- Email: [email protected]
Selective attention differentially affects brainstem auditory evoked potentials of electrodermal responders and nonresponders
227
Psychiatry Research, 16, 227-232
Elsevier
Selective Attention Differentially Affects Brainstem Auditory Evoked Potentials of Electrodermal Responders and Nonresponders Werner
Sommer
Received September 1985.
4, 1984; revised version received January 25. 1985; accepted February 13,
Abstract. Brainstem auditory evoked potentials (BAEPs) were recorded in adult male chronic schizophrenics on maintenance neuroleptic therapy, drug-free alcoholics, and normal subjects. Subjects were subdivided into electrodermal responders (Rs) and nonresponders (NRs). Attention was directed toward auditory or visual stimuli. Independent of diagnosis, latencies of BAEP wave V were longer in NRs when visual rather than auditory stimuli were attended to, while there was no task effect for Rs. This finding is interpreted as a sign of excessive selective filtering of auditory stimuli in NRs. In the alcoholic NRs, wave I-V conduction times were longer than those in any other subgroup, possibly indicating retarded neural transmission. Key Words. Alcoholism, brainstem auditory evoked nonresponding, schizophrenia, selective attention.
potentials,
electrodermal
Many schizophrenic patients are electrodermal nonresponders (NRs) to orienting stimuli (Bernstein et al., 1982). When the same stimuli are made task-relevant, however, former NRs become responders (Rs) (Gruzelier and Venables, 1973). Thus, nonresponding cannot be easily accounted for as a simple sensory deficit; more complex abnormalities of information processing must be implicated. In human information-processing studies, event-related brain potentials (ERPs) have proved to be a valuable tool (Callaway et al., 1978). While short latency ERPs reflect the encoding of stimuli in the afferent sensory systems, the later components are related to the subjective evaluation of the prior input. To date, however, no published study of which I am aware has addressed itself to ERPs in subjects differing in their Electrodermal Response Type. In this investigation, ERPs were recorded in the same subjects from the early brainstem auditory evoked potentials (BAEPs) up to the long latency, endogenous components, varying the direction of attention within repeatedly presented sets of auditory and visual stimuli. Results concerning the late ERP components have been published elsewhere (Cohen et al., 1981). In this article, BAEP correlates of electrodermal responding and nonresponding are described.
Dr. Werner Sommer, Dip]. Psych., is at Universiik Konstanz, FG Psychologie, Postfach Konstanr, W. Germany. (Reprint requests to Dr. W. Sommer at the address above.) 0165-1781~85:
$03.30 0 1985 Elsevier Science Publishers
B.V.
5560, D-7750
228 Methods The subjects, all males, were 26 nonparanoid schizophrenics, 22 chronic alcoholics, and 22 healthy employees of a German psychiatric state hospital. The alcoholics had been abstinent for at least 3 weeks and were off tranquilizers. Twenty-two of the schizophrenics were receiving maintenance neuroleptic therapy. Psychiatric classification was based on the consensus of two experienced psychiatrists and DSM-IZI(American Psychiatric Association, 1980) criteria. The mean duration of hospitalization was 8 months for the alcoholics and 4.4 years for the schizophrenics. All the subjects had participated in a similar experiment a few days before in which skin conductance orienting responses (SCORs) and long latency ERPs were recorded. The groups were matched for age (mean q 41 years) and educational status. As it was the aim of this study to investigate the processing of auditory stimuli in such Rs and NRs as usually studied in psychophysiological research, no attempt was made to select the subjects or to adjust the stimulus intensity levels according to the psychometrical “hearing” performance of the subjects. To confirm subjects’ electrodermal response types, five tones (7%dB, I-kHz, l-second duration, 25-ms rise time) were presented at intervals between 25 and 40 seconds. The subjects were classified as Rs or NRs depending on whether they showed any SCORs greater than 0.05 micromho to the first three stimuli. All subjects showed shifts in skin conductance in the social exchange before and after the experiment. For the BAEPs, bursts of 10 clicks (0. I-ms duration, lOO-ms interstimulus interval, 650-ms interburst interval) were presented binaurally. The intensity of the rare (10%) click bursts was always 50 dB, while the frequent bursts were either of 70 dB (high intensity condition) or of 30 dB (low intensity). The intensity of the clicks was related to the threshold of a listener with normal hearing. Simultaneously, sets of light bars of l-second duration were presented on a screen with the same temporal distribution as the rare click bursts. In counterbalanced order, the subjects either counted the rare click bursts (count clicks) or the light stimuli (count lights). The number of targets counted had to be reported at the end of each task. The electroencephalogram (EEG) was recorded with Ag/AgCl electrodes from vertex referred to linked mastoids and differentially amplified (x 100,000, bandpass: 10 to 3,000 Hz, 6 dB rolloff/ octave). The responses to thefrequent clicks were sampled at 10 kHz for 10 ms after stimulus onset and averaged separately for each task, for each intensity condition, and for odd and even stimuli within the click bursts (cf. Fig. 1). For peak detection, the averages from odd and even stimuli were combined, resulting in averages of 2,000 and 4,000 responses to high- and low-intensity stimuli, respectively. The latencies of waves I and V were determined by visual inspection wherever possible, and analyses of variance (ANOVAs) were calculated for the two latency measures, as well as for the difference between them (wave I-V interpeak conduction time). To correct for nonhomogeneity of cell variance, ANOVAs were based on log-transformed data; the means, however, are reported untransformed. For a posteriori tests @ < 0.05), the Neuman Keuls procedure was used.
Results The proportion of NRs to Rs was I 1 to 15 in schizophrenics, 6 to 16 in alcoholics, and 4 to 17 in normal subjects. In comparison to the first SCOR test, this is an increase of four NRs in the normal subjects, two in alcoholics, and one in schizophrenics. The test/retest correlation for the global R/ NR dichotomy amounts to a 0 coefficient of 0.47 (x2= 15.37; df= l;p
229 Fig. 1. BAEP waveforms in a typical subject COUNT
7066
Y
2 i LJ
30dB
11
’
0
2
”
“1’ 4
1
6
8
I
IO mset
Frequent 30 dB In = 2,000) and 70 dB (n = 1.000) clicks ware heard while either rare auditory or visual stimuli were to be counted. BAEPs were averaged separately for odd and even stimuli (solid and dashed lines) within click bursts.
The ANOVA of the counting results with factors Diagnosis (schizophrenics, alcoholics, normal subjects), Eectrodermal Response Type (NRs, Rs) and repeated measures on Task (count lights, count clicks) and Intensity (30 dB, 70 dB) revealed a main effect of Diagnosis (F= 6.0 1; df = 2, 59; p < 0.0 l), with schizophrenics deviating from the correct result by 11.5% while alcoholics and normal subjects had only a 3.3% and 4.6% average deviation, respectively. Even though wave V could be determined most reliably, there were many instances, especially in the low-intensity condition, where this was not possible. The remaining number of subjects with wave V measures for both task and intensity conditions, were 10 Rs and 5 NRs among the schizophrenics, 4 Rs and 10 NRs among the alcoholics, and 4 Rs and 17 NRs among the normal subjects. The ANOVA showed a very strong main effect of intensity (F = 385.61; df = 1, 42; p < 0.01) and an interaction of the Electrodermal Response Type and Task (F= 4.22, df = 1,42;p < 0.05, Fig. 2A). Mean latencies were 6.23 ms for loud and 7.56 ms for soft clicks. The interaction is due to the NRs who showed longer wave V latencies when instructed to count lights (mean 6.94 ms) as compared to counting clicks (mean = 6.85 ms). q
230 To ensure that this result was not caused by the inequality of cell frequencies, the diagnoses were pooled and the analysis was repeated. The resulting F ratio for the Electrodermal Response Type x Task interaction remained almost the same as before (F = 4.48; df= 1,46). As differences in signal reliability might also have led to such an interaction, the correlation coefficients between the BAEP waveforms from odd and even trials were subjected to the same type of ANOVA as the wave V amplitudes. As the Electrodermal Response Type x Task interaction, significant for wave V amplitudes, remained nonsignificant for the correlation coefficients (F = 0.50; df= 1, 63; p = 0.48) there seems to be no relationship between the wave V effect and signal reliability. To enhance the number of detectable peaks, wave I latencies and wave I-V conduction times were based on BAEP signals, which were averaged for the two tasks, and the ANOVA was restricted to the high-intensity condition. Thus, only factors Diagnosis and Electrodermal Response Type remained, with 8 NRs and 11 Rs among schizophrenics, 5 NRs and 7 Rs among alcoholics, and 4 NRs and 14 Rs among normal subjects. There were no significant effects for wave I latencies, but for the wave I-V conduction times, there was an interaction of Diagnosis and Electrodermal Response Type (F = 4.77; df= 2,43;p < 0.05; Fig. 2B). Alcoholic NRs showed longer mean conduction times (4.63 ms) than all the other five subgroups, with further differences only between means for the schizophrenic Rs (4.3 ms) and the normal Rs (4.05 ms). Fig. 2A. Wave V latencies for electrodermal responders and nonresponders
Fig. 28. Wave I-V conduction times B)
A)
R
NR 6.8
40
LLIGHTS
COUNT
Latencies were derived from either auditory or visual stimuli
CLICKS
conditions in which were to be attended.
LSCHIZOPHRENICS ALCOHOLICS
NORMALS
Subjects in the 3 diagnostic groups were categorized into electrodermal responders and nonresponders.
Discussion In electrodermal NRs the latencies of BAEP wave V, elicited by frequent auditory nontarget stimuli, are longer when attention is directed toward visual stimuli than when subjects are detecting rare target stimuli within the auditory modality; there is no such effect for the Rs.
231 Previous studies of attentional effects on human cochlear nerve and brainstem activity have mostly reported negative results (e.g., Woods and Hillyard, 1978). It is only recently that Lukas (1980) reported empirical evidence for a peripheral gating mechanism at the brainstem level in humans. He found the amplitudes of BAEP waves I and V to be reduced and the latencies prolonged if the subjects performed a visual discrimination task during the click presentation as compared to “listening to” the auditory stimuli. Lukas (1980) attributed the negative results of the previous human studies to factors such as slow stimulus presentation rates, allowance for the simultaneous processing of both sensory channels, use of high-frequency stimuli, or use of long, experimenter-initiated trials. As the conditions of this study resemble those with negative results, it is not surprising to find no influence of the direction of attention on the BAEP of Rs, who constitute the majority of “normal” populations. In NRs, however, the findings of Lukas (1980) are replicated for wave V latencies. Lukas interpreted the prolonged wave V latencies under distraction from the BAEP eliciting stimuli as a sign of efferent inhibition from the olivocochlear bundle on auditory input, allowing relevant stimuli to pass freely, whereas irrelevant stimuli are attenuated by preset filters. On the basis of this line of argument, the task effect in NRs and its absence in Rs might be seen as a sign of more accentuated gating of sensory stimuli in NRs than in Rs-that is, NRs shut out irrelevant stimuli and/or admit attended stimuli more readily into the afferent sensory system than Rs do. This is quite what one would expect from the SCORs, where NRs are seemingly insensitive to irrelevant orienting stimuli but readily respond when the same stimuli are to be attended (Gruzelier and Venables, 1973). The present data indicate that this excessive selective filtering of sensory stimuli in NRs is already in effect at the brainstem level, involving efferent inhibition of the olivocochlear bundle on structures up to the inferior colliculus. An alternative explanation for the task effect in NRs might be that the task was more difficult for the NRs, a condition which, according to Lukas, promotes the effects of attention on the BAEP. This hypothesis finds some support in a nonsignificant trend toward poorer counting results in NRs (F= 3.1 I; df= 1, 59;~ 0.08). There are no differences between the diagnostic groups in BAEP latencies, either as a main effect or in interaction with Electrodermal Response Type or Task. The former normal auditory is in line with the literature (e.g., Roth et al., 1979), indicating functions in schizophrenics and alcoholics. The lack of interaction between Diagnosis and Electrodermal Response Type suggests that at the brainstem level electrodermal nonresponding in schizophrenics, alcoholics, and normal subjects is related to the same excessive filtering of irrelevant sensory input. However, under the probably nonoptimal conditions of the present experiment, in which the attentional effect on the BAEP was rather small, one should not be surprised at the inability to discern possible, but even more subtle variations of this effect between the diagnostic groups. Clearly, the findings call for replication under conditions such as high stimulation rates, low channel separation, and/ or short, subject-initiated trials. q
232 The lack of differences between the diagnostic groups in wave V latencies contrasts to the prolonged wave I-V conduction times in the alcoholic Rs. This measure is usually considered as a global sign of auditory pathway function from the eighth nerve action potential to the inferior colliculus (Starr et al., 1978). As the number of subjects in this group is only five and the inspection of individual data reveals that the effect is due to only three of these subjects, the results should best be considered as preliminary and requiring further investigation. In some alcoholic NRs, electrodermal nonresponding might be related to a retarded neural transmission of the auditory input. Acknowledgments. The author deeply appreciates Psychiatrisches Landeskrankenhaus Reichenau.
the help of the staff and the patients
of the
References American Psychiatric Association. DSM-III: Diagnostic and Statistical Manual of Mental Disorders. 3rd ed. APA, Washington, DC (1980). Bernstein, A.S., Frith, C.D., Gruzelier, J.H., Patterson, T., Straube, E., Venables, P.H., and Zahn, T.P. An analysis of the skin conductance orienting response in samples of American, British, and German schizophrenics. Biological Psychology, 14, 155 (1982). Callaway, E., Tueting, P., and Koslow, S.H., eds. Event-Related Brain Potentials in Man. Academic Press, New York (1978). Cohen, R., Sommer, W., and Hermanutz, M. Auditory event-related potentials in chronic schizophrenics: Effects of electrodermal response type and demands on selective attention. In: Mendlewicz, J., and van Praag, H.M., eds. Advances in Biological Psychiatry. Vol. 6. S. Karger, Base1 (198 1). Gruzelier, J.H., and Venables, P.H. Skin conductance responses to tones with and without attentional significance in schizophrenic and non-schizophrenic patients. Neuropsychologia, 11,221 (1973). Lukas, J.H. Human auditory attention: The olivocochlear bundle may function as a peripheral filter. Psychophysiology, 17, 444 (1980). Roth, W.T., Ford, J.M., Pfefferbaum, A., Horvath, T.B., Doyle, C.M., and Kopell, B.S. Event-related potential research in psychiatry. In: Lehmann, D., and Callaway, E., eds. Human Evoked Potentials: Applications and Problems. Plenum Press, New York (1979). Starr, A., Sohmer, H., and Celesia, G.G. Some applications of evoked potentials to patients with neurological and sensory impairment. In: Callaway, E., Tueting, P., and Koslow, S.H., eds. Event-Related Brain Potentials in Man. Academic Press, New York (1978). Woods, D.L., and Hillyard, S.A. Attention at the cocktail party: Brainstemevoked responses reveal no peripheral gating. In: Otto, D.A., ed. Multidisciplinary Perspectives in Event-Related Brain Potential Research. U.S. Environmental Protection Agency (EPA-600/9-77-043), Washington, DC (1978).
Psychiatry Research, 16, 227-232
Elsevier
Selective Attention Differentially Affects Brainstem Auditory Evoked Potentials of Electrodermal Responders and Nonresponders Werner
Sommer
Received September 1985.
4, 1984; revised version received January 25. 1985; accepted February 13,
Abstract. Brainstem auditory evoked potentials (BAEPs) were recorded in adult male chronic schizophrenics on maintenance neuroleptic therapy, drug-free alcoholics, and normal subjects. Subjects were subdivided into electrodermal responders (Rs) and nonresponders (NRs). Attention was directed toward auditory or visual stimuli. Independent of diagnosis, latencies of BAEP wave V were longer in NRs when visual rather than auditory stimuli were attended to, while there was no task effect for Rs. This finding is interpreted as a sign of excessive selective filtering of auditory stimuli in NRs. In the alcoholic NRs, wave I-V conduction times were longer than those in any other subgroup, possibly indicating retarded neural transmission. Key Words. Alcoholism, brainstem auditory evoked nonresponding, schizophrenia, selective attention.
potentials,
electrodermal
Many schizophrenic patients are electrodermal nonresponders (NRs) to orienting stimuli (Bernstein et al., 1982). When the same stimuli are made task-relevant, however, former NRs become responders (Rs) (Gruzelier and Venables, 1973). Thus, nonresponding cannot be easily accounted for as a simple sensory deficit; more complex abnormalities of information processing must be implicated. In human information-processing studies, event-related brain potentials (ERPs) have proved to be a valuable tool (Callaway et al., 1978). While short latency ERPs reflect the encoding of stimuli in the afferent sensory systems, the later components are related to the subjective evaluation of the prior input. To date, however, no published study of which I am aware has addressed itself to ERPs in subjects differing in their Electrodermal Response Type. In this investigation, ERPs were recorded in the same subjects from the early brainstem auditory evoked potentials (BAEPs) up to the long latency, endogenous components, varying the direction of attention within repeatedly presented sets of auditory and visual stimuli. Results concerning the late ERP components have been published elsewhere (Cohen et al., 1981). In this article, BAEP correlates of electrodermal responding and nonresponding are described.
Dr. Werner Sommer, Dip]. Psych., is at Universiik Konstanz, FG Psychologie, Postfach Konstanr, W. Germany. (Reprint requests to Dr. W. Sommer at the address above.) 0165-1781~85:
$03.30 0 1985 Elsevier Science Publishers
B.V.
5560, D-7750
228 Methods The subjects, all males, were 26 nonparanoid schizophrenics, 22 chronic alcoholics, and 22 healthy employees of a German psychiatric state hospital. The alcoholics had been abstinent for at least 3 weeks and were off tranquilizers. Twenty-two of the schizophrenics were receiving maintenance neuroleptic therapy. Psychiatric classification was based on the consensus of two experienced psychiatrists and DSM-IZI(American Psychiatric Association, 1980) criteria. The mean duration of hospitalization was 8 months for the alcoholics and 4.4 years for the schizophrenics. All the subjects had participated in a similar experiment a few days before in which skin conductance orienting responses (SCORs) and long latency ERPs were recorded. The groups were matched for age (mean q 41 years) and educational status. As it was the aim of this study to investigate the processing of auditory stimuli in such Rs and NRs as usually studied in psychophysiological research, no attempt was made to select the subjects or to adjust the stimulus intensity levels according to the psychometrical “hearing” performance of the subjects. To confirm subjects’ electrodermal response types, five tones (7%dB, I-kHz, l-second duration, 25-ms rise time) were presented at intervals between 25 and 40 seconds. The subjects were classified as Rs or NRs depending on whether they showed any SCORs greater than 0.05 micromho to the first three stimuli. All subjects showed shifts in skin conductance in the social exchange before and after the experiment. For the BAEPs, bursts of 10 clicks (0. I-ms duration, lOO-ms interstimulus interval, 650-ms interburst interval) were presented binaurally. The intensity of the rare (10%) click bursts was always 50 dB, while the frequent bursts were either of 70 dB (high intensity condition) or of 30 dB (low intensity). The intensity of the clicks was related to the threshold of a listener with normal hearing. Simultaneously, sets of light bars of l-second duration were presented on a screen with the same temporal distribution as the rare click bursts. In counterbalanced order, the subjects either counted the rare click bursts (count clicks) or the light stimuli (count lights). The number of targets counted had to be reported at the end of each task. The electroencephalogram (EEG) was recorded with Ag/AgCl electrodes from vertex referred to linked mastoids and differentially amplified (x 100,000, bandpass: 10 to 3,000 Hz, 6 dB rolloff/ octave). The responses to thefrequent clicks were sampled at 10 kHz for 10 ms after stimulus onset and averaged separately for each task, for each intensity condition, and for odd and even stimuli within the click bursts (cf. Fig. 1). For peak detection, the averages from odd and even stimuli were combined, resulting in averages of 2,000 and 4,000 responses to high- and low-intensity stimuli, respectively. The latencies of waves I and V were determined by visual inspection wherever possible, and analyses of variance (ANOVAs) were calculated for the two latency measures, as well as for the difference between them (wave I-V interpeak conduction time). To correct for nonhomogeneity of cell variance, ANOVAs were based on log-transformed data; the means, however, are reported untransformed. For a posteriori tests @ < 0.05), the Neuman Keuls procedure was used.
Results The proportion of NRs to Rs was I 1 to 15 in schizophrenics, 6 to 16 in alcoholics, and 4 to 17 in normal subjects. In comparison to the first SCOR test, this is an increase of four NRs in the normal subjects, two in alcoholics, and one in schizophrenics. The test/retest correlation for the global R/ NR dichotomy amounts to a 0 coefficient of 0.47 (x2= 15.37; df= l;p
229 Fig. 1. BAEP waveforms in a typical subject COUNT
7066
Y
2 i LJ
30dB
11
’
0
2
”
“1’ 4
1
6
8
I
IO mset
Frequent 30 dB In = 2,000) and 70 dB (n = 1.000) clicks ware heard while either rare auditory or visual stimuli were to be counted. BAEPs were averaged separately for odd and even stimuli (solid and dashed lines) within click bursts.
The ANOVA of the counting results with factors Diagnosis (schizophrenics, alcoholics, normal subjects), Eectrodermal Response Type (NRs, Rs) and repeated measures on Task (count lights, count clicks) and Intensity (30 dB, 70 dB) revealed a main effect of Diagnosis (F= 6.0 1; df = 2, 59; p < 0.0 l), with schizophrenics deviating from the correct result by 11.5% while alcoholics and normal subjects had only a 3.3% and 4.6% average deviation, respectively. Even though wave V could be determined most reliably, there were many instances, especially in the low-intensity condition, where this was not possible. The remaining number of subjects with wave V measures for both task and intensity conditions, were 10 Rs and 5 NRs among the schizophrenics, 4 Rs and 10 NRs among the alcoholics, and 4 Rs and 17 NRs among the normal subjects. The ANOVA showed a very strong main effect of intensity (F = 385.61; df = 1, 42; p < 0.01) and an interaction of the Electrodermal Response Type and Task (F= 4.22, df = 1,42;p < 0.05, Fig. 2A). Mean latencies were 6.23 ms for loud and 7.56 ms for soft clicks. The interaction is due to the NRs who showed longer wave V latencies when instructed to count lights (mean 6.94 ms) as compared to counting clicks (mean = 6.85 ms). q
230 To ensure that this result was not caused by the inequality of cell frequencies, the diagnoses were pooled and the analysis was repeated. The resulting F ratio for the Electrodermal Response Type x Task interaction remained almost the same as before (F = 4.48; df= 1,46). As differences in signal reliability might also have led to such an interaction, the correlation coefficients between the BAEP waveforms from odd and even trials were subjected to the same type of ANOVA as the wave V amplitudes. As the Electrodermal Response Type x Task interaction, significant for wave V amplitudes, remained nonsignificant for the correlation coefficients (F = 0.50; df= 1, 63; p = 0.48) there seems to be no relationship between the wave V effect and signal reliability. To enhance the number of detectable peaks, wave I latencies and wave I-V conduction times were based on BAEP signals, which were averaged for the two tasks, and the ANOVA was restricted to the high-intensity condition. Thus, only factors Diagnosis and Electrodermal Response Type remained, with 8 NRs and 11 Rs among schizophrenics, 5 NRs and 7 Rs among alcoholics, and 4 NRs and 14 Rs among normal subjects. There were no significant effects for wave I latencies, but for the wave I-V conduction times, there was an interaction of Diagnosis and Electrodermal Response Type (F = 4.77; df= 2,43;p < 0.05; Fig. 2B). Alcoholic NRs showed longer mean conduction times (4.63 ms) than all the other five subgroups, with further differences only between means for the schizophrenic Rs (4.3 ms) and the normal Rs (4.05 ms). Fig. 2A. Wave V latencies for electrodermal responders and nonresponders
Fig. 28. Wave I-V conduction times B)
A)
R
NR 6.8
40
LLIGHTS
COUNT
Latencies were derived from either auditory or visual stimuli
CLICKS
conditions in which were to be attended.
LSCHIZOPHRENICS ALCOHOLICS
NORMALS
Subjects in the 3 diagnostic groups were categorized into electrodermal responders and nonresponders.
Discussion In electrodermal NRs the latencies of BAEP wave V, elicited by frequent auditory nontarget stimuli, are longer when attention is directed toward visual stimuli than when subjects are detecting rare target stimuli within the auditory modality; there is no such effect for the Rs.
231 Previous studies of attentional effects on human cochlear nerve and brainstem activity have mostly reported negative results (e.g., Woods and Hillyard, 1978). It is only recently that Lukas (1980) reported empirical evidence for a peripheral gating mechanism at the brainstem level in humans. He found the amplitudes of BAEP waves I and V to be reduced and the latencies prolonged if the subjects performed a visual discrimination task during the click presentation as compared to “listening to” the auditory stimuli. Lukas (1980) attributed the negative results of the previous human studies to factors such as slow stimulus presentation rates, allowance for the simultaneous processing of both sensory channels, use of high-frequency stimuli, or use of long, experimenter-initiated trials. As the conditions of this study resemble those with negative results, it is not surprising to find no influence of the direction of attention on the BAEP of Rs, who constitute the majority of “normal” populations. In NRs, however, the findings of Lukas (1980) are replicated for wave V latencies. Lukas interpreted the prolonged wave V latencies under distraction from the BAEP eliciting stimuli as a sign of efferent inhibition from the olivocochlear bundle on auditory input, allowing relevant stimuli to pass freely, whereas irrelevant stimuli are attenuated by preset filters. On the basis of this line of argument, the task effect in NRs and its absence in Rs might be seen as a sign of more accentuated gating of sensory stimuli in NRs than in Rs-that is, NRs shut out irrelevant stimuli and/or admit attended stimuli more readily into the afferent sensory system than Rs do. This is quite what one would expect from the SCORs, where NRs are seemingly insensitive to irrelevant orienting stimuli but readily respond when the same stimuli are to be attended (Gruzelier and Venables, 1973). The present data indicate that this excessive selective filtering of sensory stimuli in NRs is already in effect at the brainstem level, involving efferent inhibition of the olivocochlear bundle on structures up to the inferior colliculus. An alternative explanation for the task effect in NRs might be that the task was more difficult for the NRs, a condition which, according to Lukas, promotes the effects of attention on the BAEP. This hypothesis finds some support in a nonsignificant trend toward poorer counting results in NRs (F= 3.1 I; df= 1, 59;~ 0.08). There are no differences between the diagnostic groups in BAEP latencies, either as a main effect or in interaction with Electrodermal Response Type or Task. The former normal auditory is in line with the literature (e.g., Roth et al., 1979), indicating functions in schizophrenics and alcoholics. The lack of interaction between Diagnosis and Electrodermal Response Type suggests that at the brainstem level electrodermal nonresponding in schizophrenics, alcoholics, and normal subjects is related to the same excessive filtering of irrelevant sensory input. However, under the probably nonoptimal conditions of the present experiment, in which the attentional effect on the BAEP was rather small, one should not be surprised at the inability to discern possible, but even more subtle variations of this effect between the diagnostic groups. Clearly, the findings call for replication under conditions such as high stimulation rates, low channel separation, and/ or short, subject-initiated trials. q
232 The lack of differences between the diagnostic groups in wave V latencies contrasts to the prolonged wave I-V conduction times in the alcoholic Rs. This measure is usually considered as a global sign of auditory pathway function from the eighth nerve action potential to the inferior colliculus (Starr et al., 1978). As the number of subjects in this group is only five and the inspection of individual data reveals that the effect is due to only three of these subjects, the results should best be considered as preliminary and requiring further investigation. In some alcoholic NRs, electrodermal nonresponding might be related to a retarded neural transmission of the auditory input. Acknowledgments. The author deeply appreciates Psychiatrisches Landeskrankenhaus Reichenau.
the help of the staff and the patients
of the
References American Psychiatric Association. DSM-III: Diagnostic and Statistical Manual of Mental Disorders. 3rd ed. APA, Washington, DC (1980). Bernstein, A.S., Frith, C.D., Gruzelier, J.H., Patterson, T., Straube, E., Venables, P.H., and Zahn, T.P. An analysis of the skin conductance orienting response in samples of American, British, and German schizophrenics. Biological Psychology, 14, 155 (1982). Callaway, E., Tueting, P., and Koslow, S.H., eds. Event-Related Brain Potentials in Man. Academic Press, New York (1978). Cohen, R., Sommer, W., and Hermanutz, M. Auditory event-related potentials in chronic schizophrenics: Effects of electrodermal response type and demands on selective attention. In: Mendlewicz, J., and van Praag, H.M., eds. Advances in Biological Psychiatry. Vol. 6. S. Karger, Base1 (198 1). Gruzelier, J.H., and Venables, P.H. Skin conductance responses to tones with and without attentional significance in schizophrenic and non-schizophrenic patients. Neuropsychologia, 11,221 (1973). Lukas, J.H. Human auditory attention: The olivocochlear bundle may function as a peripheral filter. Psychophysiology, 17, 444 (1980). Roth, W.T., Ford, J.M., Pfefferbaum, A., Horvath, T.B., Doyle, C.M., and Kopell, B.S. Event-related potential research in psychiatry. In: Lehmann, D., and Callaway, E., eds. Human Evoked Potentials: Applications and Problems. Plenum Press, New York (1979). Starr, A., Sohmer, H., and Celesia, G.G. Some applications of evoked potentials to patients with neurological and sensory impairment. In: Callaway, E., Tueting, P., and Koslow, S.H., eds. Event-Related Brain Potentials in Man. Academic Press, New York (1978). Woods, D.L., and Hillyard, S.A. Attention at the cocktail party: Brainstemevoked responses reveal no peripheral gating. In: Otto, D.A., ed. Multidisciplinary Perspectives in Event-Related Brain Potential Research. U.S. Environmental Protection Agency (EPA-600/9-77-043), Washington, DC (1978).