394 Electroencephalography and clinical Neurophysiology, 87 (1993) 394-402 © 1993 Elsevier Scientific Publishers Ireland, Ltd. 0013-4694/93/$06.00

EEG93131

Electrical source analysis of auditory ERPs in medial temporal lobe amnestic syndrome

Brian F. O'Donnel l Ronald A. Cohen b Hiroto Hokama a, B.N. Cuffin c, Carol Lippa b, Martha E. Shenton a and David A. Drachman b

a Department of Psychiatry, 116,4, Harvard Medical School and Brockton Veterans Administration Medical Center, 940 Belmont Street, Brockton, MA 02401 (USA), b Department of Neurology, University of Massachusetts Medical Center, Boston, MA (USA),

and c Francis Bitter Magnet Laboratory, MIT, Cambridge, MA (USA)

(Accepted for publication: 21 July 1993)

Summary Auditory event-related potential (ERP) components have been anatomically linked to temporal lobe structures and functionally related to attentional and memory processes. We recorded auditory ERPs using an "oddball" paradigm from two patients with amnestic syndromes secondary to medial temporal lobe encephalitic infections. The oddball paradigm elicits the exogenous NI and P2 components, and the endogenous N2 and P3 components. Electrical source analysis was used to test for alterations in source strength and orientation in these patients compared to control subjects. Symmetric dipoles placed in the temporal lobe region were used to measure ERP component activity. In the patient with a lesion confined to the left, medial temporal lobe, including the posterior hippocampus, dipole orientation was displaced anteriorally. In the patient with lesions to the anterior medial temporal lobe, temporal poles, and orbital frontal cortex, the negative components of the ERP (N1 and N2) were reduced in the right hemisphere, accompanied by disturbed orientation. These findings are consistent with other evidence that the different components of the auditory ERP can be dissociated on the basis of lesion effects, and that the antero-posterior extent of encephalitic lesions may play an important role in modulating ERP abnormalities.

Key words: Auditory event-related potential; Electrical source analysis; Neural generators; Encephalitis

Even t - r e l a t ed po ten t i a l s ( E R P s ) a re changes in the E E G el ic i ted by sensory s t imula t ion or synchron ized with a behav io ra l ou tput . The aud i to ry E R P e l ic i ted dur ing a d i sc r imina t ion task with low-probab i l i ty ta rge t s t imuli varies m a r k e d l y with the s t imulus task re le- vance. In the aud i to ry " o d d b a l l " pa r ad igm, bo th non- t a rge t and t a rge t tones el ici t a negat ive de f lec t ion in the E R P at abou t 100 msec, the N1 or N100 compo- nent , fo l lowed by a posi t ive de f lec t ion at abou t 200 msec, the P2 or P200 componen t . Ta rge t s t imuli elicit 2 add i t iona l componen t s : a negat ive de f lec t ion at 200 - 400 msec, the N2 or N200 componen t ; and a posi t ive def lec t ion at 300 msec or la ter , the P3 or P300 compo- nent . Because the la tency and a m p l i t u d e of the N2 and P3 c o m p o n e n t s a re highly react ive to factors such as s t imulus probabi l i ty , expectancy, s t imulus discr im- inabil i ty, and dec is ion conf idence , they have been used as non-invasive, neurophys io log ica l p robes of pe rcep - tual and cogni t ive p rocess ing in humans (Sut ton et al. 1965; Donch in and Coles 1988; Novak et al. 1990). Never the less , the pe r cep tua l and cogni t ive funct ions,

* Corresponding author. Tel.: (508) 583 4500, ext. 3935.

r e p r e s e n t e d by this ser ies of componen t s , r emain poor ly def ined . Severa l theor ies have l inked the P3 compo- nent to m e m o r y processes . Donch in and Coles (1988) have a rgued tha t the P3 c o m p o n e n t indexes the pro- cess of context upda t ing . Gros sbe rg (1984) has sug- ges ted tha t it ref lects a reset of sho r t - t e rm memory .

Both the N2 and P3 c o m p o n e n t s a re also af fec ted by b ra in dysfunct ion. N2 and P3 la tencies increase with age (at the ra te of 1 - 2 m s e c / y e a r ) and are fu r the r p r o l o n g e d in d e m e n t i n g d i so rde r s ( G o o d i n et al. 1978; P fe f f e rbaum et al. 1984a,b; O ' D o n n e l l et al. 1990, 1992; Polich 1991). Moreover , the la tency of P3 shows an inverse co r re la t ion with in te l lec tua l p e r f o r m a n c e in no rma l subjects (Pol ich et al. 1983; E m m e r s o n et al. 1990; O ' D o n n e l l et al. 1992), in pa t i en t s with Park in- son 's d i sease (Hansch et al. 1982; O ' D o n n e l l et al. 1987), and in d e m e n t i a (Pol ich et al. 1986; O ' D o n n e l l et al. 1990). R e d u c t i o n of the amp l i t ude of the audi tory P3 c o m p o n e n t is one of the most robus t physiological abnormal i t i e s r e p o r t e d in sch izophren ic pa t ien t s on or off med ica t i on (e.g., Ro th et al. 1980; P fe f fe rbaum et al. 1984b, 1989; Faux et al. 1990, 1993).

The psychological s ignif icance of the aud i to ry ERP, and its sensit ivity to b ra in dysfunct ion, has mot iva ted a

ERPs IN M E D I A L T E M P O R A L A M N E S I A 395

variety of approaches over the past decade seeking to identify the neural electrical sources of scalp-recorded ERP components in humans. Vaughan and Arezzo (1988) have argued on the basis of scalp topography and functional neuroanatomy that both the N1 and P2 components are primarily generated from the supra- temporal plane. However, the N1 component may be spared after extensive bilateral lesions to the temporal lobe, unless the lesions extend into parietal tissue, suggesting non-temporal sources of this component (Woods et al. 1987). Depth field potential recordings in patients with seizure disorders indicate that P3-1ike potentials show polarity inversions in medial temporal region structures, most likely reflecting a hippocampal generator (Halgren et al. 1980, 1986; Stapleton et al. 1987; McCarthy et al. 1989). Other lesion studies in humans suggest that medial temporal lesions have min- imal impact on the surface-recorded P3 component. A patient with anterograde amnesia due to a left infiltrat- ing glioma which affected the left medial temporal lobe showed normal and symmetrical auditory and visual P300 components (Rugg et al. 1991). Patients with unilateral temporal lobe resections for seizure control do not show reductions in auditory P3 amplitude or right-left asymmetries of P3 topography that might be expected if these structures made a significant contri- bution to the scalp P3 (Stapleton et al. 1987; Johnson 1988). It should be noted that this procedure usually spares the posterior hippocampal formation where depth-recorded P3-1ike potentials have been most prominent (Halgren et al. 1986; McCarthy et al. 1989; Knight 1990). Knight et al. (1989) have reported that lesions of the superior temporal lobe and adjacent parietal tissue reduce P3 amplitude on scalp record- ings, but do not alter the N2 component. In schizophrenic patients, reduced left posterior superior temporal gyrus gray volume has been associated with P3 amplitude reduction at scalp electrode sites adja- cent to the temporal lobe, while N2 amplitude reduc- tion has been related to reduction in medial temporal lobe structures as well (McCarley et al. 1990; O'Don- nell et al. 1993). In conjunction, these studies suggest that while temporal lobe structures are important for generation of the scalp-recorded auditory ERP, the sequential components that make it up may have sepa- rable generator circuits and electrical sources. In addi- tion, studies of patients with focal lesions may be helpful in dissociating these ERP components, which appear as an ensemble in healthy individuals.

Patients who have suffered medial temporal lobe amnestic syndromes secondary to encephalitic infec- tions provide a means of testing both structural and functional hypotheses related to auditory ERP compo- nents: whether they are sensitive to lesions of the hippocampus, amygdala, and parahippocampal struc- tures; and whether disrupted memory processes in the

presence of otherwise intact intellectual function result in scalp-recorded ERP abnormalities. Ford et al. (1991) described auditory ERPs recorded from a patient who suffered severe left temporal lobe damage secondary to herpes simplex encephalitis. This patient suffered uni- lateral destruction of the tissue of both medial and neocortical left temporal lobe, as well as pathology of the medial, left orbital frontal area. The patient was left-handed, suffered a transient aphasia, and the pathology affected the superior plane of the temporal lobe. Verbal memory was more impaired than spatial memory performance. This patient had a small but symmetrical auditory P3 component compared to a control subject. The N2 component was absent, but only over the electrode sites which overlay the lesioned left temporal lesion. Onofrj et al. (1991, 1992) de- scribed a series of 5 patients with bilateral lesions to the anterior and midtemporal lobes secondary to en- cephalitis. All 5 patients were globally amnestic. In these patients, the N1, P2 and N2 components were of normal amplitude and latency. The P3 component was bilaterally reduced at lateral sites over the temporal and frontal lobes (left side: F7 and T3; right side: F8 and T4) and was in the low normal range at midline and posterior electrode sites. P3 latency was prolonged in 1 patient. This series of cases suggests that the P3 component persists in medial temporal lobe amnestic syndrome, but that topographic abnormalities may oc- cur. This finding is consistent with the presence of multiple electrical sources for the P3 component, prob- ably including structures within the medial temporal lobe (Halgren et al. 1986).

One problem in the analysis of subtle topographic differences in patients is the procedure for statistical testing of individual or group differences. With large numbers of electrodes, multiple t tests across individ- ual electrode sites will result in a significant number of type I errors. In addition, a topographic difference due to alteration in the geometry of underlying sources may not be reflected in a significant difference at any one electrode site, resulting in a type II error. Electrical source analysis, or the use of topographic electroen- cephalographic (EEG) data to obtain information about the strength, location, and orientation of electrical sources within the brain, is a recently developed tech- nique which can address these problems (Nunez 1981; Fender 1987; Scherg 1990; Williamson and Kaufman 1990; Cohen and Cuffin 1991; Rogers et al. 1991). The determination of the electrical sources of scalp-re- corded EEG voltage distribution is usually referred to as the inverse problem. Sources are often represented as equivalent dipoles. A current dipole is an electrical source, where its strength or moment is measured in units of current x length. While procedures to accu- rately localize single and multiple dipoles are still being validated (Cohen and Cuffin 1991), a dipole model can

396 B.F. O'DONNELL ET AL.

be used to reduce the voltage values recorded across the entire scalp to a small set of parameters , reducing the likelihood of type I errors. In addition, by providing separate values for dipole strength and orientation, differences in source orientation can be detected even when the dipole strength remains within normal range.

In a clinical setting, patients with discrete neuro- logic lesions to putative source structures may be used in conjunction with source analysis techniques to test hypotheses regarding neural generators (Scherg and Von Cramon 1986). We report auditory ERP abnor- malities evaluated using electrical source analysis in two patients with post-encephalitic medial temporal lobe lesions and amnestic syndromes due to herpes encephalitic infections. Since these patients had intact skulls, the presence or absence of ERP abnormalities could not be due to atypical volume conduction from the brain. Since both patients had amnestic syndromes, it was possible to consider whether disturbed memory processes affected the P3 or other components of the auditory ERP.

Methods and patients

Patient A The patient was a 46-year-old white right-handed

woman who developed an amnestic syndrome at the age of 16 subsequent to a herpes simplex encephalitic infection. Her short- term course and amnestic syn- drome have been previously described (case 2, Drach- man and Adams 1962). An arsenic scan in 1959 showed bilateral medial- temporal and orbital frontal damage. Recent MRI T-2 weighted images (Fig. 1A and B) showed widespread lesions affecting the inferior and medial aspects of both temporal lobes, right worse than left, anterior worse than posterior. There was enlarge- ment of both sylvian fissures, without diffuse cortical atrophy. In addition, lesions of the basal frontal lobe and olfactory tracts were observed. While the patient has shown some recovery of memory function, her current Wechsler Memory Quotient of 99 (Wechsler 1945) remains depressed compared to her WAIS-R Full Scale IQ of 122 (Wechsler 1981). She was most severely impaired on tests of delayed recall of verbal and visual material. In contrast, working memory as measured by the number of digits she could recall forward (8) and backward (6) was intact. Other residual disturbances include anosmia, ageusia, compulsive eat- ing, flat affect, behavioral inertia, social withdrawal, and impoverished ideation. In summary, her radiologi- cal and clinical findings were consistent with bilateral lesions to the medial temporal and basal frontal re- gions. (MRI scans were evaluated by D.A.D. and C.L., who were blind to the E R P results on both patients.)

Fig. 1. A and B: horizontal T-2 weighted MRI sections of patient A showing severe, bilateral atrophy of medial temporal lobes. The right temporal lobe (viewer left) is more affected than the left tempo-

ral lobe.

Patient B Patient B was a 60-year-old right-handed civil engi-

neer who suffered an amnestic syndrome after a herpes simplex encephalitic infection at the age of 59. He subsequently showed a dense retrograde amnesia and was unable to continue working after the infection. Despite insight into his memory impairment, he was affable and cheerful after discharge from the hospital.

ERPs IN MEDIAL TEMPORAL AMNESIA 397

Fig. 2. Horizontal T-2 weighted MRI section of patient B showing left medial temporal lesion (high intensity signal) in patient B,

affecting the hippocampus and posterior amygdala.

T-2 weighted M R I images showed high intensity signal in the left medial temporal lobe involving the hip- pocampus and posterior amygdala (Fig. 2). Neocortical temporal lobe structures, including the superior tempo- ral gyrus, were spared. In contrast to patient A, the right temporal lobe and frontal lobe appeared intact. At the time of ERP evaluation, his WAIS-R Full Scale IQ was 103, while his WMS Quotient was 93. Like patient A, his most profound deficit was on the de- layed recall of verbal or visual material. For example, on the Logical Memory subtest of the WMS, he re- membered 3.5 items on immediate recall and none after a 5 min delay. Working memory span was intact (7 digits forward, 6 digits backward).

Control subjects Twelve control subjects were evaluated who were

between 40 and 70 years of age (mean age = 49 _+ 8 S.D.). All had Mini-Mental Status scores greater than 27 (Folstein et al. 1975). All control subjects were right-handed. Eleven were male and one was female. None of the control subjects was under medical treat-

ment, and none had a history of head trauma, neuro- logic illness, or alcohol abuse.

ERP recording procedures Event-related potentials (ERPs) were recorded us-

ing an auditory oddball paradigm (Faux et al. 1990). ERPs were elicited by tone pips of 40 msec duration (10 msec r i se / fa l l time) with a 1.2 sec ISI. Infrequent ( P = 0.15) high pitched tones (1500 Hz, 97 dB SPL) were presented pseudorandomly, interspersed among frequent low pitched tones (1000 Hz, 97 dB SPL). Tones were presented through Etymotic insert ear- phones against a background of continuous 70 dB binaural white noise. Subjects were asked to silently count the number of infrequent, high pitched tones. All subjects were able to perform the task with 97% or bet ter detection of target tones. While listening to the tones, subjects stared at a central fixation point to reduce eye movements. Subjects received 800 tones in this condition.

ERPs were recorded from 28 tin plate scalp elec- trodes using an Electro-Cap International electrode cap. Scalp electrode placements included all electrodes in the international 10-20 system with 8 additional interpolated electrodes. Fpl , Fp2 and Cz sites were located manually by international 10-20 measure- ments, and all other electrodes were positioned auto- matically at standard distances. A vertical electro-oc- ulogram (EOG) was recorded using right eye supra- and infra-orbital electrodes. Horizontal E O G was recorded from electrodes at the right and left external canthi. Electrode impedance was maintained at less than 4 kl2. Right and left ear impedance was matched within 1 kg2. The E E G was filtered using a bandpass of 0.15-40 Hz, with 36 d B / o c t a v e roll-off for low pass and 6 d B / o c t a v e for high pass. Single-trial epochs were digitized and stored on hard disk for later off-line processing. Each ERP consisted of 71)0 msec epoch, sampled at a rate of 2.734 msec / s amp le , with a 100 msec prestimulus baseline interval.

Data processing All single-trial epochs were baseline corrected prior

to subsequent processing. ERP responses with vertical E O G artifact were corrected through individually com- puted weighting coefficients at each electrode site us- ing the Semlitsch et al. (1986) procedure. After correc- tion for vertical E O G artifact, all epochs with voltages in excess of +50 ~V at any electrode site were re- jected. Epochs were then sorted into infrequent and frequent ERP averages. Averages were low-pass fil- tered at 16 Hz prior to source analysis.

Source analysis Peak component latencies were measured at Cz,

since all subjects showed the 4 components of interest

3 9 8

at this electrode site. Peak N1 latency was measured at the most negative voltage from 80 to 140 msec, and peak P2 latency at the most positive voltage between 180 and 290 msec from the ERP averaged to frequent tones. Peak N2 latency was measured at the most negative voltage between 200 and 300 msec in the difference ERP. Peak P3 latency was measured at the most positive peak between 300 and 450 msec in the rare, target ERP (Simson et al. 1977; Novak et al. 1990; O'Donnell et al. 1990, 1992, 1993).

Source analysis was computed using the Brain Elec- trical Source Analysis (BESA) program developed by Scherg and colleagues (Scherg and Von Cramon 1986; Scherg 1990; Scherg and Berg 1991). BESA uses a 3-shell head model and a temporo-spatial representa- tion of dipole activity (dipoles are typically fixed in location and vary in strength and orientation over time). Dipoles were positioned in symmetric locations within a spherical head model within the superior, posterior temporal lobe region, since the sources of all 4 components of interest (N1, P2, N2, P3) have been attributed to temporal lobe structures. The BESA co- ordinates for the right hemisphere dipole were: 61% eccentricity, 84 ° theta, and 0 ° phi; and for the left hemisphere: 61% eccentricity, - 8 4 ° theta, and 0 ° phi (Fig. 3). For the subsequent analysis, the locations of these dipoles were fixed, and the orientation and dipole moment (strength) were adjusted to fit the scalp topog- raphy over a 20 msec time interval for each compo- nent. The interval for the solution was based on the peak latency of the component minus 20 msec. The dipole, therefore, reflected the voltage distribution over the scalp during the ascending phase and peak of each ERP component.

An iterative series of calculations, going from coarse (20 ° steps) to fine (2 ° steps) adjustments of orientation, was used to arrive at the inverse solution for each subject and ERP component, using the simplex method to minimize residual variance. Closeness of fit between the measured ERP component and the dipole solution was measured as percentage residual variance. Dipole magnitude was measured as /ZVef f (Scherg and Von Cramon 1986; Scherg and Berg 1991). Dipole orienta- tion was initially represented as a unit vector using spherical coordinates. Theta was the azimuth angle in degrees from a vertical z axis through the center of the head. Angular displacement was measured between 0 and 180 °, with 0 ° at the vertex of the head and 180 ° pointing directly downward. Theta values were positive for left displacement, and negative for right displace- ment. Phi measures the latitude angle in the horizontal plane, with values varying from 0 ° (corresponding to the right mastoid) to 360 ° (180 ° corresponds to the left mastoid). A phi value of 90 ° was oriented toward the anterior of the head, and 270 ° to the posterior of the head. This coordinate system is shown in Fig. 4, with

F R E Q U E N T

B . F . O ' D O N N E L L E T A L

TARGET T A R G E T - F R E Q U E N T

+ , 4

-2 t t

-S

Cz

10 P2

4 i:

i 2 I "

i+

lO Ix l \

, v/ ',

2 ¢

-e l e P3

111 ,o ,;,,, ,,,,

-e

4 I " I ee I ~

16 P3

l!,,

I ~ t I , ! 'l '~ i l l ,

II : 11

Pz ControlB _++ II •

,o /i, ,t 1o Pa t i en t A . . . . . . 10 ] v t l t I

i P a t i e n t II . . . . . . . . . . . . . r ff / • I

-2 2 .~,,. ,,' -

--eloo 0 l O 0 2 0 0 s o o 4 0 0 s o o e o o --8100 0 l o 0 2 0 0 a o o 4 0 0 s o o e o o e loo 0 I 00200300400s00~00

MS MS MS

Fig. 3. Upper figure: ERPs elicited by rare, target tones, frequent, non-target tones, and subtraction wave forms (rare-frequent) from control subjects (solid line, average of 12 subjects), patient A (fine dotted line), and patient B (coarse dotted line) at midline electrode sites. Positive is upward on all figures. Frequent tones elicit the N1 and P2 components. Rare tones elicit these components, as well as the N2 and P3 components. Subtraction of the ERPs to frequent tones from the ERPs to rare tones isolates the N2 component. Note the absence of the N1 and N2 components in patient A (bilateral anterior temporal and orbital frontal lesions), and the absence of the P3 component at Pz in patient B (left medial temporal lesion). Lower figure: position of symmetric dipoles in superior, posterior temporal region of the left and right hemisphere. Dipole positions were fixed for the source analysis, and dipole magnitude and orientation were

adjusted to fit the surface voltage distribution.

the orientation of the control and patients shown for the P3 component. Phi and theta angles are shown for both the left and right hemisphere dipoles.

Because the arithmetic mean of a set of orientations can yield gross errors in central tendency and standard deviation because of the discontinuity at 0 /360 °, vector components along Cartesian axes were used to mea- sure dipole displacement in the z, x, and y directions for each dipole. Cosine theta measured displacement along the z (vertical) axis, sine phi along the x (right-left) axis, and cosine phi along the y (antero-posterior) axis. The mean and standard deviation of the control values for each of these vector components were used to

ERPs IN M E D I A L T E M P O R A L AM NE S IA

P 3 D I P O L E O R I E N T A T I O N S N1 Topography

399

P2 Topography

H E M I S P H E R E LEFT R I G H T

Control

PHI

180

anter ior anter ior

90 90 @0 "@O 270 270

posterior poster ior

T H E T A up up o o -90@90 -90@

- 180/180 - 180/180

down down

90

Fig. 4. P3 dipole orientation and magni tude values for control subjects (C) and patients (A, B) in theta (azimuth) and phi (latitude) angles. The circles indicate the orientation values for the dipoles placed in the left and right temporal regions. (The circles do not represent the head.) The length of each dipole represents dipole magnitude, with the radius value of 1 /xVef r Mean control orienta-

tion was computed on the basis of the vector sum.

A

÷2 .,_~J i ,5

W 1.,~ I Z , . 5 • o.,) i ~ . 5 i I . e I ~ . ~ , m~.o I215

• 4, ~

• 7 .,]uU

I 4 . 5

I;:)

B

Fig. 5. Interpolated voltage distribution of the peak N1 and P2 components, measured from the control group average, and patients

A and B.

tients. N2 amplitude appeared to be reduced and dis- placed to the posterior region of the scalp in patient A. P3 amplitude and topography were WNL for patient A, but the maximum P3 amplitude appeared to be shifted forward in patient B.

N2 Topography P3 Topography

evaluate the deviance of values obtained from each of the patients, with a _+ 2 S.D. criterion for abnormality.

Sontrol

Results

Fig. 3 shows the ERPs to the rare, target tones and frequent tones from the 2 patients and the grand average ERP from the 12 control subjects. The latency of P3 measured at Cz in the patients (330 msec in both patients) was earlier (> 2 S.D.) than control subjects (391 +_ 29 msec). The latencies of other components were within normal limits.

Fig. 5 shows the distribution of peak N1 and P2 voltages measured from the averaged ERP to the fre- quent tones. Fig. 6 shows the voltage distribution for the N2 and P3 component peaks. N1 amplitude ap- peared smaller in patient A, and shifted to the left frontal region of the scalp. P2 amplitude appeared to be of normal amplitude and topography in both pa-

A

2.'-

II,O m,.s no., ml,i ml.~ m2.4 m.~.~

I : : ) ms.~ mm.~.~,

+Z3.8,J 12.!

Wu.' - m le.2 m)~ • ~,4 m-.4 m~,5 • 5,¢ m4.~

I ] : ; m~.~

B

Fig. 6. Interpolated voltage distribution of the peak N2 and P3 components measured from the control group averages, and from patients A and B. N2 was measured from the target minus non-target

difference ERP, and P3 from the target ERP.

400

T A B L E I

D i p o l e m a g n i t u d e (/.~Veff) a n d o r i e n t a t i o n c o m p o n e n t s .

C o n t r o l P a t i e n t A P a t i e n t B

M e a n (S.D.)

Right N1

P2

N2

P3

N1

P2

N2

P3

m a g n i t u d e - 0.24 (0.05) - 0 .07 * 0.21

cos t h e t a (z) 0.75 (0.16) - 0.31 * 0.84

sin ph i (x) 0.82 (0.18) - 0.74 * 0.99

cos ph i (y) 0.39 (0.39) 0.67 - 0 . 1 6

m a g n i t u d e 0.31 (0.22) 0.37 0.25

cos t h e t a 0.82 (0.13) 0.85 0.71

sin ph i 0.00 (0.78) 1.00 0.99

cos ph i 0.51 (0.45) 0.04 - 0.10

m a g n i t u d e - 0 . 3 3 (0.11) - 0 . 0 7 * - 0 . 3 0

cos t h e t a 0.90 (0.20) - 0.28 * 0.91

sin phi 0.02 (0.64) - 0.92 0.31

cos ph i 0.67 (0.44) 0.39 0.95

m a g n i t u d e 0.42 (0.18) 0.67 0.32

cos t h e t a 0.85 (0.13) 0.97 0.39

sin ph i - 0.62 (0.56) - 0.73 0.99

cos ph i 0.13 (0.59) 0.68 - 0.14

m a g n i t u d e - 0.23 (0.06) - 0.14 - 0.27

cos t h e t a 0.81 (0.16) 0.67 0.95

sin phi 0.82 (0.20) 0.91 0.99

cos ph i 0.28 (0.47) - 0.41 0.11

m a g n i t u d e 0.33 (0.19) 0.30 0.25

cos t h e t a 0.90 (0.07) 0.95 0.99

sin ph i 0.09 (0.83) 1.00 0.44

cos phi - 0.35 (0.50) 0.01 0.90

m a g n i t u d e - 0.39 (0.22) - 0 .10 - 0.45

cos t h e t a 0.67 (0.41) 0.49 0.84

sin phi - 0.35 (0.65) - 0.77 0.97

cos ph i - 0.36 (0.63) 0.64 - 0.24

m a g n i t u d e 0.35 (0.20) 0.58 0.21

cos t h e t a 0.82 (0.17) 0.87 0.36

sin phi - 0.52 (0.41) - 0.59 0.92

cos ph i 0.20 (0.76) - 0 . 8 1 0.40

* > 2 S.D.

Table I shows dipole magnitude values for each component. Right N1 and N2 dipole moments were reduced in patient A. Orientation was analyzed in terms of x, y and z vector components. The vector component values in the antero-posterior (sine phi; y axis), left-right (cosine phi; x axis), and up-down (sine theta; z axis) directions are also listed in Table I.

Both patients showed abnormal dipole orientations. Patient A showed a more negative cosine theta value for N1 on the right side, indicating that this component was displaced from vertical toward the left side. The sine phi value was also more negative than in the control group, indicating that N1 was displaced in the posterior direction. Patient A had a more negative cosine theta for the right side N2 dipole, indicating that this dipole was displaced from vertical compared to the control group.

Patient B showed a more positive cosine phi value for the right P2 dipole, indicating that the P2 compo- nent was displaced to the right. P3 was abnormal on

B.F. O ' D O N N E L L E T AL.

both sides, showing smaller cosine theta values and large, positive sine phi values. These indicate greater displacement from the vertical axis, and an anterior shift in dipole orientation.

In summary, patient A showed a diminution of dipole strength for the N1 and N2 components located in the right temporal lobe, coupled with abnormal orientations. While patient B showed normal dipole strength bilaterally, the orientation of the P3 dipoles was altered. The orientation of the P2 dipole on the left side in patient B was also disturbed.

Discussion

Auditory ERP abnormalities were present in both patients with encephalitic damage, but the components affected differed in each case. In patient B, in whom the lesion was confined to the left medial temporal lobe and included the posterior hippocampus, P2 ori- entation on the left side and P3 orientation bilaterally were altered, with no significant change in dipole strength. These findings are consistent with those of Onofrj et al. (1992), who reported that 5 patients with medial temporal lobe lesions showed abnormal P3 to- pography, with reduction in amplitude confined to the most lateral electrode recording sites. Seizure patients with left temporal lobe foci have also been reported to show P3 amplitude asymmetries at lateral electrode sites (McCarthy et al. 1987). Such findings are consis- tent with depth electrode evidence (Halgren et al. 1980, 1986; McCarthy et al. 1989) that the medial temporal lobe, particularly the posterior hippocampal region, may contribute to the scalp-recorded P3 com- ponent in humans. This contribution to the scalp-re- corded P3 appears to be small in magnitude and largely confined to lateral and possibly posterior electrode sites.

In patient A, who had severe bilateral lesions of the anterior medial temporal lobes, temporal poles and orbital frontal damage, both P2 and P3 were intact. This finding appears to be consistent with results ob- tained from patients with unilateral temporal lobe re- sections for seizure control, who fail to show P3 ampli- tude reduction or asymmetries (Stapleton et al. 1987; Johnson 1988). The posterior hippocampus is usually spared in this procedure, as is also likely in patient A. Patient A did show reductions in magnitude and dis- turbances of orientation of the negative ERP compo- nents, N1 and N2. These abnormalities were significant only in the dipole located in the right hemisphere, which also incurred the most severe damage from the encephalitic infection. Reduction of N2 amplitude over the lesioned hemisphere has also been reported in a post-encephalitic case by Ford et al. (1991). This pa- tient, like patient A, had damage to the orbital frontal

ERPs IN MEDIAL TEMPORAL AMNESIA 401

lobe, suggesting that the N2 c o m p o n e n t might reflect damage to s t ructures in this region. Only one of the

group of 5 pat ients repor ted by Onofrj et al. (1992) had frontal damage, and in this pa t i en t the lesion was conf ined to the right media l f rontal region.

P3 latency was earl ier in both patients . This f inding suggests that the cognitive processes represen ted by P3, such as s t imulus evaluat ion and context updat ing, may be p remature ly t e rmina ted in these amnest ic pa- t ients compared to control subjects. Alternat ively, this f inding may be the result of changes in over lapping wave forms which act to increase P3 ampl i tude in the earl ier phase of the component . This explanat ion is most plausible in pa t ien t A, in whom the N2 compo- nen t was essentially absent . These latency effects may have been specific to the parad igm used in this lab, in which tones were p resen ted against a background of con t inuous white noise. Because this decreases the discriminabil i ty of the stimuli, P300 latency is later compared to the paradigms used by Ford et al. (1991) and Onofrj et al. (1992).

These data also suggest that the ampl i tude of the P3 componen t elicited by a tonal d iscr iminat ion is mini- mally affected by severe d is turbances of memory pro- cesses. Meador et al. (1987) also found that tonal P3 componen t s could be elicited in 7 pat ients with disor- ders of recent memory of diverse etiologies, but that their P3 c o m p o n e n t to phonemic and semant ic process- ing of spoken words was decreased. Therefore , the P3 componen t elicited by tonal and verbal d iscr iminat ions may reflect different neura l processes and be differen- tially affected by memory dis turbance.

The source analysis methodology employed in this study shows promise in the evaluat ion of cognitive event - re la ted potent ia l abnormal i t ies in pat ients with b ra in lesions and may be especially useful when large number s of recording electrodes are used. Conjoin t use of anatomic, neuropsychological and source analysis data may provide a powerful methodology to be t te r def ine the ana tomic genera tors and funct ional signifi- cance of specific E R P components .

This work was supported by the Department of Veterans Affairs Merit Review Program, NIMH 40,799, and the NIH Research Scien- tist Career Development Award KO4-MH00746-04.

The authors greatly appreciate the technical support of R. Scott Smith, Cora Lee and Matthew Kimble in data collection and process- ing.

R e f e r e n c e s

Cohen, D. and Cuffin, B.N. EEG versus MEG localization accuracy: theory and experiment. Brain Topogr., 1991, 4: 95-103.

Cuffin, B.N., Cohen, D., Yunokuchi, K., Maniewski, R., Purcell, C., Cosgrove, G.R., Ives, J., Kennedy, J. and Schomer, D. Tests of EEG localization accuracy using implanted sources in the human brain. Ann. Neurol., 1991, 29: 132-138.

Donchin, E. and Coles, M.G.H. Is the P300 component a manifesta- tion of context updating? Behav. Brain Sci., 1988, 11: 357-373.

Drachman, D.A. and Adams, R.D. Herpes simplex and acute inclu- sion-body encephalitis. Arch. Neurol., 1962, 7: 61-79.

Emmerson, R.Y., Dustman, R.E., Shearer, D.E. and Turner, C.W. P3 latency and symbol digit performance correlations in aging. Exp. Aging Res., 1990, 15: 151-159.

Faux, S.F., Shenton, M.E., McCarley, R.W., Nestor, P.G., Marcy, B. and Ludwig, A. Preservation of P300 event-related potential topographic asymmetries in schizophrenia with use of either linked-ear or nose reference sites. Electroenceph. clin. Neuro- physiol., 1990, 75: 378-391.

Faux, S.F., McCarley, R.W., Nestor, P.G., Shenton, M.E., Pollak, S.D., Penhune, V., Mondrow, E., Marcy, B., Peterson, A., Hor- vath, T. and Davis, K.L. P300 topographic asymmetries are pres- ent in unmedicated schizophrenics. Electroenceph. clin. Neuro- physiol., 1993, 88: 32-41.

Fender, D.H. Source localization of brain electrical activity. In: A.S. Gevins and A. R6mond (Eds.), Methods of Analysis of Brain Electrical and Magnetic Signals. Elsevier, Amsterdam, 1987: 355 -403.

Folstein, M.F., Folstein, S.E. and McHugh, P.R. "Mini-mental state" - a practical method for grading the cognitive state of patients for the clinician. J. Psychiat. Res., 1975, 12: 189-198.

Ford, J.M., Rosenbloom, M.J., Sullivan, E.V. and Pfefferbaum, A. ERPs and brain structure: relationships across the adult age span in alcoholics and in a patient with herpes simplex encephalitis. In: C.H.M. Brunia, G. Mulder and M.N. Verbaten (Eds.), Event-Re- lated Brain Research (EEG Suppl. 42). Elsevier, Amsterdam, 1991: 342-354.

Goodin, D.S., Squires, K.C., Henderson, B.H. and Starr, A. Age-re- lated variations in evoked potentials to auditory stimuli in normal subjects. Electroenceph. clin. Neurophysiol., 1978, 44: 447-458.

Grossberg, S. Some psychophysiological and pharmacological corre- lates of a developmental, cognitive, and motivational theory. In: R. Karrer, J. Cohen and P. Tueting (Eds.), Brain and Informa- tion: Event-Related Potentials. Ann. NY Acad. Sci., 1984, 425: 58-151.

Halgren, E.J., Squires, N.K., Wilson, C.L., Rohrbaugh, J.W., Babb, T.L. and Crandall, P.H. Endogenous potentials generated in the human hippocampal formation and amygdala by infrequent events. Science, 1980, 210: 803-805.

Halgren, E., Stapleton, J.M., Smith, M. and Altafullah, I. Generators of the human scalp P3(s). In: R.Q. Cracco and I. Bodis-Wollner (Eds.), Evoked Potentials. Alan Liss, New York, 1986: 269-284.

Hansch, E.C., Syndulko, K., Cohen, S.N., Goldberg, Z.I., Potvin, A.R. and Tourtellotte, W.W. Cognition in Parkinson disease: an event-related potential perspective. Ann. Neurol., 1982, 11: 599- 607.

Johnson, R. Scalp-recorded P300 activity in patients following unilat- eral temporal lobectomy. Brain, 1988, 111: 1517-1529.

Knight, R.T. Neural mechanisms of event-related potentials: evi- dence from human lesion studies. In: J.W. Rohrbaugh, R. Para- suraman and R. Johnson, Jr. (Eds.), Event-Related Brain Poten- tials. Oxford Univ. Press, New York, 1990: 3-18.

Knight, R.N., Scabini, D., Woods, D.L. and Clayworth, C.C. Contri- butions of temporal-parietal junction to the human auditory P3. Brain Res., 1989, 502: 109-116.

McCarley, R.W, Shenton, M.E., O'Donnell, B.F., Faux, S.F., Kikinis, R., Nestor, P.G. and Jolesz, F.A. Auditory P300 abnormalities and left posterior temporal gyrus volume reduction in schizophre- nia. Arch. Gen. Psychiat., 1990, 50: 190-197.

McCarthy, G., Darcey, T.M., Wood, C.C., Williamson, P.D. and Spencer, D.D. Asymmetries in scalp and intracranial endogenous ERPS in patients with complex partial epilepsy. In: J. Engel, Jr. et al. (Eds.), Fundamental Mechanisms of Human Brain Func- tion. Raven Press, New York, 1987: 51-59.

402 B.F. O'DONNELL ET AL.

McCarthy, G., Wood, C.C., Williamson, P.D. and Spencer, D.D. Task-dependent field potentials in human hippocampal forma- tion. J. Neurosci., 1989: 4523-4268.

Meador, K.J., Hammond, E.J., Loring, D.W., Allen, M., Bowers, D. and Heilman, K.M. Cognitive evoked potentials and disorders of recent memory. Neurology, 1987, 37: 526-529.

Novak, G.P., Ritter, W., Vaughan, H.G. and Winitzer, M.L. Differ- entiation of negative event-related potentials in an auditory dis- crimination task. Electroencaph. clin. Neurophysiol., 1990, 75: 255-275.

Nunez, P.L. Electric Fields of the Brain: the Neurophysics of EEG. Oxford Univ. Press, New York, 1981.

O'Donnell, B.F., Squires, N.K., Martz, M.J., Chen, J.R. and Phay, A.J. Evoked potential abnormalities and neuropsychological per- formance in Parkinson's disease. Biol. Psychol., 1987, 24: 23-37.

O'Donnell, B.F., Friedman, S., Squires, N.K., Maloon, A., Drach- man, D.A. and Swearer, J.M. Active and passive P3 latency in dementia: relationship to psychometric, EEG and CT measures. Neuropsychiat. Neuropsychol. Behav. Neurol., 1990, 3: 164-179.

O'Donnell, B.F., Friedman, S., Maloon, A. and Drachman, D.A. P3 latency and neuropsychological performance: influence of age and individual differences. Int. J. Psychophysiol., 1992, 12: 187- 195.

O'Donnell, B.F., Shenton, M.E., McCarley, R.W., Faux, S.F., Smith, R.S., Salisbury, D.F., Nestor, P.G., Pollak, S.D., Kikinis, R. and Jolesz, F.A. The auditory N2 component in schizophrenia: rela- tionship to MRI temporal lobe grey matter and to other ERP abnormalities. Biol. Psychiat., 1993, 34: 26-40.

Onofrj, M., Gambi, D., Fulgente, T., Bazzano, S. and Colamartino, P. Persistence of a P3 component in severe amnestic syndrome. Electroenceph. clin. Neurophysiol., 1991, 78: 480-484.

Onofrj, M., Fulgente, T., Nobilio, D., Malatesta, G., Bazzano, S., Colamartino, P. and Gambi, D. P3 recordings in patients with bilateral temporal lobe lesions. Neurology, 1992, 42: 1762-1767.

Pfefferbaum, A., Ford, J.M., Wenegrat, B.G., Walton, W.T. and Kopell, B.S. Clinical application of the P3 component of event-re- lated potentials. I. Normal aging. Electroenceph. clin. Neurophys- iol., 1984a, 59: 95-103.

Pfefferbaum, A., Wenegrat, B.G., Ford, J.M., Roth, W.T. and Kopell, B.S. Clinical application of the P3 component of event-related potentials. II. Dementia, depression and schizophrenia. Elec- troenceph, clin. Neurophysiol., 1984b, 59: 104-124.

Pffefferbaum, A., Ford, J.M., White, P.M. and Rother, W.T. P3 in schizophrenia is affected by stimulus modality, response require- ments, medication status, and negative symptoms. Arch. Gen. Psychiat., 1989, 46: 1035-1044.

Polich, J. P300 in the evaluation of aging and dementia. In: C.H.M. Brunia, G. Mulder and M.N. Verbaten (Eds.), Event-Related Brain Potential Research (EEG Suppl. 42). Elsevier, Amsterdam, 1991.

Polich, J., Howard, L. and Starr, A. P300 latency correlates with digit span. Psychophysiology, 1983, 20: 665-669.

Polich, J., Ehlers, C.L., Otis, S., Mandell, A.J. and Bloom, F.E. P300 latency reflects the degree of cognitive decline in dementing illness. Electroenceph. clin. Neurophysiol., 1986, 63: 138-144.

Rogers, R.L., Bauman, S.B., Papanicolaou, A.C., Bourbon, T.W., Alagarsamy, S. and Eisenberg, H.M. Localization of the P3 sources using magnetoencephalography and magnetic resonance imaging. Electroenceph. clin. Neurophysiol., 1991, 79: 308-321.

Roth, W.T., Horvath, T.B., Pfefferbaum, A. and Kopell, B.S. Event- related potentials in schizophrenics. Electroenceph. clin. Neuro- physiol., 1980, 8: 127-139.

Rugg, M.D., Pickles, C.D., Potter, D.D. and Roberts, R.C. Normal P300 following extensive damage to the left medial temporal lobe. J. Neurol. Neurosurg. Psychiat., 1991, 54: 217-222.

Scherg, M. Fundamentals of dipole source potential analysis. In: F. Grandori et al. (Eds.), Auditory Evoked Magnetic Fields and Electric Potentials. Advances in Audiology, Vol. 6. Karger, Basel, 1990: 40-69.

Scherg, M. and Berg, P. BESA: Brain Electrical Source Analysis Handbook (Ver. 1.8). Neuroscan, Herndon, VA, 1991.

Scherg, M. and Von Cramon, D. Evoked dipole source potentials of the human auditory cortex. Electroenceph. clin. Neurophysiol., 1986, 65: 344-360.

Semlitsch, H.V., Anderer, P., Schuster, P. and Presslich, O. A solution for reliable and valid reduction of ocular artifacts, ap- plied to the P300 ERP. Psychophysiology, 1986, 23: 695-7(13.

Simson, R., Vaughan, H.G. and Ritter, W. The scalp topography of potentials in auditory and visual discrimination tasks. Electroen- ceph. clin. Neurophysiol., 1977, 42: 528-535.

Stapleton, J.M., Halgren, E. and Moreno, K.A. Endogenous poten- tials after anterior temporal lobectomy. Neuropsychologia, 1987, 24: 459-557.

Sutton, S., Braren, M., Zubin, J. and John, E.R. Evoked-potential correlates of stimulus uncertainty. Science, 1965, 150: l 187-1188.

Vaughan, H.G. and Arezzo, J.C. The neural basis of event-related potentials. In: T.W. Picton (Ed.), EEG Handbook (Revised Se- ries, Vol. 3). Elsevier, Amsterdam, 1988: 45-96.

Wechsler, D. A standardized scale for clinical use. J. Psychol., 1945. 19: 87-95.

Wechsler, D. The Wechsler Intelligence Scale - Revised Manual. Psychological Corp., New York, 1981.

Williamson, S.J. and Kaufman, L. Theory of neuroelectric and neu- romagnetic fields. In: F. Grandori et al. (Eds.), Auditory Evoked Magnetic Fields and Electric Potentials. Advances in Audiology, Vol. 6. Karger, Basel, 1990: 1-39.

Woods, D.L., Clayworth, C.C., Knight, R.T., Simpson, G.V. and Naeser, M.A. Generators of middle- and long-latency auditory evoked potentials: implications from studies of patients with bitemporal lesions. Electroenceph. clin. Neurophysiol., 1987, 68: 132-148.