46).inally, we found no differences between UHR groups in inclusionriteria at baseline (�2 � 6.8, p � .554).

RPThe mean number of usable trials for the target tone was 47.4

UHR � T: mean � 46.0, SD � 7.6, range � 29–57; UHR � NT:ean � 46.4, SD � 10.1, range � 27–59; control subjects: mean �

48.6, SD � 9.3, range � 29–60). The mean of usable trials for thetandard tone was 189.2 (UHR � T: mean � 183.9, SD � 36.3,

range 103–236; UHR � NT: mean � 188.2, SD � 40.3, range �94–238; control subjects: mean � 193.9, SD � 33.6, range �122–241). The groups did not differ significantly on the numberof usable trials for the target tone (F � .58, p � .56) or the

Table 2. ERP Parameters of UHR and Control Subjects at Baseline

UHR � T (n � 18)

Amplitude N100 Cz �6.9 (2.8)Amplitude P200 Pz 3.7 (3.3)Amplitude N200 Target Tone Cz �.25 (6.7)Amplitude N200 Standard Tone Cz �.97 (2.9)N2b Component Cz .72 (5.9)Amplitude P300 Pz 10.5 (2.3)Hit Rate (%) 99.4Reaction Times (sec) .43 (.1)

Values are mean (SD). The N100 and P200 amplitudes were calculated frones. The N2b component was calculated by subtracting the N200 amplitu

ERP, event-related potentials; other abbreviations as in Table 1.ap � 0.03.bp � .0001.

nontarget tone (F � .41, p � .66). a

ERP amplitudes, hit rates, and reaction times are presented inable 2. There were no differences in reaction times and hit ratesetween the groups. Baseline analyses revealed a significantifference between the groups in P3 amplitude (F � 22.2, p �0001). Subsequent t tests for independent samples showedignificantly smaller P3 amplitudes of UHR � NT subjectsompared with control subjects (t � 3.99, p � .001). An eventronger effect was found when comparing P3 amplitudes be-ween UHR � T subjects and control subjects (t � 6.73, p �0001). The P3 amplitudes were also significantly smaller inHR � T subjects, compared with UHR � NT subjects (t � 4.16,� .001). We found no group differences in the N1, P2, or N2

omponents. However, our analyses did reveal significant differ-nces in the N2b component between groups (F � 3.66, p �030). Subsequent t test showed that the N2b difference wasarger in control subjects compared with both UHR � T (t � 2.32,� .025) and UHR � NT (t � 2.44, p � .017) subjects. We foundo differences in the N2b component between UHR � T andHR � NT subjects. We found no differences in ERP measuresetween the distinct inclusion groups.

Grand average ERP of the three groups after target and nontargetones are presented in Figures 1A and 1B, respectively. See Figure1 in Supplement 1 for the topographic maps of P3 distribution.

3 LateralityWe found no baseline differences in P3 normalized ampli-

udes at T3 or T4 or in the T3/T4 ratio scores and T4 � T3ifference scores between the groups (Table 3).

3 and Medication UseWe found no significant differences in medication use in the

our categories (as presented in Table 1) between UHR groups�2 � 1.9, p � .75). Nor did we find any associations between these of medication and ERP amplitudes.

3 and Psychiatric SymptomsSee Table S1 in Supplement 1 for the SIPS, PANSS, and GAF

ata of the three groups. The score on the subscale negativeymptoms of the SIPS was higher in UHR � T compared withHR � NT subjects (U � 251.0, p � .031). In addition, UHR � Tubjects showed a trend toward a higher score on the negativeymptoms subscale of the PANSS (U � 257.0, p � .063). Weound no differences between the groups on the positive symp-om scales. Reduced P3 amplitude was related to increased socialnhedonia and withdrawal as assessed with Item N1 of the SIPS� � �.313, p � .014) and a higher rate of active social avoidance

UHR � NT (n � 43) Control Subjects (n � 28)

�6.8 (2.8) �7.8 (4.0)3.3 (2.7) 4.5 (2.6)

�.57 (6.1) �2.1 (5.8)�.80 (2.7) 1.1 (2.7)

.23 (6.1) 3.2 (5.4)a

16.5 (6.0) 22.9 (7.6)b

99.7 99.2.42 (.1) .40 (.1)

andard tones. The N200 and P300 amplitudes were calculated from targetter standard (low) tones from the N200 amplitude after target (high) tones.

om stde af

s assessed with Item G16 of the PANSS (� � �.372, p � .003).

www.sobp.org/journal

rFhy“

E

Ptat1

iSisg[sceotra

D

tcafs

Ft

TS

TTTT

C

arw

646 BIOL PSYCHIATRY 2010;68:642–648 M.J. van Tricht et al.

In addition, UHR subjects with smaller P3 amplitudes scoredhigher on the global assessment of the highest level of function-ing of the PAS (� � �.395, p � .002), where the highest scoreeflects “unable to function in or enjoy any aspect of life.”urthermore, UHR subjects with smaller P3 amplitudes scoredigher at the item “social personal adjustment after the age of 15ears” (� � � .267, p � .045), where the highest score reflectsno desire to be with peers or others. Either asocial or antisocial.”

ffect of ERP and Clinical Variables on Survival TimeThe following variables were entered in the final Cox model:

3 amplitude, N2b score, and negative symptoms subscales ofhe PANSS and the SIPS. With the Cox regression survivalnalysis, the strongest predictor of developing a psychosis washe P3 amplitude (B � �.23, Wald � 10.4, p � .001, hazard ratio �

Figure 1. (A) Grand average target waveforms for each group at parietal(Pz). (B) Grand average nontarget waveforms for each group at Pz. (C) Grand

verage nontarget waveforms for each group at central midline. Ultra highisk subjects with transition to psychosis � red lines. Ultra high risk subjectsithout transition � green lines. Control group � dashed gray lines.

.37). This means that the relative risk of developing a psychosisda

www.sobp.org/journal

ncreases 37% with a decrease of the P3 amplitude of 1 �V (Table2 in Supplement 1). None of the other covariates contributedndependently to the prediction of a first psychotic episode. Theurvival distribution (Figure 2) differed significantly between theroups with above and below UHR mean amplitudes (log rankMantel Cox] � 10.48, p � .001). We also attempted to assess theensitivity and specificity of the P3 amplitude. With the defaultut value, the sensitivity and specificity of the P3 amplitudes wasstimated as 83.3% and 79.1%, respectively. For the aim ofbtaining a high sensitivity (thereby avoiding false negatives),he cut value for predicted group membership was set at .24,esulting in a sensitivity of 100% and a specificity of 69.8% for P3mplitudes.

iscussion

Our study shows that UHR subjects who eventually made aransition to psychosis have smaller parietal P3 amplitudesompared with subjects who did not make this transition. Inddition, P3 amplitudes were the best predictor of a developingirst psychotic episode in our UHR group. As such, P3 amplitudeseem to be a neurobiological predictor of future transition to

igure 2. Survival curves for ultra high risk (UHR) subjects with P3 ampli-udes above or below UHR group mean (14.7). The survival distribution

able 3. P3 Laterality on T3 and T4 Scalp Positions of UHR and Controlubjects at Baseline

UHR � T(n � 18)

UHR � NT(n � 43)

Control Subjects(n � 28)

3 Amplitudea .69 (.3) .57 (.3) .46 (.1)4 Amplitudea .67 (.2) .50 (.2) .47 (.2)3/T4 Ratio 1.03 (.3) 1.14 (.5) 1.02 (.8)4 � T3 Difference �.02 (.3) �.07 (.5) .01 (.4)

SD, Values are mean (SD). Abbreviations as in Table 1.aNormalized peak data: lateral voltage (�V) divided by the mean Group

z amplitude.

iffered significantly between the groups with above or below UHR meanmplitude (log rank [mantel Cox] � 10.48, p � .001).

i(b

pm

tehtg#B(Cctt

p

o

1

1

1

1

1

1

1

M.J. van Tricht et al. BIOL PSYCHIATRY 2010;68:642–648 647

psychosis. Our findings concur with other studies in high-riskpopulations, reporting midline P3 amplitudes reductions (15,25–27). We are the first, however, to report on the predictive valueof P3 abnormalities.

We found no differences between the groups in N1 or P2components. We did find a higher N2 difference score (N2b) incontrol subjects. The N2b component is used to index targetdetection and stimulus classification and is elicited by controlleddeviance detection (51,52). This component, however, did notdiffer between the UHR groups. Thus, although both UHRgroups showed impaired deviance detection, the N2b compo-nent did not attribute to predicting psychosis.

We found no evidence of P3 amplitude asymmetry differ-ences between the groups. That our paradigm involved a buttonpress, which is thought to diminish the abnormal P3 asymmetryin schizophrenia, could explain the absence of asymmetrydifferences (26,44,53). However, our results replicate the find-ngs of other studies on P3 asymmetry in UHR subjects13,24,25). Given that the laterality changes have repeatedlyeen found in later stages of the disease (17), it is likely that these

changes represent a marker of illness progression rather than amarker of susceptibility to schizophrenia. Replication studies areneeded before any firm conclusions can be drawn.

As for the relations among P3, psychiatric symptoms, andsocial functions, some potential meaningful associations werefound. Smaller P3 amplitudes were associated with a higherrating of social anhedonia and withdrawal as assessed with theSIPS and the PANSS. Moreover, reduced P3 amplitudes wererelated to a greater inability to function in or enjoy any aspect oflife and a lower level of social personal adjustment as assessedwith the PAS. Although these associations were modest, theypoint in the same direction, thus suggesting a relationshipbetween P3 amplitudes and a decline in social functioning indaily life, which is a characteristic phenomenon in the prodromalphase of schizophrenia. Similar associations between ERPs, inparticular Mismatch Negativity and the amplitude of the N2bcomponent, and daily social functioning have been reported inschizophrenia patients (54,55). Our analysis failed to showassociations between P3 parameters and positive UHR symp-toms. This might be due to a more modest symptom presentationin UHR subjects compared with schizophrenia patients.

In the present study, three subjects were lost to follow-up.Therefore, we cannot be certain about their current state offunctioning. A transition to psychosis is highly unlikely, becausewe asked the practitioner who referred the subjects to contact uswhen they suspected a psychosis. Hence, we do not believe thatwe have incorrectly labeled these subjects as UHR � NT. Finally,these subjects did not differ from the other UHR subjects on anyof the clinical or neurophysiological measures at baseline. It istherefore very unlike that these cases have influenced our results.

There is converging evidence that early intervention mightprevent or delay the onset of psychosis (4,56,57). Our resultssuggest that P3 amplitudes might be helpful in improvingprediction and therefore preventing transition to psychosis. Ourforthcoming report will reveal whether P3 abnormalities show aprogression after conversion to psychosis, as suggested by recentneuroimaging findings (58).

We conclude that smaller parietal P3 amplitudes might behelpful in identifying an impending psychotic episode. Clinicalvariables, such as social withdrawal might also be a helpfulmeasure, although these variables did not independently con-tribute to the prediction of a first psychotic episode in our study.

If replicated, our findings might contribute to a more accurate

rediction of a first psychotic episode in addition to othereasures, such as neuroimaging and neurocognition.

The authors thank the electroencephalography technicians ofhe Clinical Neurophysiology Unit for their assistance in thevent-related potentials recordings. We also thank T. Boeree foris technical assistance. This study was supported by a grant for

he Dutch Prediction of Psychosis Study from ZON-MW (Zor-Onderzoek Nederland/NWO-Medische Wetenschappen, project2630.0001) and a grant from the European Commission in

russels, Belgium, for the European Prediction of Psychosis studyGrant QLGU-CT-2001-01081). The ZON-MW and the Europeanommission had no further role in the study design; in theollection, analysis, and interpretation of data; in the writing ofhe report; and in the decision to submit the paper for publica-ion.

The authors reported no biomedical financial interests orotential conflicts of interest.

Supplementary material cited in this article is availablenline.

1. Yung AR, Phillips LJ, Yuen HP, Francey SM, McFarlane CA, Hallgren M,McGorry PD (2003): Psychosis prediction: 12-Month follow up of a high-risk (“prodromal”) group. Schizophr Res 60:21–32.

2. Nieman D, Becker H, van de Fliert R, Plat N, Bour L, Koelman H, et al.(2007): Antisaccade task performance in patients at ultra high risk fordeveloping psychosis. Schizophr Res 95:54 – 60.

3. Yung AR, Yuen HP, Berger G, Francey S, Hung TC, Nelson B, et al. (2007):Declining transition rate in ultra high risk (prodromal) Services: Dilutionor reduction of risk? Schizophr Bull 33:673– 681.

4. McGorry PD, Yung AR, Phillips LJ (2003): The “close-in” or ultra high-riskmodel: A safe and effective strategy for research and clinical interven-tion in prepsychotic mental disorder. Schizophr Bull 29:771–790.

5. Yung AR, Stanford C, Cosgrave E, Killackey E, Phillips L, Nelson B,McGorry PD (2006): Testing the ultra high risk (prodromal) criteria forthe prediction of psychosis in a clinical sample of young people. Schizo-phr Res 84:57– 66.

6. Woods SW, Addington J, Cadenhead KS, Cannon TD, Cornblatt BA,Heinssen R, et al. (2009): Validity of the prodromal risk Syndrome for firstpsychosis: Findings from the North American Prodrome LongitudinalStudy. Schizophr Bull 35:894 –908.

7. Cannon TD, Cornblatt B, McGorry P (2007): Editor’s introduction: Theempirical status of the ultra high-risk (prodromal) research paradigm.Schizophr Bull 33:661– 664.

8. Takahashi T, Wood SJ, Yung AR, Phillips LJ, Soulsby B, McGorry PD, et al.(2009): Insular cortex gray matter changes in individuals at ultra-high-risk of developing psychosis. Schizophr Res 111:94 –102.

9. Job DE, Whalley HC, McConnell S, Glabus M, Johnstone EC, Lawrie SM(2003): Voxel-based morphometry of grey matter densities in subjectsat high risk of schizophrenia. Schizophr Res 64:1–13.

0. Pantelis C, Velakoulis D, Wood SJ, Yücel M, Yung AR, Phillips LJ, et al.(2007): Neuroimaging and emerging psychotic disorders: The Mel-bourne ultra-high risk studies. Int Rev Psychiatry 19:371–379.

1. Niendam TA, Bearden CE, Zinberg J, Johnson JK, O’Brien M, Cannon TD(2007): The course of neurocognition and social functioning in individ-uals at ultra high risk for psychosis. Schizophr Bull 33:772–781.

2. Brewer WJ, Wood SJ, Phillips LJ, Francey SM, Pantelis C, Yung AR, et al.(2006): Generalized and specific cognitive performance in clinical high-risk cohorts: A review highlighting potential vulnerability markers forpsychosis. Schizophr Bull 32:538 –555.

3. Bramon E, Rabe-Hesketh S, Sham P, Murray RM, Frangou S (2004): Meta-analy-sis of the P300 and P50 waveforms in schizophrenia. Schizophr Res 70:315–329.

4. Jeon Y, Polich J (2003): Meta-analysis of P300 and schizophrenia: Patients, par-adigms, and practical implications. Psychophysiology 40:684–701.

5. van der Stelt O, Lieberman JA, Belger A (2005): Auditory P300 in high-risk, recent-onset and chronic schizophrenia. Schizophr Res 77:309 –320.

6. Shin KS, Kim JS, Kang DH, Koh Y, Choi JS, O’Donnell BF, et al. (2009):

Pre-attentive auditory processing in ultra-high-risk for schizophreniawith magnetoencephalography. Biol Psychiatry 65:1071–1078.

www.sobp.org/journal

3

3

4

4

4

4

4

4

4

4

4

4

5

5

5

5

5

5

5

5

5

648 BIOL PSYCHIATRY 2010;68:642–648 M.J. van Tricht et al.

17. Jeon YW, Polich J (2001): P300 asymmetry in schizophrenia: A meta-analysis. Psychiatry Res 104:61–74.

18. Squires NK, Squires KC, Hillyard SA (1975): Two varieties of long-latencypositive waves evoked by unpredictable auditory stimuli in man. Elec-troencephalogr Clin Neurophysiol 38:387– 401.

19. Demiralp T, Ucok A, Devrim M, Isoglu-Alkac U, Tecer A, Polich J (2002):N2 and P3 components of event-related potential in first-episodeschizophrenic patients: Scalp topography, medication, and latency ef-fects. Psychiatry Res 111:167–179.

20. Donchin E, Coles MGH (1988): Is the P300 component a manifestation ofcontext updating. Behav Brain Sci 11:357–374.

21. Kok A (2001): On the utility of P3 amplitude as a measure of processingcapacity. Psychophysiology 38:557–577.

22. Turetsky BI, Cannon TD, Gur RE (2000): P300 subcomponent abnormal-ities in schizophrenia. III. Deficits in unaffected siblings of schizophrenicprobands. Biol Psychiatry 47:380 –390.

23. Jeon YW, Polich W (2008): Meta-analysis of P300 and schizophrenia: Patients,paradigms, and practical implications. Psychophysiology 40:684–701.

24. Bramon E, McDonald C, Croft RJ, Landau S, Filbey F, Gruzelier JH, et al.(2005): Is the P300 wave an endophenotype for schizophrenia? A meta-analysis and a family study. Neuroimage 27:960 –968.

25. Frommann I, Brinkmeyer J, Ruhrmann S, Hack E, Brockhaus-Dumke A,Bechdolf A (2008): Auditory P300 in individuals clinically at risk forpsychosis. Int J Psychophysiol 70:192–205.

26. Bramon E, Shaikh M, Broome M, Lappin J, Bergé D, Day F, et al. (2008):Abnormal P300 in people with high risk of developing psychosis. Neu-roimage 41:553–560.

27. Ozgürdal S, Gudlowski Y, Witthaus H, Kawohl W, Uhl I, Hauser M (2008):Reduction of auditory event-related P300 amplitude in subjects withat-risk mental state for schizophrenia. Schizoph Resour 105:272–278.

28. Potts GF, Hirayasu Y, O’Donnell BF, Shenton ME, McCarley RW (1998):High-density recording and topographic analysis of the auditory odd-ball event-related potential in patients with schizophrenia. Biol Psychi-atry 44:982–989.

29. O’Donnell BF, Hokama H, McCarley RW, Smith RS, Salisbury DF, Mond-row E, et al. (1994): Auditory ERPs to non-target stimuli in schizophrenia:Relationship to probability, task-demands, and target ERPs. Int J Psycho-physiol 17:219 –231.

30. Salisbury DF, Collins KC, McCarley RW (2009): Reductions in the N1 and P2auditory event-related potentials in first-hospitalized and chronic schizophre-nia [published online ahead of print March 12]. Schizophr Bull.

31. Salisbury DF, O’Donnell BF, McCarley RW, Shenton ME, Benavage A(1994): The N2 event-related potential reflects attention deficit inschizophrenia. Biol Psych 39:1–13.

32. van der Stelt O, Frye J, Lieberman JA, Belger A (2004): Impaired P3Generation reflects high-level and progressive neurocognitive dysfunc-tion in Schizophrenia. Arch Gen Psychiatry 61:237–248.

33. Winterer G, Egan MF, Radler T, Coppola R, Weinberger DR (2001): Event-related potentials and genetic risk for schizophrenia. Biol Psychiatry50:407– 417.

34. O’Donnell BF, Shenton ME, McCarley RW, Faux SF, Smith RS, SalisburyDF, et al. (1993): The auditory N2 component in schizophrenia: Relation-ship to MRI temporal lobe gray matter and to other ERP abnormalities.Biol Psychiatry 34:26 – 40.

35. Spitzer RL, Williams JB, Gibber M, First MB (1992): The Structured clinicalinterview for DSM-III-R (SCID). I: History, rationale, and description. ArchGen Psychiatry 49:624 – 629.

36. Wittchen HU (2001): Reliability and validity studies of the WHO-Com-posite International Diagnostic Interview (CIDI): A critical review. J Psy-chiatr Res 28:57– 84.

37. Schmand BA, Bakker D, Saan R, Louman J (1991): The Dutch reading testfor adults: A measure of premorbid intelligence level. Tijdschr Gerontol

Geriatr 22:15–19.

www.sobp.org/journal

8. Gratton G, Coles M, Donchin E (1982): A new method for off-line removalof ocular artifact. Electroencephalogr Clin Neurophysiol 55:468 – 484.

9. Duncan CC, Barry RJ, Connolly JF, Fischer C, Michie PT, Näätänen R, et al.(2009): Event-related potentials in clinical research: Guidelines for elic-iting, recording, and quantifying mismatch negativity, P300, and N400.Clin Neurophysiol 120:1883–1908.

0. Turetsky BI, Bilker WB, Siegel SJ, Kohler CG, Gur RE (2009): Profile of auditoryinformation-processing deficits in schizophrenia. Psychiatry Res 165:27–37.

1. Rosburg T, Boutros NN, Ford JM (2008): Reduced auditory evoked po-tential component N100 in schizophrenia—a critical review. PsychiatryRes 161:259 –274.

2. Buchsbaum MS (1977): The middle evoked response components andschizophrenia. Schizophr Bull 3:93–104.

3. Golob EJ, Ringman JM, Irimajiri R, Bright S, Schaffer B, Medina LD, Starr A(2009): Cortical event-related potentials in preclinical familial Alzheimerdisease. Neurology 73:1649 –1655.

4. Renoult L, Prévost M, Brodeur M, Lionnet C, Joober R, Malla A, DebruilleJB (2007): P300 asymmetry and positive symptom severity: A study inthe early stage of a first episode of psychosis. Schizophr Res 93:366 –373.

5. Salisbury DF, Shenton ME, Sherwood AR, Fischer IA, Yurgelun-Todd DA,Tohen M, McCarley RW (1998): First-episode Schizophrenic psychosisdiffers from first-episode affective psychosis and controls in P300 am-plitude over left temporal lobe. Arch Gen Psychiatry 55:173–180.

6. Miller TJ, McGlashan TH, Rosen JL, Cadenhead K, Cannon T, Ventura J, et al.(2003): Prodromal assessment with the structured interview for prodromalsyndromesandthescaleofprodromalsymptoms:Predictivevalidity, interraterreliability, and training to reliability. Schizophr Bull 29:703–715.

7. Kay SR, Fiszbein A, Opler LA (1987): The Positive and Negative SyndromeScale (PANSS) for schizophrenia. Schizophr Bull 13:261–276.

8. Cannon-Spoor H, Potkin G, Wyatt J (1982): Measurement of premorbidadjustment in chronic schizophrenia. Schizophr Bull 8:470 – 484.

9. Lindenmayer JP, Grochowski S, Hyman RB (1995): Five factor model ofschizophrenia: Replication across samples. Schizophr Res 14:229 –234.

0. Bryson G, Bell M, Greig T, Kaplan E (1999): Internal consistency, temporalstability and neuropsychological correlates of three cognitive compo-nents of the Positive and Negative Syndrome Scale (PANSS). SchizophrRes 27:27–35.

1. Wood SM, Potts GF, Martin LE, Kothmann D, Hall JF, Ulanday JB (2007):Disruption of auditory and visual attention in schizophrenia. PsychiatryRes Neuroimaging 156:105–116.

2. Potts GF, O’Donnell BF, Hirayasu Y, McCarley RW (2002): Disruption ofneural systems of visual attention in schizophrenia. Arch Gen Psychiatry59:418 – 424.

3. Pfefferbaum A, Ford JM, White PM, Roth WT (1989): P3 in schizophreniais affected by stimulus modality, response requirements, medicationstatus, and negative symptoms. Arch Gen Psychiatry 46:1035–1044.

4. Light GA, Braff DL (2005): Mismatch negativity deficits are associatedwith poor functioning in schizophrenia patients. Arch Gen Psychiatry62:127–136.

5. Maeda K, Nakagome K, Kasai K, Ikebuchi E, Iwanami A, Kudoh N, et al. (2005):Electrophysiological indices associated with social functioning outcome inschizophrenia: A 5-year follow-up study. Neurosci Res 51:215–218.

6. Harris MG, Henry LP, Harrigan SM, Purcell R, Schwartz OS, Farrelly SE, etal. (2005): The relationship between duration of untreated psychosisand outcome: An eight-year prospective study. Schizophr Res 79:85–93.

7. McGorry PD, Yung AR, Phillips LJ, Yuen HP, Francey S, Cosgrave EM, et al.(2002): Randomized controlled trial of interventions designed to reducethe risk of progression to first-episode psychosis in a clinical samplewith subthreshold symptoms. Arch Gen Psychiatry 59:921–928.

8. Takahashi T, Wood SJ, Yung AR, Soulsby B, McGorry PD, Suzuki M, et al.(2009): Progressive gray matter reduction of the superior temporal gy-

rus during transition to psychosis. Arch Gen Psychiatry 66:366 –376.