Gating of auditory response in schizophrenics and normal controls

Schizophrenia Research, 4 Elsevier

SCHIZO

(1991) 3 l-40

31

00147

Gating of auditory response in schizophrenics and normal controls Effects of recording Herbert

T. Nagamoto,

site and stimulation

Lawrence

E. Adler, Merilyne Freedman

interval C. Waldo,

on the P50 wave Jay Griffith

and

Robert

Denver Veterans Administration Medical Cenler, and Departments of Psychiatry and Pharmacology, University of Colorado Health Sciences Center, Denver, CO 80262, U.S.A. (Received

30 April

1990, accepted

29 June

1990)

Auditory evoked potentials were recorded using a paired stimulus, conditioning-testing paradigm from 14 schizophrenic patients and 13 normal subjects with no family history of psychotic disorder. Previous studies of the vertex P50 wave using this paradigm have demonstrated a possible sensory gating deficit in schizophrenics, as shown by their failure to diminish the response to a test stimulus presented 500 ms after a conditioning stimulus. Recordings were made at Cz, Fz, C3, T3, C4, and T4, to compare effects at different recording sites with this paradigm. Schizophrenics had significantly poorer sensory gating than normals, with the most significant difference between the groups at Cz. In addition to the 500 ms interval, subjects were also recorded at a conditioning-testing interval of 100 ms. Most schizophrenics showed normal sensory gating at the 100 ms interval, despite their abnormalities at 500 ms. The results indicate that Cz is an optimal recording site for this paradigm, and that gating abnormalities in schizophrenic subjects are limited to specific interstimulus intervals. Key words; PSO; Auditory

evoked

potential;

(Schizophrenia)

INTRODUCTION

It has been hypothesized that problems with attention and perception in schizophrenia are the result of deficits in filtering or gating of sensory input that lead to hypervigilance and difficulties in discrimination (McGhie and Chapman, 1961; Roth and Cannon, 1972; Gruzelier, 1975; Landau et al., 1975; Zahn, 1976; Spohn et al., 1977; Roemer et al., 1979; Braff and Saccuzzo, 1982). Venables (1964) proposed that inability to gate or control responsiveness to sensory stimuli could lead to a Correspondence to: H.T. Nagamoto, Department of Psychiatry, C-268, University of Colorado Health Sciences Center, 4200 East Ninth Avenue, Denver, CO 80262, U.S.A. (Tel: (303)-270-5440).

0920-9964/91/$03.50

0

1991 Elsevier Science Publishers

psychotic disorganized state in which patients are ‘flooded’ by an overabundance of sensory stimulation. Clinical and psychometric data support the hypothesis that schizophrenics have disordered sensory gating mechanisms. They perform poorly on tests requiring focus of attention, such as the continuous performance test (Mirsky and Kornetsky, 1968) or smooth pursuit eye movement tasks (Holzman et al., 1973). Their poor sensory gating is further demonstrated by their diminished inhibition of acoustic startle in response to preceding warning tones (Braff et al., 1978). Sensory gating deficits in schizophrenics can also be demonstrated with the auditory P50 evoked potential, by a conditioning-testing paradigm often used to test local brain inhibitory pathways (Eccles, 1969). In this paradigm, subjects are presented with paired auditory stimuli. The first, or

B.V. (Biomedical

Division)

32

‘conditioning’ stimulus is hypothesized to activate inhibitory or other gating mechanisms that are tested by the second or ‘test’ stimulus. Normal control subjects have a decrease in the amplitude of the P50 waveform in response to a test stimulus, whereas schizophrenic subjects show little or no decrement of P50 amplitude to the test stimulus. This decreased gating of the P50 response appears to be a stable trait in schizophrenics, present whether unmedicated and acutely ill (Adler et al., 1982) or medicated and clinically stable (Freedman et al., 1983). This deficit is not exclusive to schizophrenics, as it also appears in the acute phases of a variety of psychiatric illnesses such as mania (Franks et al., 1983; Baker et al., 1987a) and in a substantial portion of the first degree relatives of schizophrenics (Siegel et al., 1984). The aim of this study was to extend this finding in two ways: (I) to consider the behavior of P50 in the conditioning-testing paradigm at different scalp locations, and (2) to examine the effects of shorter interstimulus intervals. In our previous studies using the conditioningtesting paradigm, we have chosen to measure the P50 wave, as it is less affected by changes in the subject’s interest than other auditory waves (Hillyard et al., 1973). The vertex location was chosen for these investigations, because P50 has a large amplitude at Cz (Wood and Wolpaw, 1982) and Cz is less subject to muscle contamination than other scalp leads. While a vertex recording site has been successfully used in this paradigm, it is not clear whether it is the optimal recording site to use to demonstrate this deficit in schizophrenics. Therefore, one aim of this study was to compare recordings at vertex with other scalp leads. Presently, animal and human data suggest that P50 recorded from the scalp is the summation of multiple intracranial sources. These include superficial temporal (Knight et al., 1988; Reite et al., 1988) hippocampal (Woods et al., 1987; Bickford-Wimer et al., 1989) and thalamic (Hinman and Buchwald, 1983; Chen and Buchwald, 1986) sites. For a lateral temporal source, optimal recordings might be obtained from a temporal scalp lead. For a deeper source, such as hippocampus or thalamus, a temporal lead might not offer any advantages over vertex. Another consideration is the geometry of the source in addition to its location. Currents radial to a particular electrode

generate waves of much greater amplitude than those that are tangential (Vaughan, 1974). Although the scalp surface distribution of P50 is known (Wood and Wolpaw, 1982) the possibility of several intracranial generators makes it necessary to examine its gating at different sites. Different intracranial generators might behave differently in the conditioning-testing paradigm, with some showing more suppression of the test response than others. For this study we chose Cz, Fz, C3, T3, C4, and T4 as recording leads, to examine if differences between schizophrenics and normals in sensory gating were more prominent in temporal leads, central leads. or at the vertex. We compared the spatial distribution of the P50 response itself, as measured by its amplitude and latency in the initial conditioning response, with the distribution of the sensory gating mechanisms that differentiate normals from schizophrenics, which we measured as the ratio of test to conditioning response amplitudes. In our previous studies, we demonstrated that schizophrenics have a deficit in sensory gating at intervals of 150-2000 ms (Adler et al., 1982) but not at 75 or 100 ms (Nagamoto et al., 1989). We have investigated these different conditioning-testing intervals for two reasons. First, it is likely that there are different neuronal mechanisms operating at varying intervals that contribute to sensory gating (Nagamoto et al., 1989). Gating at the 100 and 75 ms intervals is likely to reflect inhibitory activity of local GABAergic interneurons (Eccles. 1969); gating at longer intervals may reflect multisynaptic mechanisms such as inputs from the brainstem (Hinman and Buchwald, 1983). Second, a few schizophrenics show normal gating at the 500 ms interval but poor gating at the 75 and 100 ms intervals. Patients with this less common sensory gating pattern might comprise a subgroup whose sensory gating deficit has a different underlying neuronal mechanism. In this study, we therefore further extended our comparison of conditioning-testing intervals of 500 and 100 ms.

METHODS

15 schizophrenic subjects were recruited from clinical services of the Denver Veterans Administration

33

Medical Center and the Colorado Psychiatric Hospital. All of these subjects were clinically stable medicated with neuroleptics. 13 outpatients, patients were male and two were female. All patients fulfilled DSM-IIIR criteria for active schizophrenia and did not have current or recent histories of alcoholism or drug abuse. Diagnoses were made by board certified psychiatrists familiar with the patient. If the patient had not been followed for at least a year, a Schedule for Affective Disorders and Schizophrenia-Lifetime Version (SADS-L) interview was performed. 14 normal control subjects were recruited from the staffs of the Denver Veterans Administration and Colorado Psychiatric Hospitals. 12 controls were male and two were female. All controls were screened by a psychiatrist. Exclusion criteria for controls included a history of mental illness, drug or alcohol abuse, use of psychotropic medications, or a family history of mental illness as determined by methods adapted from Baker et al. (1987b). The mean age of the patients (33.6 f 6.5 years, range 24440 years old) was not significantly different from that of controls (33.6f 5.5 years, range 24-40 years old). Subjects with mental retardation were excluded. Education level was significantly higher for controls (18.5 +2.0 years) than for schizophrenics (13.4 f 1.4 years). Subjects were tested for handedness with the D’Elia dominance test battery (D’Elia, 1970). 13 controls and ten schizophrenics were right handed, one control and three schizophrenics showed mixed dominance, and no controls and two schizophrenics were left handed. The distribution of handedness was not significantly different between the two groups by chi-squared test (x2 = 3.4 n.s.). Recording methods All recordings were obtained with subjects supine and relaxed, but awake with eyes focused on a spot on the ceiling above. Monopolar recordings were made with gold-disc electrodes affixed to the scalp with electrode paste. Recording electrodes were placed at CZ, C3, C4, and 0.5 cm above T3 and T4. The modified T3/T4 positions were used to decrease temporal muscle artifact that tended to obscure P50. These recording electrodes were referenced to the tip of the nose (S.F. Faux, personal communication) with a forehead ground. Noncephalic references could not be used because of

excessive muscle artifact in schizophrenic subjects that caused difficulty in resolving P50. Electrode resistance was less than 10 kSZ. The evoked electroencephalogram (EEG) activity was amplified 20,000 times by a 22 MS2 impedance amplifier with filters ( - 50%) at 1.O and 300 Hz (model 7~511, Grass Instruments, Quincy, MA). The EEG was monitored on an IBM AT-compatible computer (Epson Equity III, Epson America, Torrence, CA) with a 7 channel analog interface (Data Translation DTlOl, Data Translation, MA). Data was acquired at a 1000 Hz digitization rate. Individual trials were rejected when there was evidence of muscle activity or other artifacts. The electrooculogram (EOG) from the superior orbital ridge referenced to the lateral canthus was also recorded. Auditory stimulation A 0.04 ms duration pulse was amplified with a bandwidth from 20 to 12,000 Hz and was delivered through an 8-inch loudspeaker placed above the subject’s head, directed parallel to the body. Peak intensity was 110 dB sound-pressure level (SPL), as measured at the subject’s ear by a sound meter (Tandy Corporation, Houston, TX) monitored on an oscilloscope for impulse SPL readings. For two normal control subjects who showed a startle response to the clicks, the mean intensity was decreased by 5 dB. The recording room was quiet, but not acoustically isolated. Conditioning-testing paradigm Stimuli were presented as a series of pairs. of conditioning and testing clicks. The interval between conditioning stimuli was 10 s. Three averaged evoked responses to 32 pairs of stimulus presentations were obtained from subjects at a conditioning-testing interval of 500 ms. In addition to the 500 ms interval recordings, 12 control and nine schizophrenic subjects were also recorded at a 100 ms conditioning-testing interval. Response analysis The averaged auditory evoked responses were analyzed by a computer using an algorithm described in detail in a previous paper (Nagamoto et al., 1989). Averaged evoked responses were first reviewed to determine that they were free of muscle, EOG, or other artifacts, and contained a recognizable conditioning P50 wave. Grand averages of the

34

evoked responses at each recording interval were digitally filtered twice using a low pass filter that averaged each point with the 3 points before and after, yielding a 70% reduction of frequencies above 250 Hz. The computer then selected the conditioning P50 response as the most positive peak between 40 and 90 ms after the conditioning stimulus. Amplitude was measured relative to the preceding negativity. In previous studies, test responses were read if the test latency was within 10 ms of the conditioning latency, to help insure that paired responses represented the same waveform. To verify the basis for this criterion, test responses in this study were paired with the conditioning response that was closest in latency, and the difference in conditioning and test P50 latency was calculated. This difference in conditioning and test P50 latency was not calculated for tracings in which a test P50 wave was not present. The latency differences between conditioning and testing waves at each interval and lead for both groups are presented in Table 1. The 10 ms criterion is adequate for recordings at Cz at the 500 ms conditioning-testing interval, but needs to be increased slightly for other leads. Differences in latency are greater when the stimuli are closer together, as has been found by others (Surwillo, 1981). The conditioning-testing paradigm ratio, expressed as a percentage, is the test amplitude divided by the conditioning amplitude, then multiplied by 100. For data analysis, conditioningtesting paradigm ratios above 200% were truncated to 200%, to prevent outliers from having a disproportionate effect on group means. ArtifactTABLE

1

Latency

differences

between conditioniq

and test PSO responses

Values given are range; n, mean k standard deviation. Cases in which a test wave was entirely absent are not included in this analysis. Site

Conditioning-testing

free recordings at the 500 ms conditioning-testing interval were obtained in all but one control and one schizophrenic subject, leaving 12 male and one female controls and 12 male and two female schizophrenics. To obtain the grand averages of the evoked potentials obtained at each recording lead at the 500 ms interval shown in Fig. 1, the condition and test responses were shifted in time, so that for each subject the latency of the Cz conditioning response was at the midpoint between the shortest and longest latencies in their diagnostic group. This procedure was used because of the high intersubject variance in response latency, particularly among the schizophrenics, which obscured the responses in the population grand averages. For the 100 ms conditioning-testing interval, a subtraction procedure was used to remove the effect of any later waves from the first response that overlapped with the second response (Schwartz and Shagass, 1963). This procedure was not used for the 500 ms conditioning-testing interval, since there are no prominent waves from the first response occurring 500 ms later (Callaway, 1975). The conditioning response of the 500 ms interval was subtracted from the potentials recorded at the 100 ms interval, as illustrated in Fig. 2. In our previous study using conditioningtesting intervals of 75 and 150 ms, subtraction changed test P50 amplitude in only one case. In this study, subtraction had a significant effect in the majority of subjects (eight of 11 controls and six of eight schizophrenics). Subtraction was helpful when a large Pl80 response after the first stimulus obscured the P50 response from the test stimulus presented at 100 ms. Artifact-free recordings were obtained for all but one of the controls recorded at the 100 ms conditioning-testing interval, leaving ten males and one female. Of the nine schizophrenics recorded at 100 ms, six males and two females had artifact-free recordings.

interval

RESULTS 100 ms

500 ms CZ FZ c3 T3 c4 T4

O-10 O-12 O-9 O-l I O-12 I-10

ms; ms; ms; ms; ms; ms;

25, 21, 23, 20, 24, 22.

3.8k2.9 4.1 k3.0 3.1 +2.5 4.6+3.1 3.8k2.6 4.8k2.9

O-II 1-12 3-14 l-16 O-17 1-12

ms; ms; ms; ms; ms; ms;

13, 14, 13, 11, 13, 11,

5.9k3.4 7.Ok2.9 7.3k3.2 6.6i4.8 7.Ok4.6 6.61.3.4

Normal controls showed significant suppression of the vertex auditory P50 response in the conditioning-testing paradigm with a 500 ms interval between stimuli, while schizophrenic subjects showed significantly less suppression, as reported in previ-

35

Normal controls

Schizophrenics EOG

__15llV 50ms Fig. 1. Grand averages of the auditory evoked response in a conditioning-testing paradigm in 13 normal controls (top) and 14 schizophrenic subjects (bottom), recorded at Cz, Fz, C3, T3, C4, and T4. Paired condition (C) and test (T) averages at each lead are presented. The P50 wave in normal controls is significantly decreased in the second (test) response when compared with the first (conditioning) response. The schizophrenics fail to show significant decrement in amplitude of the P50 waveform from the conditioning to the test response. Amplitude was measured by computer from the P50 wave peak (cursor below each trace) to the preceding negative trough (cursor above each trace). The auditory stimuli occur at the beginning of each trace; the conditioning-testing interval was 500 ms. Grand averages of the electrooculogram (EOG) responses to the clicks for each group are also presented. There was no significant EOG contribution to the P50 waveforms for either group. The EOG cursors mark the site of the Cz PSO waveforms. Head display for this figure adapted from and used with permission of Robert Knight, M.D. Vertical calibration is 5 PV (positive down), horizontal 50 ms.

ous studies. Normal controls had a low mean conditioning-testing paradigm ratio of 28.5% f 4.5 SEM, indicative of intact gating of the P50 response. Schizophrenic subjects had diminished

Fig. 2. Effect of subtraction procedure on auditory evoked potentials in a normal control. Top tracing shows conditioning (left) and test (right) averaged responses to auditory clicks (arrows) played 500 ms apart (CTSOO). The test P50 response (marked by computer cursors) is diminished in amplitude, compared to the conditioning response. For the 100 ms interval, a subtraction procedure was used to remove waves in the conditioning response that overlapped with the test response. CT100 is the unsubtracted response to the conditioning stimulus and test stimulus at 100 ms. A large P180 response to the conditioning stimulus (asterisk) obscures the test CT100 response. The CT500 response was subtracted from the CT100 response to remove the effect of the conditioning response on the CT100 test response, yielding the bottom tracing. The bottom tracing shows the subtracted test response, with its test P50 marked by computer placed cursors. As with the 500 ms response, this subject shows significant decrement of the test 100 ms response compared to the conditioning response. Calibration is 5 PV (positive down) and 50 ms.

sensory gating, with a significantly higher mean conditioning-testing ratio of 94.7% f 16.1 SEM (ANOVA, F=14.6, df(1,25), P
Conditioning-Testing 3

5

500

Interval:

msec

100

msec

200

1PL i l-d-id 100

7

T3

c3

C4

0

T4

LEAD

FZ

c3

73

c4

14

LEAD

Fig. 3. Mean ratio of test P50 response to conditioning response amplitude (%, f SEM) for each group at conditioning-testing intervals of 500 ms (left) and 100 ms (right). At the 500 ms interval these ratios are significantly higher for schizophrenics than normals at Cz and C4 by ANOVA with Bonferroni’s correction (alpha set at 0.0083). These differences approached significance at Fz and C3, but at the T3 and T4 leads at 500 ms and all leads at 100 ms. schizophrenics were not significantly different from normal controls. ***. P
P>O.2), and T4 (F= 1.1, df(1,24), P>O.3). In our previous work, 95% of normal controls have had a conditioning-testing paradigm ratio of 50% or less. At Cz, using this definition of intact sensory gating (Siegel et al., 1984) one control showed decreased gating and two schizophrenics showed intact gating at the 500 ms interval. Using this criterion, 24 of 27 subjects, or 88.9%, were correctly grouped by diagnosis, which was significant by chi-squared test (x2= 13.5, df= 1, P
500 ms interval, and two others showed decreased gating at the 100 ms interval. Of the schizophrenics, two showed intact gating at 500 ms, while three had a gating deficit at 100 ms. In this study, using stable outpatient schizophrenics, conditioning P50 amplitude was not significantly different between patients and controls at any of the leads recorded (ANOVA with Bonferroni correction for multiple comparisons). Conditioning P50 latency was also not significantly different between the groups. The distribution of the amplitude and latency of the conditioning P50 response for each group is presented in Table 2. While P50 was maximal in amplitude at Cz for the control group and in the temporal leads for schizophrenics, the differences between recording leads were not statistically significant. An analysis of variance of amplitude and latency did not show any significant lead effects for controls, schizophrenics, or both groups combined. The electrophysiological data were also analyzed with respect to BPRS scores, total and subscales by Spearman’s rank order correlations, with no significant effect. While the normal controls had significantly higher number of years of formal education than the schizophrenics, this parameter was

31

TABLE

2

Condirioning

PSV ctmplitude rend luterq

Values given are: amplitude, dard deviation, respectively. 500 ms. S/c

cz FZ C3 T3 C-4 T4

by recording .slte

pV: latency, ms; both mean k stanConditioning-testing interval was

Normul c~onrrols

Schi:ophrenics

Amp/i&de

n

LUtetIc>

Amplitude

n

LUtWKJ

2.5* 1.2 1.9* I.1 l.6iI.I 1.4* I.1 2.0* 1.3 2.0+0.8

13 13 I3 I2 I3 I2

54k3.7 55k5.4 53k4.9 52+4.1 52k3.9 5li3.5

2.2& 1.3 2.2? 1.2 1.7*1.1 2.4& I.5 1.7+ 1.0 2.4* 1.4

14 I4 I4 I4 I4 I4

s7+7.2 58k6.7 56110.1 56k9.9 57k8.4 55i9.3

not correlated with any of the P50 parameters measured. Hemispheric dominance, as measured by the D’Elia handedness battery, did not appear to be related to any of the P50 parameters measured. There also did not appear to be any significant right/left hemispheric differences in P50 measurements, and hemispheric dominance was not related to any hemispheric differences.

DISCUSSION

The data in this study are consistent with our previous demonstration of a sensory gating deficit in schizophrenics in the conditioning-testing paradigm. In this paradigm, subjects are presented with pairs of auditory clicks 500 ms apart; normal controls suppress the P50 response to the second or test stimulus, while schizophrenics do not. This phenomenon has been demonstrated in a series of studies using the vertex P50 response (Adler et al., 1982, 1985; Freedman et al., 1983, 1987; Baker et al., 1987a; Braff and Judd, 1987; Nagamoto et al., 1989). In this study, we have addressed two major methodological questions: which recording sites and which interstimulus intervals optimally differentiate schizophrenics from normals with this paradigm? Compared to Fz, C3, C4, T3, and T4, the vertex site better differentiates schizophrenics from normal controls with regard to sensory gating as measured by the conditioning-testing ratio. There was little difference between the groups at T3/T4.

Although the P50 amplitude and latency in the initial or conditioning response did not show any significant differences between recording sites, gating of P50 was significantly different. The data presented here would seem to be consistent with the hypothesis that P50 has multiple intracranial sources. The recordings obtained at the temporal leads may primarily reflect the activity of a source that shows similar gating patterns in normal controls and schizophrenics. One candidate for this source is superior temporal gyrus, as suggested by magnetoencephalography (Reite et al., 1988) and AEP recordings obtained from stroke patients (Knight et al., 1988). Recordings from the vertex may primarily reflect the activity of a different source. which exhibits intact sensory gating in normal controls and a gating deficit in schizophrenics. A likely candidate for such a generator would be a deeper source such as hippocampus (Woods et al., 1987; Bickford-Wimer et al., 1989) or thalamus (Hinman and Buchwald, 1983; Chen and Buchwald, 1986). The second issue addressed in this study is which conditioning-testing interval is optimal for this paradigm and what further information might be gained at other intervals. It is likely that there are different neuronal mechanisms operating at various interstimulus intervals that contribute to sensory gating. At intervals between stimuli of I ms or less, neurons do not respond at all to a test stimulus because of membrane conductance changes during the absolute refractory period. From I ms to approximately 50 ms, the response to the second stimulus is greatly increased (Morison and Dempsey, l942), perhaps due to facilitation of activity of excitatory synapses, similar to that seen in tetanic stimulation at the neuromuscular junction (Katz, 1969). It has been shown that schizophrenic subjects have decreased facilitation of somatosensory input (Shagass and Schwartz, 1963) and visual input (Speck et al., 1966; Ishikawa, 1968). At intervals of approximately 50&150 ms, the response to the second stimulus is decreased compared to the first stimulus (Ishikawa, 1968), possibly due to activation of local inhibitory neurons, such as the GABAergic neurons of the cerebral cortex, by the first stimulus (Eccles, 1969). Sensory gating at longer intervals, such as the 500 ms interval studied here, may be due to activation of multisynaptic afferents, pos-

sibly from areas such as the reticular formation and the nonspecific thalamic nuclei (Lindsley and Adey, 1961; Fuster and Dotter, 1962; Schwartz and Shagass, 1963). In previous studies, we have found mean conditioning-testing ratios to be significantly different between schizophrenics and controls at intervals from 1SO to 2000 ms (Adler et al., 1982; Freedman et al., 1982; Nagamoto et al., 1989). In this range, the most significant differences in group means were at the 500 ms interval. We have also found that normal controls and schizophrenics were not significantly different at an interval of 75 ms (Nagamoto et al., 1989) and in this study we report that the same is true at an interval of 100 ms. Thus, the data are most consistent with a defect in multisynaptic gating mechanisms. While most schizophrenics show diminished gating at an interval of 500 ms and normal gating at an interval of 75 or 100 ms, there are exceptions at each interval. We have hypothesized that these exceptions might reflect the long-suspected biological diversity in the pathophysiology of schizophrenia (Bleuler, 1950; Nagamoto et al., 1989). There may be a common deficiency in sensory gating in most schizophrenics, but possible differences in the neuronal mechanisms underlying these deficiencies. In examining these patients, we have looked for clinical characteristics that might be correlated with their unusual sensory gating patterns. There was one patient in the first study and one in the present study with normal gating at the 500 ms interval but abnormal gating at the shorter interval. These two subjects had predominantly negative symptoms, but other predominantly negative patients had deficits only at the 500 ms interval. A third subject was normal at both intervals; the most unusual characteristic about this patient was that she had a family history strongly positive for affective disorders. In addition, there were four patients evenly divided between the two studies with abnormal gating at both intervals. In the first study, these two patients were the only acutely ill subjects, in the second study the two patients had BPRS scores that were in the highest range. It is possible that gating at the 75 and 100 ms intervals is state dependent in some schizophrenics, worsening with increased symptomatology. One possible mechanism might be increased noradrenergic neurotransmission. Norepinephrine interferes with in-

hibitory GABAergic mechanisms that could underly sensory gating at the 75 and 100 ms intervals (Kehl and McLennan, 1985). Plasma MHPG measurements suggest that noradrenergic metabolism is increased in acutely ill schizophrenics (Ko et al., 1988) which could cause this aspect of disordered gating. The principal rationale for this study was to improve the methodology of an electrophysiological measurement that was previously shown to distinguish schizophrenics from normals. A problem with evoked potentials is that they do not unambiguously indicate underlying neuronal mechanisms. Specification of the scalp surface distributions and of the effects of changes in interstimulus intervals in themselves contribute only limited additional information. However, because of the inaccessibility of direct measurements of neuronal physiology in humans, such data, acquired for methodological reasons, may also have some use in the assessment of candidate mechanisms for phenomena like sensory gating.

ACKNOWLEDGEMENTS

The authors would like to thank Gary Zerbe, Ph.D. and Karen Stevens, Ph.D. for statistical consultation, and Robert Knight, M.D. for use of the head display for Fig. 1. Supported by USPHS Grants MH-38231 and MH-44212 and the VA Medical Research Service.

REFERENCES

Adler, L.E.. Pachtman, E.. Franks, R.D.. Pecevich, M., Waldo, M.C. and Freedman, R. (1982) Neurophysiological evidence for a defect in neuronal mechanisms involved in sensory gating and schizophrenia. Biol. Psychiatry 17, 639-654. Adler, L.E., Waldo, M. and Freedman, R. (1985) Neurophysiological studies of sensory gating in schizophrenia: Comparisons of auditory and visual responses. Biol. Psychiatry 20, 1284-1296. Baker, N., Adler, L.E.. Franks, R.D., Waldo. M.C., Berry, S., Nagamoto, H., Muckle, A. and Freedman, R. (1987a) Neurophysiological assessment of sensory gating in psychiatric inpatients; Comparison between schizophrenia and other diagnoses. Biol. Psychiatry 22, 6033617. Baker, N.J., Berry, S.L. and Adler. L.E. (1987b) Family

39 diagnoses missed on a clinical inpatient service. Am. J. Psychiatry 144, 630-632. Bickford-Wimer, P.C., Nagamoto, H., Johnson, R., Adler, L.E., Egan, M., Rose, G.M. and Freedman, R. (1989) Auditory sensory gating in hippocampal neurons: A model system in the rat. Biol. Psychiatry 27, 1833192. Bleuler, E. (1950) Dementia Praecox or the Group of Schizophrenias. Intl. Univ. Press, New York. Braff, D. and Judd, L. (1987) Regional erp/attention deficits in schizophrenia. Abstracts from Annual Meeting of the Society of Biological Psychiatry, 109. Braff, D. and Saccuzzo, D. (1982) Effect of antipsychotic medication: Speed of information processing in schizophrenic patients, Am. J. Psychiatry 139, 1127-l 130. Braff, D.. Stone, C.. Callaway, E., Geyer, M.. Glick, I. and Bali, L. (1982) Prestimulus effects on human startle reflex in normals and schizophrenics. Psychophysiology 15, 3399343. Callaway, E. (1975) Brain Electrical Potentials and Individual Psychological Differences. Grune and Stratton, Orlando, FL. Chen, B.M. and Buchwald, J.S. (1986) Midlatency auditory evoked responses: Differential effects of sleep in the cat. Electroencephalogr. Clin. Neurophysiol. 65, 373-382. D’Elia, G. (1970) Unilateral Electroconvulsive Therapy. Acta Psychiatr. Stand. 2!5 (Suppl.), 64465. Eccles, J.C. (1969) The Inhibitory Pathways of the Central Nervous System. University Press, Liverpool. Franks, R.D., Adler, L.E., Waldo, M.C., Alpert, J. and Freedman, R. (1983) Neurophysiological studies of sensory gating in mania: Comparison with schizophrenia. Biol. Psychiatry 18, 98991005. Freedman, R., Adler, L.E., Waldo, M.C., Pachtman, E. and Franks, R.D. (1983) Neurophysiological evidence for a defect in inhibitory pathways in schizophrenia: Comparison of medicated and drug-free patients. Biol. Psychiatry 18, 537-551. Freedman. R., Adler, L.E., Gerhardt, G.A., Waldo, M., Baker, N., Rose, G.M., Drebing, C., Nagamoto, H., BickfordWimer, P. and Franks, R. (1987) Neurobiological studies of sensory gating in schizophrenia. Schizophr. Bull. 13, 669-678. Fuster. J.M. and Dotter, R.F. (1962) Variation of optic evoked potentials as a function of reticular activity in rabbits with chronically implanted electrodes. J. Neural Transm. 25, 3244336. Gruzelier, J.H. (1975) The cardiac responses of schizophrenics to orienting, signal, and non-signal tones. Biol. Psychiatry 3, 143-155. Hillyard, S.A., Hink, R.F., Schment, V.L. and Picton, T.W. (1973) Electrical signs of selective attention in the human brain. Science 181, 1777180. Hinman, C.L. and Buchwald, J.S. (1983) Depth evoked potential and single unit correlates of vertex midlatency auditory evoked responses. Brain Res. 264, 57-67. Holzman, P.S., Proctor, L.R. and Hughes, D.W. (1973) Eyetracking performance in schizophrenia. Science 181, 179% 18 I. Ishikawa, K. (1968) Studies on the visual evoked responses to paired light flashes in schizophrenics. Kurume Med. J. 15, 153. Katz, B. (1969) Release of Neural Transmitter Substances. University Press, Liverpool. Kehl, S.J. and McLennan, H. (1985) A pharmacological char-

acterization of chlorideand potassium-dependent inhibitions in the CA3 region of the rat hippocampus in vitro. Exp. Brain Res. 60, 309-317. Knight, R.T., Scabini, D., Woods, D.L. and Clayworth, C. (I 988) The effects of lesions of superior temporal gyrus and inferior parietal lobe on temporal and vertex components of the human AEP. Electroencephalogr. Clin. Neurophysiol. 70, 4999509. Ko, G.N., Jimerson, D.C., Wyatt, R.J. and Bigelow, L.B. (1988) Plasma 3-methoxy-4-hydroxyphenylglycol changes associated with clinical state and schizophrenic subtype. Arch. Gen. Psychiatry 45. 8422846. Landau, S.G., Buchsbaum, M.S.. Carpenter, W., Strauss, J. and Sacks, M. (1975) Schizophrenia and stimulus intensity control. Arch. Gen. Psychiatry 32, 123991245. Lindsley, D.F. and Adey, W.R. (1961) Availability of peripheral input to the midbrain reticular formation. Exp. Neurol. 4, 358. McGhie, A. and Chapman, J.S. (1961) Disorders of attention and perception in early schizophrenia. Br. J. Med. Psychol. 34, 103. Mirsky, A.F. and Kornetsky, C. (1968) The effect of centrally acting drugs on attention. In: D.H. Efron (Ed.), Psychopharmacology: A Review of Progress. USPHS, Bethesda. Morison, R.S. and Dempsey, E.W. (1942) A study of thalamocortical relations. Am. J Physiol. 35, 281-292. Nagamoto, H.T., Adler, L.E., Waldo, M.C. and Freedman, R. (1989) Sensory gating in schizophrenics and normal controls: Effects of changing stimulation interval. Biol. Psychiatry 25, 5499561. Reite. M.. Teale, P., Zimmerman, J.. Davis, K.. Whalen. J. and Edrich, J. (1988) Source origin of a 50-msec latency auditory evoked field component in young schizophrenic men. Biol. Psychiatry 24, 4955506. Roemer, R.A., Shagass, C., Straumanis. J. and Amadeo, M. (1979) Somatosensory and auditory evoked potential studies of functional differences between the cerebral hemispheres in psychosis. Biol. Psychiatry 14, 3577373. Roth, W.T. and Cannon, E.H. (1972) Some features of the auditory evoked response in schizophrenics. Arch. Gen. Psychiatry 27, 466-47 1. Schwartz, M. and Shagass, C. (1963) Reticular modification of somatosensory cortical recovery function. Electroencephalogr. Clin. Neurophysiol. 15, 2655271, Shagass, C. and Schwartz, M. (1963) Psychiatric correlates of evoked cerebral cortical potentials. Am. J. Psychiatry 119, 105551061. Siegel. C., Waldo, M.. Mizner, G., Adler, L.E. and Freedman, R. (1984) Deficits in sensory gating in schizophrenic patients and their relatives: Evidence obtained with auditory evoked responses. Arch. Gen. Psychiatry 41, 607-612. Speck, L.B.. Dim, B. and Merce, M. (1966) Visual evoked responses of psychiatric patients. Arch. Gen. Psychiatry 15, 59963. Spohn. H.E., Lacoursiere, R.B., Thompson, K. and Coyne, L. (I 977) Phenothiazine effects on psychological dysfunction in chronic schizophrenics. Arch. Gen. Psychiatry 34, 633-644. Surwillo, W.W. (1981) Recovery of the cortical evoked potential from auditory stimulation in children and adults. Dev. Psychobiol. 14. I.

40 Vaughan, H.G. (1974) The analysis of scalp-recorded brain potentials. In: R. Thompson and M.M. Patterson (Eds.). Bioelectric Recording Techniques. Part B, Electroencephalography and Human Brain Potentials. Academic Press, New York. Venables, P. (1964) Input dysfunction in schizophrenia. In: B.A. Maher (Ed.), Progress in Experimental Personality Research. Academic Press, Orlando, FL, pp. I-47. Wood, C. C.. and Wolpaw, J.R. (1982) Scalp distribution of human auditory evoked potentials. II. Evidence for overlap-

ping sources and involvement of auditory cortex. Electroencephalogr. Clin. Neurophysiol. 54. 25-38. Woods, D.L., Clayworth, C.C.. Knight, R.T., Simpson. G.V. and Naeser, M.A. (1987) Generators of middle- and longlatency auditory evoked potentials: implications from studies of patients with bitemporal lesions. Electroencephalogr. Clin. Neurophysiol. 68, 1322148. Zahn, T.P. (1976) On the bimodality of the distribution of electrodermal orienting responses in schizophrenic patients. J. Nerv. Ment. Dis. 162. 1955199.