Psychophysiology, 38 ~2001!, 669–677. Cambridge University Press. Printed in the USA. Copyright © 2001 Society for Psychophysiological Research

Attentional stages of information processing during a continuous performance test: A startle modification analysis

ERIN A. HAZLETT,a MICHAEL E. DAWSON,b ANNE M. SCHELL,c and KEITH H. NUECHTERLEIN b a

Department of Psychiatry, Mount Sinai School of Medicine, New York, USA Department of Psychology, University of Southern California, Los Angeles, USA c Department of Psychology, Occidental College, Los Angeles, USA d Department of Psychiatry and Biobehavioral Sciences, University of California, Los Angeles, USA b

Abstract This study of 31 college students employed the startle eye-blink modification ~SEM! technique to index both early and later stages of attentional processing during a memory-load version of the Continuous Performance Test ~CPT!. Participants viewed a series of digits and pressed a button after the digit 7 of each 3–7 sequence. A startling noise burst was presented either 120 or 1,200 ms following three preselected prepulses: target ~3!, nontarget ~non-3 and non-7 digits!, or target plus distractor ~3 and simultaneous tone distractor!. Greater startle inhibition occurred 120 ms following target and target-plus-distractor prepulses compared with nontargets, indicating early selective attention. No difference was observed between SEM during target and target-plus-distractor prepulses, suggesting the distractor was effectively ignored. At 1,200 ms, the three prepulse types produced nondifferential inhibition, suggesting that modalityspecific selective attention occurs in anticipation of the presentation of the next CPT prepulse. These findings indicate that SEM distinguishes between different early selective attention and later anticipatory attention subprocesses underlying the CPT. Descriptors: Startle eye-blink modification, Selective attention, Prepulse inhibition, Continuous performance test, Skin conductance

interval ~referred to as the lead interval! between the prepulse and the startle stimulus is relatively short ~between 30 and 300 ms!, a process referred to as prepulse inhibition ~PPI!. At long lead intervals ~greater than 500 ms!, startle amplitude is facilitated if attention is directed toward the modality of the startle-eliciting stimulus, but may be inhibited if attention is directed away from the modality of the startle-eliciting stimulus ~Putnam & Vanman, 1999!. Thus, SEM at long lead intervals may be either facilitatory or inhibitory depending on the direction of attention, and therefore may be a sensitive index of sustained modality-specific attention. In a passive attention paradigm without demands to actively engage attentional mechanisms, SEM may index automatic sensorimotor processes at short lead intervals and automatic arousal processes at long lead intervals ~Graham, 1975!. In an active attention paradigm, SEM may index, in addition to automatic processing, early controlled attentional processes at short lead intervals and later controlled processes at long lead intervals ~Dawson, Schell, Swerdlow, & Filion, 1997!. Several investigators have shown that during selective attention tasks, both short and long lead interval SEM is modified by controlled attentional processes, with greater SEM following an attended prepulse than an ignored prepulse ~see review by Filion, Dawson, & Schell, 1998!. Taken together, the results of several studies suggest that the SEM technique can be used as a sensitive “probe” of ongoing automatic and

The startle eye-blink modification ~SEM! technique has been widely used by cognitive psychophysiologists during the past three decades to further understand information processing ~Dawson, Schell, & Böhmelt, 1999; Hackley, 1999!. This technique has several theoretical and methodological strengths. These include its potential to provide separate evaluation of automatic processes and controlled attentional processes on a within-task basis without interference with an ongoing task, noninvasiveness, its amenability to animal modeling, modulation by drug effects, and ease of measurement in clinical populations due to a reflexive nature. SEM occurs when innocuous, nonstartling stimuli ~prepulses! are presented shortly before startle-eliciting stimuli; prepulses reliably inhibit or facilitate the amplitude of the startle reflex, including the startle eye-blink reflex. The startle eye-blink is inhibited if the This research was supported by grants from the National Institute of Mental Health ~MH10381 to E.A.H.; MH46433 and K02 MH01086 to M.E.D.; MH37705 to K.H.N. and MH30911 to R.P. Liberman at UCLA!. We thank Dr. Diane Filion for consulting on various technical aspects of this research, William Troyer for software development, and Steven Hackley and two anonymous reviewers for their thoughtful comments on an earlier draft of this manuscript. Address reprint requests to: Erin A. Hazlett, Ph.D., Department of Psychiatry, Box 1505, Mount Sinai School of Medicine, 1 Gustave L. Levy Place, New York, NY 10029, USA. E-mail: [email protected].

669

670 controlled information processing in various paradigms ~see reviews by Anthony, 1985; Filion et al., 1998; Putnam, 1990!. To date, studies that have reported attentional modulation effects in SEM have employed discrete-trial paradigms with long prepulse interstimulus intervals ~e.g., 15 s or more!. Although it has been suggested that methods using rapid presentation rates of stimuli provide inherently stronger manipulations of attention ~Graham & Hackley, 1991!, SEM research which has examined selective attention has focused on the discrete-trial format with one exception. Hackley, Woldorff, and Hillyard ~1987! used an auditory evoked-potential experimental design with rapid stimulus presentations and reported a selective sensory attentional effect for PPI of the postauricular reflex with greater PPI produced by prepulses presented to the attended ear than prepulses presented to the unattended ear. The present study extends the earlier work of Hackley et al. in several ways. First, because the Hackley et al. study did not have a no-prepulse control condition, they could not firmly conclude that the attended- versus ignored-ear differences were due to enhancement of inhibition triggered by the attendedear prepulses. Secondly, the relevant dependent measure in the Hackley et al. study was the pinna-flexion ~i.e. postauricular! component of startle, rather than the eye-blink ~orbicularis oculi! component. There is very little animal or human research on PPI of the pinna-flexion component of startle, whereas PPI of the eyeblink component is a thoroughly investigated model system. Finally, Hackley et al. used a complicated design in which each stimulus served both as a relexogenic stimulus and, at the same time, as a prepulse to modulate the response to the subsequent stimulus. The present study relates more simply and directly to the existing literature on attentional modulation of the startle eyeblink. In light of these methodological issues, it is unclear whether selective attention effects in SEM are limited to discrete-trial paradigms or whether they extend to paradigms such as a continuous performance test ~CPT!, which uses an ongoing stream of stimuli and involves a relatively high and continuous processing load. The present research addressed this issue and optimized experimental conditions by using prepulse interstimulus intervals that were longer than those employed by Hackley et al., yet substantially shorter than those used in previous discrete-trial SEM research. In a thorough review of psychophysiological evidence, Graham and Hackley ~1991! concluded that SEM as well as brainwave evoked-potential data indicate that attention does affect the early stages of sensory-perceptual processing in healthy individuals, based on the divergence in the processing of attended and ignored stimuli. The present study employed an adaption of a visual memoryload continuous performance test ~3–7 CPT; Nuechterlein, Edell, Norris, & Dawson, 1986!, which is an information-processing task thought to involve rapid perceptual discrimination as well as working-memory processes. Versions of the CPT are widely used as indicators of cognitive dysfunction in schizophrenia and attentiondeficit0hyperactivity disorder. The CPT involves a quasi-random presentation of a sequence of target and nontarget stimuli. Typically, the stimuli are visual and presented briefly ~40–200 ms!, and occur once every 1 or 2 s. Subjects respond to targets by pressing a button. It is still unclear, however, exactly what this test measures. The overarching goal of the present study was to use SEM to better understand the nature of early and later cognitive subprocesses tapped by this vigilance task thought to demand active attention. Specifically, this experiment had two specific aims: ~1! to evaluate automatic and controlled information processing by examining SEM at short and long lead intervals following the

E.A. Hazlett et al. onset of target, nontarget, and target-plus-distractor prepulses during performance of the CPT and ~2! to examine the association between behavioral and SEM measures during CPT performance. Two lead intervals were employed to examine early and later attentional processing during the CPT. Short lead interval SEM is a useful measure for probing into the early processing stages of attention, whereas long lead interval SEM is used to probe later stages on a within-task basis. As discussed elsewhere ~Filion, Dawson, & Schell, 1993!, the SEM technique allows a method for probing attentional processing in close to “real-time,” in that the startle response can reflect cognitive activity occurring within less than 100 ms from the time that the startle-eliciting stimulus is presented. Dawson et al. ~1997! proposed that in an attention demanding task, PPI at very short lead intervals ~e.g., 60 ms! reflects automatic stimulus detection and identification, whereas PPI at somewhat longer lead intervals ~e.g., 120 ms! reflects additional controlled attentional processes. SEM at even longer lead intervals ~e.g., 1,200 ms! reflects general activation plus sustained modalityselective controlled attention. Based on this model, we predicted that several cognitively relevant processes would affect SEM during the CPT. At the 120-ms probe position during the target and the target-plus-distractor prepulses, early controlled attentional modulation of processing would be engaged and is hypothesized to modulate PPI, with greater PPI occurring after the target compared with the nontarget prepulses. Because we expected that the subjects would effectively ignore the auditory distractor, we hypothesized that the target and the target-plus-distractor prepulses would receive equal controlled attentional processing. Therefore, similar PPI modulation would occur during these two prepulse types. At the long lead interval ~i.e. 1,200 ms!, sustained modality-selective attention to the visual modality is hypothesized to be the dominant cognitive process following the target and the target-plus-distractor prepulses. Because attention is focused on the visual modality and the startle stimulus is in the auditory modality, we hypothesized that SEM at 1,200 ms would be inhibitory. Similar to the short lead interval, we hypothesized that the target and the target-plusdistractor prepulses would produce equal SEM inhibition at 1,200 ms because the auditory distractor would be effectively ignored. We also predicted that greater controlled processing and therefore, greater PPI would be observed following the target compared with the nontarget prepulses at the long lead interval. Because no cognitive task was required during the distractor prepulse, SEM following the distractor prepulse ~alone! was expected to provide an estimate of automatic processing. Based on previous research ~Hazlett et al., 1998!, we predicted that PPI would be observed at 120 ms and prepulse facilitation would be observed at 1,200 ms. Methods Participants Thirty-one students from Introductory Psychology classes at the University of Southern California comprised the final sample of individuals who volunteered to participate in this experiment ~mean age 5 19.3, SD 5 2.4; 22 female and 9 male!. Data from an additional 13 individuals were discarded from analysis for miscellaneous reasons: equipment problems ~n 5 2!; excessive muscle artifact ~n 5 5!; failure to exhibit eye-blinks on over 25% of the baseline startle stimulus presentations ~n 5 6!. All participants received extra credit points in their Introductory Psychology class for their participation and provided written informed consent.

Attention and startle modification Design To examine the SEM effects of prepulses presented during the CPT, a 3 3 3 within-subjects design was employed for the short and long lead intervals separately. The two variables consisted of prepulse type ~target, nontarget, and target plus distractor! and trial block ~1 to 3!. To examine the SEM effects of the distractor prepulses, a one-way design was employed with baseline trial block ~1 to 2! as a repeated measure for the short and long lead intervals separately. Procedure Upon arrival at the laboratory, participants were asked to read and sign an informed consent form that provided a general description of the experiment and the physiological responses being recorded. Participants also completed a short questionnaire that included questions pertaining to general demographic information, vision and hearing impairments, family history of neurological and psychiatric disorders, and recent use of medications. None of the participants reported hearing impairments and all reported normal or corrected-to-normal vision. In addition, none of the participants reported significant neurological or psychiatric histories or recent medication usage. Participants were then seated in a comfortable chair in the testing room with their eyes 1 m from the blank computer monitor. The experimental session consisted of four phases: electrode placement and task instructions, rest period, CPT practice, and the SEM-CPT experiment. First, electrodes were attached using standard methods for the recording of skin conductance ~Dawson, Schell, & Filion, 2000! and startle eye-blink ~Berg & Balaban, 1999!. Next, participants were presented tape-recorded instructions that provided information about the nature of the stimuli to be used and a description of their task throughout the experimental session. Participants were instructed to watch a series of single digits that would be presented rapidly on the computer monitor in front of them. More specifically, they were told that their task would be to concentrate on the numbers appearing on the screen and press a response button every time they saw the number 3 followed by the number 7 and to refrain from pressing the button at any other time. Participants were also instructed that two types of auditory distractors ~a pure tone and0or a brief burst of static noise! would be presented occasionally throughout the experiment and that they should simply ignore these distractors and concentrate on the visual task. Following these instructions, participants were asked to sit quietly with their eyes open while skin conductance was recorded for a 3-min resting phase. Throughout the entire experimental session, participants were monitored through a one-way mirror for any movements. Next, some additional details regarding the 3–7 CPT task were described and 10 examples of the 3–7 sequence were shown to the participants. They were asked to practice pressing the button after each of these 3–7 sequences. Next, participants performed one block ~160 trials! of the CPT, which served as practice to familiarize them with the visual task and ensure that they understood the task instructions. No distractors were presented during this practice block. Participants were then given three warned examples of the startle stimulus for the purpose of adjusting polygraph sensitivity. Following the examples of the noise bursts, the SEM-CPT portion of the experimental session began. The SEM-CPT phase lasted 21.5 min and consisted of three CPT blocks with a baseline period preceding and following each block. Each CPT block consisted of 160 stimulus presentations or

671 “trials” made up of twenty 3–7 sequences ~forty stimuli!, eighty nontargets, twenty 3s not followed by a 7, and twenty 7s not preceded by a 3 over a continuous observation period of approximately 4.5 min ~interstimulus interval 5 1.65 s!. On 20% of the CPT trials ~32 trials!, a tone distractor was presented through headphones simultaneously with the CPT stimulus. To measure blink modification during the processing of the CPT stimuli, 12 preselected trials were “probed” with a startle-eliciting stimulus. It is important to note that in order to measure blink modification effects to “target” stimuli without interference from the task motor response, the startle-eliciting stimulus was presented at one of two possible lead intervals following the onset of the digit 3 of a 3–7 trial sequence, while the participant was to press the response button after the digit 7 of the same sequence. Figure 1 shows a schematic of a portion of the trial sequence during the baseline and CPT periods. To determine blink modification effects of “nontarget” stimuli, the startle-eliciting stimulus was presented following the onset of a nontarget digit ~i.e. 0s and 1s were chosen to be the least easily confused with the targets!. Last, to determine the blink modification effects of distractability, startle-eliciting probes were presented on target trials that had the simultaneously presented auditory distractor. In sum, during each CPT block, there were two startle-eliciting probes presented at both the short and long lead intervals following preselected target, nontarget, and target-plus-distractor stimuli ~“prepulses”!, yielding a total of 12 probe trials. Six of the probe trials were presented during the first 80 CPT stimuli and the remaining six were presented during the last 80 stimuli. The amount of time between the startle-eliciting probes varied from 13 to 36 s, with a mean interval of 24 s. The CPT trial sequence was varied across the three CPT blocks by rotating the starting point of the 160 trials. Although the order of probed CPT trials was pseudo-random within each participant’s test session, it did not vary across individuals. However, probe position ~120 vs. 1,200 ms! was counterbalanced across subjects. Each CPT block was preceded and followed by a baseline period that resulted in a total of four baseline periods ~see Figure 1!. The purpose of the 2-min baseline periods was twofold. First, to determine whether participants showed blink modification during the distractor prepulse ~i.e. tone prepulse without the presence of a visual CPT prepulse! during each baseline period, a startle probe was presented at both the short and the long lead

Figure 1. Schematic of a portion of the experimental session. During the baseline blocks, there were two stimulus conditions ~startle probe alone, distractor tone with a startle probe presented at either the 120 ms or 1,200 ms lead interval!. During the CPT blocks, there were three stimulus conditions ~target, nontarget, and target plus the distractor tone! that had a startle probe presented at either the 120 ms or 1,200 ms lead interval.

672 interval following the onset of the distractor prepulse. Second, during each baseline period, two startle stimuli were presented at preselected times in the absence of a prepulse stimulus. Responses to these probes served as baseline measures with which to compare blink modification produced by the prepulses. No stimuli were presented for at least 15 s before the distractor prepulses that contained probes or before the baseline probes. To match the individual’s arousal level during the baseline and CPT periods as closely as possible, CPT digits were presented during the baseline periods. These CPT digits had the same interstimulus interval as those presented during the CPT proper ~1.65 s!; however, occasionally there were relatively longer interstimulus intervals ~varying from 3 to 24 s with a mean interval of 13 s!. The intervals between digit prepulses were sometimes longer during the baseline compared with the CPT periods, in order to allow for the presentation of other stimulus conditions. It is important to note that there was no interruption between the baseline and CPT phases of the experiment. So, from the participant’s point of view, the baseline period was identical to the CPT except for the occasional longerthan-usual gap between CPT stimuli. The participants were told that occasionally there would be brief periods where no numbers would appear, but they should focus their gaze on the computer monitor and wait for the next number to appear. In sum, during each baseline period, there was one startle-eliciting probe presented at the short ~120 ms! and long ~1,200 ms! lead intervals following the onset of the distractor tone and two startle probes presented in the absence of any prepulse stimuli, yielding a total of four probes. Stimulus Materials and Apparatus The CPT was administered using a computer with a NEC Multisync 3D monitor. The computerized version of the CPT used in this project was a modification of the 3–7 CPT program developed by Nuechterlein and Asarnow ~1993! and used in the UCLA longitudinal study of the early phases of schizophrenia ~Nuechterlein et al., 1992!. Specifically, CPT stimuli consisted of a series of single digits ~0–9! presented visually one at a time with an exposure time of 50 ms and an interstimulus interval of 1.65 s. As described earlier, this 3–7 version of the CPT employs a memoryload target ~3 followed by 7!. The computer recorded correct target detections, false positive responses, and reaction time in milliseconds. The auditory distractor stimulus was a 1000 Hz 70 dB~A! tone, 1,400 ms in duration with a controlled rise0fall time of 25 ms with the frequency controlled by the VCO input to a Coulbourn tone generator. The startle-eliciting stimulus was a 100 dB~A! white noise burst 40 ms in duration with a near instantaneous rise0fall time generated by a Grason Stadler Model 901B noise generator. All auditory distractors and startle-eliciting stimuli were presented through headphones ~Realistic NOVA-40 Model!. The onsets, durations, and intervals between stimuli were controlled on-line by ASYST computer software. Physiological responses were recorded using a Grass Model 7 polygraph equipped with a wide-band integrator0preamplifier ~Model 7P3! for the electromyographic ~EMG! recording of eyeblink responses and a Wheatstone Bridge ~e.g., see Venables & Christie, 1973! together with a DC preamplifier0driver amplifier ~Model 7P22! for the recording of skin conductance. Startle eye-blink was recorded as EMG from a pair of 4-mm silver–silver chloride electrodes filled with TECA electrode paste and placed directly over the orbicularis oculi muscle, the flexor muscle responsible for eyelid closure. These electrodes were placed

E.A. Hazlett et al. just below and to the left of the lower left eyelid. All impedances were less than or equal to 5 KV. The EMG signal was recorded at full wave rectification with an integration time constant of 20 ms. The eye-blink responses were digitized at a rate of 1000 Hz for a period of 200 ms before to 300 ms following the presentation of each startle-eliciting stimulus. Skin conductance levels ~SCLs! were recorded from the volar surface of the distal phalanges of the first and second fingers of the participant’s nonpreferred hand, using 10-mm silver–silver chloride electrodes and a 0.05 molar NaCl paste. Skin conductance levels were hand-scored from the paper record. Dependent Measures The three primary dependent variables were the modification of blink magnitude, skin conductance level, and behavioral performance on the CPT ~indexed by the transformed A9 score!. Startle blink magnitude was measured under five stimulus conditions: probe-alone presentations ~i.e. baseline magnitude! and probes presented at the two lead intervals following target, nontarget, target-plus-distractor, and distractor-alone presentations. Blink amplitude was computer scored off-line based on the criteria suggested by Balaban, Losito, Simons, & Graham ~1986!. Blink amplitude was scored in microvolt units for eye-blinks beginning within a latency window of 21 to 120 ms and reaching a peak within 150 ms of response onset. To examine blink modification across the three CPT blocks, separate baseline blink amplitudes were calculated for each CPT block. Specifically, the two baseline probes preceding and following each CPT block were averaged together, which resulted in three baseline mean values ~i.e. Baseline 1 5 mean of probes presented during baselines 1 and 2, Baseline 2 5 mean of probes presented during baselines 2 and 3, and Baseline 3 5 mean of probes presented during baselines 3 and 4!. The result of this averaging was that two of the four baseline startle probes for the Baseline 1 calculation were common to Baseline 2, and Baselines 2 and 3 had two common values as well. A few comments should be made regarding this strategy. First, this is not optimal because it creates a dependency between the three mean baseline values for blink magnitude. However, the presence of highly correlated baseline blink magnitudes is not unique to the present study. Typically, baseline blink magnitudes are highly intercorrelated, with contiguous blocks being more correlated ~e.g., blocks 1 and 2! than those that are not contiguous ~e.g., blocks 1 and 3!. Second, these correlations are corrected to some degree in statistical analyses involving repeated measures with more than two levels by using Greenhouse–Geisser epsilon corrections to adjust probabilities for repeated measures F values. Finally, this analytic strategy would not affect comparisons between the three prepulse types. The magnitude of each startle eye-blink elicited under a prepulse condition was expressed as a percent change score from the startle alone stimulus presented during the baseline periods. That is, for each of the three CPT blocks, the mean blink magnitude to startle probes presented during the respective baseline period was subtracted from the magnitude of the blinks elicited during the target, nontarget, and target-plus-distractor prepulses, resulting in a blink modification score for each critical lead interval, separately for each block. Similarly, for the distractor alone, the mean blink magnitude to startle stimuli alone presented during the same baseline period was subtracted from the magnitude of the blinks elicited during the distractor-alone prepulses. All scores were then converted to percent change from baseline values so as to control for individual variability in absolute blink amplitude. A positive

Attention and startle modification SEM score indicates startle facilitation relative to baseline, whereas a negative SEM score indicates startle inhibition relative to baseline. To track arousal levels during the baseline and CPT blocks and ensure that tonic arousal levels were similar, we recorded skin conductance levels. They were scored in microsiemens twice per minute for the entire session. More specifically, during the initial 3-min rest phase, SCL was scored at 30-s intervals, and throughout the remaining SEM-CPT experimental session, SCL was scored 3 s prior to every startle probe presented. Therefore, during each 2-min baseline period, because a total of four probes were presented, SCL was measured at approximately 30-s intervals. During each 4.5-min CPT block, because a total of 12 probes were presented, SCL was scored for only the eight probes that were in closest proximity to 30-s intervals. To determine behavioral performance on the CPT, the hit rate and false alarm rate were used to calculate A9 ~i.e. sensitivity! values across the three CPT blocks. In signal detection theory, “sensitivity” refers to an individual’s ability to discriminate the target ~signal! stimuli from the nontarget ~noise! stimuli independent of any response bias ~Green & Swets, 1966!. When an individual has a high perceptual sensitivity score he0she typically has a relatively high hit rate ~few errors of omission! and a relatively low false alarm rate ~few errors of commission!. A9 was calculated using the formula from Grier ~1971!: A9 5 0.5 1 ~HR 2 FAR!~1 1 HR 2 FAR!04~HR!~1 2 FAR!, where HR 5 hit rate and FAR 5 false alarm rate. Each subject’s A9 score was then transformed using the formula: 2 3 arcsin~#A9! in order to create a more normally distributed variable ~McNicol, 1972!. In addition, this transformation creates a variable with a nearly linear relationship to d9, which is another commonly used index of sensitivity ~McNicol, 1972!. These transformed A9 scores were used in all subsequent data analyses to index behavioral performance on the CPT. Statistical Analyses Analyses were performed using BMDP software ~BMDP, 1992!, and those analyses of variance effects with more than two levels of repeated measures used Greenhouse–Geisser epsilon corrections to adjust probabilities for repeated measures F values. The uncorrected degrees of freedom and epsilon values for these analyses are reported.

673 prepulses” and “distractor prepulse” findings are discussed separately. Baseline Startle Mean startle eye-blink amplitude to the startle alone stimulus presented during the baseline period was 22.4 mV with a SD of 22.8 mV. SEM During the CPT Prepulses Short lead interval SEM effects. The short lead interval SEM results for each prepulse type are shown in Figure 2. A Prepulse Type ~target, nontarget, target plus distractor! 3 Trial Block ~1 to 3! repeated-measures ANOVA on SEM during the CPT revealed a significant overall level of SEM, F~1,30! 5 116.75, p , .0001, indicating that the amount of blink inhibition averaged over both prepulse type and trial block was significantly different from zero. In addition, the main effects of Prepulse Type, F~2,60! 5 13.22, p , .0001, E 5 .9147, and Trial Block, F~2,60! 5 3.58, p , .05, E 5 .8286, were significant. As can be seen in Figure 2, the main effect of prepulse type was primarily due to greater blink inhibition produced by both the target and target-plus-distractor prepulses compared to the nontarget prepulse. The main effect of trial block reflects the finding that the overall amount of blink inhibition averaged over the three prepulses decreased across the three trial blocks. The Prepulse Type 3 Trial Block interaction did not approach significance, F~4,120! 5 0.97. Because the Prepulse Type 3 Trial Block interaction was nonsignificant, blink inhibition scores were averaged over the three trial blocks for further analysis. To test our specific a priori hypotheses about the effect of prepulse type, SEM during the target prepulse was compared with SEM during the nontarget prepulse and the target-plus-distractor prepulse using t tests. The target prepulse produced significantly greater blink inhibition than the nontarget prepulse, t~30! 5 3.69, p , .01. The target and target-plus-distractor prepulses were not different from each other, t~30! 5 0.68, p . .10. Long lead interval SEM effects. The long lead interval SEM results for each prepulse type are shown in Figure 3. To test for attentional modulation of long lead interval SEM, blink modification scores were submitted to a Prepulse Type ~target, nontarget, target plus distractor! 3 Trial Block ~1 to 3! repeated-measures

Results SEM during the CPT prepulses and SEM during the distractoralone prepulses were analyzed separately. The reason for conducting separate analyses was methodological in nature. The SEM scores for CPT prepulses were based on three trial blocks with each blink modification score for a trial block based on two presentations of the startle probe. The distractor-alone prepulses were presented during four baseline periods ~each CPT block was preceded and followed by a baseline period!. To have two presentations of the startle probe at each temporal position during the distractor-alone prepulse, scores from distractor-alone prepulses during baseline periods 1 and 2 were averaged together and scores from baseline periods 3 and 4 were averaged together. Thus, SEM scores for distractor prepulses were ultimately based on two trial blocks. Because the distractor prepulses were presented during the baseline period and not during the CPT, their SEM data were analyzed separately. In the following results section, the “CPT

Figure 2. Mean prepulse inhibition at the 120-ms lead interval during the CPT and baseline periods. The asterisk indicates a significant difference betweeen the target and nontarget prepulses.

674

E.A. Hazlett et al. p 5 .14. It is important to note that from the participant’s point of view, the baseline period was identical to the CPT except that occasionally there was a relatively longer time interval between CPT stimuli ~i.e. 3–24 s versus 1.65 s!. Behavioral Performance All of the participants performed well on the CPT based on their transformed A9 scores ~scores ranged from 2.63 to 3.03, M 5 2.94, SD 5 0.11!. Nine of the 31 participants ~29%! had a perfect score.

Figure 3. Mean startle eye-blink modification at the 1,200-ms lead interval during the CPT and baseline periods. A positive percent change score indicates facilitation and a negative score indicates inhibition.

ANOVA. The results revealed a significant overall level of SEM, F~1,30! 5 238.53, p , .0001, indicating that the amount of blink inhibition averaged over both prepulse type and trial block was significantly different from zero. The main effect of prepulse type and trial block failed to reach significance, as did the interaction effect ~all ps . .10!. As can be seen in Figure 3, the three prepulse types produced significant and nondifferential inhibition, suggesting that modality-specific selective attention may occur at 1,200 ms as participants anticipate the presentation of the next CPT prepulse. SEM During the Distractor Prepulse in Baseline Period Short lead interval SEM effects. A one-way analysis of variance was conducted on the SEM scores with trial block as a repeated measure to determine whether SEM was significant and whether or not habituation occurred. As can be seen in Figure 2, the results revealed a significant overall level of SEM, F~1,30! 5 129.63, p , .0001, indicating that the amount of blink inhibition averaged over trial block was significantly different from zero. There was a nonsignificant trend for the main effect of Trial Block, F~1,30! 5 3.61, p 5.07, indicating that the amount of prepulse inhibition decreased over trials. Long lead interval SEM effects. A one-way analysis of variance was conducted on the SEM scores with trial block as a repeated measure. As can be seen in Figure 3, the results revealed a significant overall level of SEM, F~1,30! 5 7.67, p , .01, indicating that the amount of blink facilitation averaged over trial block was significantly different from zero. The main effect of Trial Block failed to reach significance, F~1,30! 5 2.54, p 5 .12, indicating that the amount of blink facilitation during the first block was not significantly different from that observed during the second block. Skin Conductance Skin conductance levels were obtained per minute during rest, the three blocks of CPT, and the four interdigitated baseline periods. To rule out any differences between SEM during the distractor in the baseline and CPT periods being due to differences in tonic arousal, we conducted a paired t test on the mean SCL level during baseline and CPT periods. Mean SCL level during the baseline ~mean 5 8.63 mS, SD 5 3.01! and CPT periods ~mean 5 8.57 mS, SD 5 3.06! did not significantly differ, paired t~30! 5 1.50,

Relationship Between Behavioral Performance and SEM Pearson product moment correlations were computed between blink modification scores and CPT transformed A9 scores. The specific blink modification scores used in this analysis included the SEM scores at each lead interval for each CPT prepulse type as well as the SEM difference scores ~target minus nontarget prepulses! at each lead interval. All of these SEM scores were averaged over the three trial blocks. CPT transformed A9 scores were calculated based on performance during all three trial blocks. None of the correlations between SEM and the narrow range of behavioral performance represented here reached significance. Discussion The results in terms of the short and long lead interval SEM effects during the CPT can be summarized as follows: ~1! at the short lead interval, significant blink inhibition was produced by all three prepulse types; ~2! at the short lead interval, the target prepulse produced significantly greater blink inhibition than did the nontarget prepulse; ~3! at the short lead interval, the target and targetplus-distractor prepulses did not produce differential amounts of blink inhibition; ~4! at the long lead interval, significant but nondifferential blink inhibition was produced by all three prepulse types. The finding of overall blink inhibition at short lead intervals across the three prepulse types is consistent with a number of previous studies that employed discrete-trial active attention paradigms ~e.g., Filion et al., 1998; Hackley & Graham, 1987!. In addition, blink inhibition has previously been reported during a CPT paradigm; however, the probes were not time linked to particular CPT stimuli ~Zelson & Simons, 1986!. As described earlier, Graham ~1975! proposed the idea that short lead interval blink inhibition reflects “a wired-in negative feedback which reduces the distraction produced by reflexes such as startle, and thus protects what has been called preattentive stimulus processing” ~p. 246!. Consistent with this notion, the results of the present experiment revealed that overall blink inhibition occurred at the short lead interval. The finding of differential blink inhibition produced by the target and nontarget prepulses represents the first demonstration of attentional modulation of short lead interval SEM during a CPT, which is a relatively high processing load task when a memoryload or perceptual-load version is used ~Nuechterlein, 1991!. Furthermore, it suggests that early controlled processing is involved in stimulus discrimination during the CPT. As described earlier, several studies have reported attentional modulation of short lead interval SEM in discrete-trial paradigms ~for review, see Filion et al., 1998!. The results of the present study extend the previously reported short lead interval attentional modulation effect to a paradigm with a continuous stream of stimuli involving a high processing load and visual prepulses. The finding of differential blink inhibition produced by the target and nontarget prepulses is

Attention and startle modification consistent with the suggestion that blink inhibition reflects an obligatory stage of processing that is modifiable by controlled attentional processes ~Dawson et al., 1997!. The result of these processes is that the target prepulses receive greater protection of their initial processing ~i.e. greater PPI! compared with the nontarget prepulses. In addition, the finding of nondifferential blink inhibition produced by the target and target-plus-distractor prepulses suggests that the college student participants effectively ignored the distractor prepulse at the short lead interval. That is, participants in the present study exhibited the same amount of short lead interval inhibition during the visual target alone as during the visual target plus auditory distractor, indicating that attention was not redirected to the auditory modality. It is noteworthy that on the first trial block, there was a nonsignificant trend, t~30! 5 1.85, p 5 .07, for greater blink inhibition to the target-plus-distractor prepulse than to the target prepulse, and this effect disappeared across trial blocks. However, because the Prepulse Type 3 Trial Block interaction failed to reach significance, all subsequent analyses were conducted on SEM scores that were averaged across trial blocks for each of the prepulse types. When averaged over trial blocks, the apparent effect of more blink inhibition to the target-plus-distractor prepulse compared to the target prepulse was not confirmed statistically. To confirm that participants were successful in ignoring the distractor, we compared the level of response accuracy to target stimuli ~i.e. press after a 7 that was preceded by a 3! with the level of accuracy following target-plus-distractor stimuli ~i.e. press after a 7 that was preceded by a 3 simultaneous with the distractor tone!. In parallel with the prepulse inhibition findings, our behavioral data show that the subjects are equally accurate during these two conditions, further indicating that they effectively ignored the distractor tone and focused on the primary visual task. According to the Intermediate Theory ~Hackley, 1993!, early sensory-perceptual processes are obligatory and invariant with attention at brain-stem levels. Then, beginning at forebrain levels ~e.g., thalamus and cortex!, there is a transition from strong to partial automaticity. The findings in the present study indicate that this transition occurred by 120 ms because short lead interval SEM was modulated by controlled cognitive processes that are thought to be mediated by “top-down” influences from the cortex ~Dawson et al., 1997!. Consistent with this theory, functional neuroimaging work has demonstrated that increased glucose metabolism in prefrontal cortex is associated with greater prepulse inhibition during an attended prepulse ~Hazlett et al., 1998!. In the present study, the differential short lead interval blink inhibition to target and nontarget prepulses as well as nondifferential blink inhibition between the target and target-plus-distractor prepulses suggest the possibility that selective sensory attention both within and across modality can occur at early stages of information processing. The participants showed selective attention within the visual modality ~i.e. more blink inhibition to the visual target prepulse than to the visual nontarget prepulse!. In addition, nondifferential blink inhibition was observed between the target prepulse and the target-plusdistractor prepulse, suggesting that the participants effectively ignored the auditory prepulse that was in a modality different from the primary visual task. At the long lead interval, in addition to automatic arousal processes, SEM is also thought to reflect later controlled processes. Contrary to prediction, differential SEM did not occur at the long lead interval to the target and nontarget prepulses. It was predicted that efficient performance would be characterized by an attentional modulation pattern similar to that observed at the short

675 lead interval. Specifically, it was thought that the blink response following the startle stimulus would be most inhibited following a 3 ~which is the first stimulus in the two-sequence target! at the long lead interval because distraction at this point should have the most detrimental effect on CPT performance. That is, participants should be anticipating a visual target stimulus ~i.e. a 7! and screening out irrelevant auditory stimuli ~the startle stimulus!. However, the SEM pattern observed in the present study is more consistent with a generalized anticipatory attention strategy. That is, approximately 1,200 ms following the onset of a target or a nontarget prepulse, it appears that the participants are anticipating or preparing for the presentation of the next CPT prepulse ~which occurs quickly, i.e. 500 ms after the onset of the startle probe!. From the participant’s perspective, at the 1,200-ms probe position, they have already determined whether the preceding prepulse was a target or nontarget. If they just saw a target prepulse ~a 3!, then the next prepulse may be a 7 ~indicating a target sequence!. On the other hand, if they just saw a nontarget, then the next stimulus could be a target ~a 3!. Thus, the finding of significant yet nondifferential blink inhibition to the target and nontarget prepulse at the long lead interval suggests that the participants are preparing for the presentation of the next prepulse, which may potentially be a target. However, it is also important to point out that at the long lead interval, although a selective attention effect was not observed between the target and nontarget prepulses ~i.e. within-modality selective attention!, the participants were selectively attending to the visual modality in the target-plus-distractor condition and screening out the task-irrelevant distractor prepulse. This is evidenced by the fact that at the long lead interval, blink facilitation was observed to the distractor-alone prepulse; however, when the same distractor prepulse was presented together with the visual target prepulse, blink inhibition was observed. It is important to note that both direct as well as indirect attention can modulate SEM. On the one hand, attention to or away from the startle stimulus can directly modulate the startle eye-blink response. Consequently, screening out irrelevant auditory stimuli ~e.g., the startle stimuli! should make the resultant blinks smaller. On the other hand, attention to the prepulse ~e.g., the target prepulse! exerts an indirect effect on the eye-blink reflex. If the lead interval is appropriate to produce PPI, then attention to the target prepulse makes the blink smaller compared to when the prepulse is less important ~i.e. the nontarget prepulse!. The resultant SEM is thought to reflect the algebraic sum of indirect plus direct attention effects ~Balaban, Anthony, & Graham, 1985!. Consistent with previous work that suggests that long lead interval SEM provides a useful measure of cross-modal selective attention ~Putnam, 1990!, the participants in the present study were actively screening out the auditory distractors and attending to the visual target prepulse at the long lead interval during the CPT. In contrast with the present findings and those of others ~Anthony & Graham, 1985; Putnam, 1990!, Lipp, Siddle, and Dall ~1998! reported that attention to a long lead stimulus can enhance startle if the lead and startle-eliciting stimuli are presented in different sensory modalities. The findings of Lipp et al. are not consistent with the prediction, derived from modality-specific accounts of startle modification, that acoustic startle will be smaller during attentionengaging visual stimuli. However, an alternative interpretation is that modality-specific attentional effects on startle modification may be restricted to situations, such as those in the present study, in which the task demands are very high ~Lipp et al., 1998!. One might argue that in the present study, a myriad of processes other than attention ~e.g., arousal, working memory, response

676

E.A. Hazlett et al.

preparation! that might affect SEM are engaged 1,200 ms following the target prepulse compared with the nontarget prepulse. However, our finding of significant but nondifferential SEM following the target and nontarget prepulse at the 1,200-ms lead interval is inconsistent with this hypothesis. If one predicts that arousal or working memory is greater following the target prepulse, then one would expect that this processing would selectively increase SEM following the target prepulse compared with the nontarget prepulse and this result was not observed in our study. It might also be argued that as subjects are anticipating the presentation of the next CPT prepulse, they are motivated to suppress blinking so as not to miss a task-relevant stimulus. Consequently, the significant and nondifferential inhibition may have nothing to do with modality-specific attention. Rather, it may reflect a nonspecific motor inhibition within the eye-blink circuitry. To address the issue of nonspecific motor inhibition within the eye-blink circuitry, we went back to our raw data and scored the number of spontaneous blinks per minute during baseline periods ~2, 3, and 4! and CPT blocks ~1, 2, and 3!. If one hypothesizes that subjects are motivated to suppress blinking so as not to miss a task-relevant stimulus, then consequently, spontaneous blink rates during the CPT would be reduced compared with the rate observed during the baseline periods ~when the CPT task is less demanding due to intermittently longer CPT stimulus ISIs!. The mean number of spontaneous eye-blinks during the baseline periods for the 31 subjects was 21.8 ~SD 5 11.9! and the mean for the CPT blocks was 20.8 ~SD 5 13.1!. The mean difference score ~CPT minus baseline! was 21.03 ~SD 5 3.6!. A paired t test showed that these differences failed to reach significance, t~30! 5 21.60, p 5 0.12. Thus, the present data seem to be inconsistent with the idea that nonspecific motor inhibition within the eye-blink circuitry occurs during CPT performance. The absence of differential SEM at the 1,200-ms probe position appears to be more consistent with a modality-specific selective attention hypothesis. The present study did not find any significant correlations between SEM measures of attention and behavioral performance. This negative result is consistent with another study that examined visual evoked potentials during a similar version of the CPT ~Strandburg et al., 1990!. The ceiling effect in CPT performance for this group of college student participants mitigated against obtaining significant correlations in the present study. The range of transformed A9 scores was quite restricted ~M 5 2.94, SD 5 .11,

range: 2.63 to 3.03; 9 of 31 participants had a perfect score!, making the differential SEM to the prepulses across all subjects of more importance in this study than the individual differences in CPT behavioral performance.

Conclusion The present study has implications for the nature of cognitive subprocesses underlying both normal and abnormal CPT performance. The CPT has been used extensively in the study of various psychiatric populations, especially in studies concerned with the role of sustained attention deficits in schizophrenia ~see reviews by Nuechterlein, 1991, and Nuechterlein & Dawson, 1984!. Across several studies using various paradigmatic alterations, schizophrenia patients have consistently shown deficits in the sustained detection of brief stimuli presented at fairly brief interstimulus intervals ~e.g., 1 s! ~see Nuechterlein, 1991!. The CPT deficit in schizophrenia appears to reflect an inability to discriminate the target stimuli from the nontarget stimuli ~i.e. impaired sensitivity!, rather than a response criterion alteration. However, the exact nature of cognitive processes contributing to CPT deficits in schizophrenia is poorly understood. The present findings suggest that early controlled attentional processing following the target stimulus and later, modality-specific selective attention occur during and may be important for good CPT performance. Given the unique temporal resolution of SEM and its theoretical basis, this paradigm may help identify the cognitive processes ~automatic versus controlled! that are most impaired in schizophrenia as well as the point ~120 ms vs. 1,200 ms! in the sequence of processes activated by this task at which schizophrenia patients show abnormalities. Previous neuroimaging work with positron emission tomography indicates that schizophrenia patients have significantly lower glucose metabolism in the prefrontal cortex during a visual CPT task ~Buchsbaum et al., 1990!. Recent event-related f MRI data collected during an adapted attention-to-prepulse paradigm indicates that greater brain activation is observed in thalamic nuclei with projections to frontal regions following an attended than an ignored prepulse condition ~Hazlett et al., in press!. Future multidisciplinary work that integrates state-of-the-art neuroimaging and SEM techniques may be well-suited to integrate the information processing and neural circuitry abnormalities underlying CPT deficits in schizophrenia.

REFERENCES Anthony, B. J. ~1985!. In the blink of an eye: Implications of reflex modification for information processing. In P. K. Ackles, J. R. Jennings, & M. G. H. Coles ~Eds.!, Advances in psychophysiology ~Vol. 1, pp. 167–218!. Greenwich, CT: JAI Press. Anthony, B. J., & Graham, F. K. ~1985!. Blink reflex modification by selective attention: Evidence for the modulation of “automatic” processing. Biological Psychology, 21, 43–59. Balaban, M. T., Anthony, B. J., & Graham, F. K. ~1985!. Modalityrepetition and attentional effects on reflex blinking in infants and adults. Infant Behavior & Development, 8, 443– 457. Balaban, M. T., Losito, B. D. G., Simons, R. F., & Graham, F. K. ~1986!. Offline latency and amplitude scoring of the human reflex eyelink with Fortran IV. Psychophysiology, 23, 612. Berg, W. K., & Balaban, M. T. ~1999!. Startle elicitation: Stimulus parameters, recording techniques, and quantification. In M. E. Dawson, A. M. Schell, & A. H., Böhmelt ~Eds.!, Startle modification: Implications for neuroscience, cognitive science, and clinical science ~pp. 21–50!. Cambridge, UK: Cambridge University Press.

BMDP. ~1992!. BMDP biomedical statistical software. Los Angeles, CA: University of California. Buchsbaum, M. S., Nuechterlein, K. H., Haier, R. J., Wu, J., Sicotte, N., Hazlett, E., Asarnow, R., Potkin, S., & Guich, S. ~1990!. Glucose metabolic rate in normals and schizophrenics during the continuous performance test assessed by positron emission tomography. British Journal of Psychiatry, 156, 216–227. Dawson, M. E., Schell, A. M., & Böhmelt, A. H. ~1999!. Startle modification: Introduction and overview. In M. E. Dawson, A. M. Schell, & A. H. Böhmelt ~Eds.!, Startle modification: Implications for neuroscience, cognitive science, and clinical science ~pp. 6–18!. Cambridge, UK: Cambridge University Press. Dawson, M. E., Schell, A. M., & Filion, D. L. ~2000!. The electrodermal system. In J. T. Cacioppo, L. G. Tassinary, & G. G. Bentson ~Eds.!, Principles of psychophysiology: Physical, social, and inferential elements ~pp. 200–223!. Cambridge, UK: Cambridge University Press. Dawson, M. E., Schell, A. M., Swerdlow, N. R., & Filion, D. L. ~1997!. Cognitive, clinical, and neurophysiological implications of startle mod-

Attention and startle modification ification. In P. J. Lang, R. F., Simons, & M. T. Balaban ~Eds.!, Attention and orienting: Sensory and motivational processes ~pp. 257–275!. Hillsdale, NJ: Erlbaum. Filion, D. L., Dawson, M. E., & Schell, A. M. ~1993!. Modification of the acoustic startle-reflex eyeblink: A tool for investigating early and late attentional processes. Biological Psychology, 35, 185–200. Filion, D. L., Dawson, M. E., & Schell, A. M. ~1998!. The psychological significance of human startle eyeblink modification: A review. Biological Psychology, 47, 1– 43. Graham, F. K. ~1975!. The more or less startling effects of weak prestimulation. Psychophysiology, 12, 238–248. Graham, F. K., & Hackley, S. A. ~1991!. Passive and active attention to input. In J. R. Jennings & M. G. H. Coles ~Eds.!, Handbook of cognitive psychophysiology ~pp. 251–356!. New York: John Wiley. Green, D. M., & Swets, J. A. ~1966!. Signal detection theory and psychophysics. New York: Wiley. Grier, J. B. ~1971!. Non-parametric indexes for sensitivity and bias: Computing formulas. Psychological Bulletin, 79, 424– 429. Hackley, S. A. ~1993!. An evaluation of the automaticity of sensory processing using event-related potentials and brain-stem reflexes. Psychophysiology, 30, 415– 428. Hackley, S. A. ~1999!. Implications of blink reflex research for theories of attention and consciousness. In M. E. Dawson, A. M. Schell, & A. H. Böhmelt ~Eds.!, Startle modification: Implication for neuroscience, cognitive science, and clinical science ~pp. 137–156!. Cambridge, UK: Cambridge University Press. Hackley, S. A., & Graham, F. K. ~1987!. Effects of attending to the spatial position of reflex-eliciting and reflex modulating stimuli. Journal of Experimental Psychology: Human Perception and Performance, 13, 411– 424. Hackley, S. A., Woldorff, M., & Hillyard, S. A. ~1987!. Combined use of microreflexes and event-related brain potentials as measures of auditory selective attention. Psychophysiology, 24, 632– 647. Hazlett, E. A., Buchsbaum, M. S., Haznedar, M. M., Singer, M. B., Germans, M. K., Schnur, D. B., Jimenez, E. A., Buchsbaum, B. R., & Troyer, B. T. ~1998!. Prefrontal cortex glucose metabolism and startle eyeblink modification abnormalities in unmedicated schizophrenia patients. Psychophysiology, 35, 186–198. Hazlett, E. A., Buchsbaum, M. S., Tang, C. Y., Fleischman, M. B., Wei, T.-C., Byne, W., & Haznedar, M. M. ~in press!. Thalamic activation during an attention-to-prepulse startle modification paradigm: A functional MRI study. Biological Psychiatry. Lipp, O. V., Siddle, D. A. T., & Dall, P. J. ~1998!. Effects of stimulus modality and task condition on blink startle modification and on electrodermal responses. Psychophysiology, 35, 452– 461.

677 McNicol, D. ~1972!. A primer of signal detection theory ~pp. 118–119!. London: Allen and Unwin. Nuechterlein, K. H. ~1991!. Vigilance in schizophrenia and related disorders. In S. R. Steinhauer, J. H. Gruzelier, & J. Zubin ~Eds.!, Handbook of schizophrenia. Vol. 5: Neuropsychology, psychophysiology and information processing ~pp. 397– 433!. New York: Elsevier. Nuechterlein, K. H., & Asarnow, R. F. ~1993!. Continuous Performance Test ~CPT! Program for IBM-PC Compatible Microcomputers, 3–7 CPT Version. Los Angeles: Authors. Nuechterlein, K. H., & Dawson, M. E. ~1984!. Information processing and attentional functioning in the developmental course of schizophrenic disorders. Schizophrenia Bulletin, 10, 160–203. Nuechterlein, K. H., Dawson, M. E., Gitlin, M., Ventura, J., Goldstein, M. J., Snyder, K. S., Yee, C. M., & Mintz, J. ~1992!. Developmental processes in schizophrenic disorders: Longitudinal studies of vulnerability and stress. Schizophrenia Bulletin, 18, 387– 426. Nuechterlein, K. H., Edell, W. S., Norris, M., & Dawson, M. E. ~1986!. Attentional vulnerability indicators, thought disorder, and negative symptoms. Schizophrenia Bulletin, 12, 408– 426. Putnam, L. E. ~1990!. Great expectations: Anticipatory responses of the heart and brain. In R. Johnson, J. W. Rohrbaugh, R. Parasuraman, & R. Johnson, Jr. ~Eds.!, Event-related brain potentials: Basic issues and applications ~pp. 178–209!. New York: Oxford University Press. Putnam, L. E., & Vanman, E. J. ~1999!. Long lead interval startle modification. In M. E. Dawson, A. M. Schell, & A. H., Böhmelt ~Eds.!, Startle Modification: Implications for neuroscience, cognitive science, and clinical science ~pp. 72–92!. Cambridge, UK: Cambridge University Press. Strandburg, R. J., Marsh, J. T., Brown, W. S., Asarnow, R. F., Guthrie, D., & Higa, J. ~1990!. Event-related potential correlates of impaired attention in schizophrenic children. Biological Psychiatry, 27, 1103–1115. Venables, P. H., & Christie, M. J. ~1973!. Mechanisms, instrumentation, scoring techniques, and quantification of responses. In W. F. Prokasy & D. C. Raskin ~Eds.!, Electrodermal activity in psychological research ~pp. 1–124!. New York: Wiley. Zelson, M. F., & Simons, R. F. ~1986!. Sustained attention in type A and type B subjects: A blink reflex analysis. Psychophysiology, 23, 385–392.

~Received April 6, 2000; Accepted December 23, 2000!