International Journal of Psychopkysiologv, 13 (1992) 191-197 0 1992 Elsevier Science Publishers B.V. All rights reserved 0167-8760/92/$05.00
191
INTPSY 00406
Rick Howard a, Peter Fenwick b, Diana Brown ’ and Royan Norton d * Department of Psychological Medicine, University of C&ago, Dunedin (New Zealand), London (UK).
b Institute of &chiatry, University of Lortdon, ’ British Airways, London (UK) and d School of Psychology, The Queen’s University of Belfast, Belfast (Northern
Ireland)
(Accepted 19 May 1992)
Key words:
Lateral@; Contingent Negative Variation (CNV); Cognitive performance; Personality; Sex difference
This study used verbal and spatial variants of a cognitive ‘match/mismatch’ paradigm to explore relationships between functional brain asymmetries and individual differences in cognitive performance, personality and gender. Contingent negative variation (CNV) elicited in the ‘match/mismatch’ paradigm was recorded from central (Cz, C3 and C4) derivations in two male and one female samples. Results indicated gender differences in both the degree and direction of CNV asymmetries. Males showed a left hemisphere asymmetry in the verbal task and a right hemisphere asymmetry in the spatial task, with significant laterality (C3-C4) differences between tasks across the foreperiod. Females showed a left hemisphere asymmetry in both tasks with laterality differences between tasks confined to the early part of the foreperiod. CNV amplitude in the verbal task correlated with verbal memory performance and verbal I.Q. CNV amplitude in the spatial task correlated with visuospatial memory performance. Social extraversion was associated with greater left-hemisphere asymmetry in both !e~ks, while behavioural extraversion was associated with left-hemisphere asymmetry in the verbal task only, and with smaller verbal than spatial CNVs.
INTRODUCTION Previous studies have found that slow brain potential shifts such as the contingent negative variation. (CNV: Walter et al., 1964) and the stimulus preceding negativity (SPN: Brunia and Damen, 1988) have a task-specific topography. Thus, the SPN was reported by Roesler (1990) to show a left-hemisphere asymmetry in semantic tasks, but a right-hemisphere asymmetry in nonsemantic tasks. Similarly, the CNV elicited in a numerical ta-,k has been found to lateralise in the left hemisphere, while the CNV in a pattern matching task lateralised in the right hemisphere
Correspondence to: R. Howard, University of Otago Medical School. Department of Psychological Medicine, P.O. Box 913, Dunedin. New Zealand.
CBirbaumer et al., 1981). Pfurtscheller (1990) described lateralisation of event-related desynchronisation evoked during the fore-period in anticipation of a verbal imperative stimulus to be classified as old or new. Together these results suggest anticipatory states of cortical mobilisation in preparation for semantic and spatial information processing. They further offer the possibility that asymmetries in slow brain potential shifts might be usefully apphed to the study of individual differences. In a series of experiments reported here the authors have explored the use or verbal and spatial versions of a cognitive match/mismatch paradigm to examine the relationship between functional brain asymmetries and individual differences in cognitive performance, personality and gender. It would clearly be predicted, first, that CNVs elicited in verbal/semantic and spatial tasks re-
192
spectively would be related to individual differences in cognitive performance in verbal and visuospatial tasks. Second, given recent evidence that ‘psychopathy’ is related to deficits in semantic processing associated with left hemisphere function (Hare and McPherson, 1984), particularly in tasks which place heavy demands on semantic processes and memory (Hare and Jutai, 1988), one would expect CNV asymmetry in a verbal task to correlate with measures of the dimension, labelled ‘lmpulsivity/Bel1igerence’ by Blackburn (1974, 1980) and ‘P-Imp-USS’ by Zuckerman (1989), which describes the core traits of psychopathy. Finally, evidence has been accumulating which support the idea of sex differences in functional cerebral asymmetry, with investigators reporting either greater hemispheric specialisation for men than for women or sex differences in the direction of asymmetry (Day, 1977; Heister, 1984; McGlone, 1980). In order to explore these relationships, CNV was measured in verbal and spatial versions of a match/mismatch task in two male samples and one female sample. In sample 1, relationships were explored between CNV laterality and cognitive performance and personality. In sample 2 an attempt was made to replicate the laterality findings from sample 1. In sample 3, CNV laterality data were acquired in female subjects, in order to verify whether the laterality findings in males to explore any possible sex differences in laterality.
METHODS Subjects
(i) Sample 1. Subjects (n = 12: were recruited through an advertisement in a London weekly entertainment guide. They were all right-handed males with no previous history of head injury, seizures, drug abuse or serious illnesses. They were drug-free and reported no visual or hearing problems. Subjects gave their informed consent and were paid for participation. The mean age was 25.3 years, with a range from 20 to 36 years, and a standard deviation of 5.4 years. (ii) Sample 2. Subjects in sample 2 were all
male smokers (n = 8) who had been recruited into a study of the effects of icotine on lateralised cortical function (Norton and Howard, unpublished data). Subjects gave their informed consent and were paid for their participation. Their mean age was 25 years, with a standard deviation of 6.9 years. Two subjects were excluded on the basis of handedness, one being left-handed, the other ambidextrous, leaving six subjects whose data were analysed. All subjects were asked to refrain from smoking for 2 h before the experiment. Prior to recording subjects were asked about recent alcohol, caffeine and other drug intake. (iii) Sample 3. Sample 3 comprised 45 women with a mean age of 33.7 years (S.D. 7.8 years), recruited from the local community for a study of event-related potential changes during the menstrual cycle (Howard et al., 1992). All gave their informed consent to participate.
PROCEDURE (i) Sample I. On their arrival at 9.30 a.m. subjects were questioned about recent intake of alcohol, caffeine, nicotine and other drugs. Subjects who had taken any of these drugs were rejected. Silver/silver chloride electrodes were then attached to the scalp with collodion. Electrode placements were CZ, C3, and C4 referred to the tip of the nose. Eye movements were recorded from a midfrontal electrode referred to the outer canthus of the left eye. During electrode application, the subjects filied ia four personality inventories as follows: (i) Eysenck Personality Questionnaire (EPQ), comprising scales for Extraversion (El, Psychoticism (P) and Neuroticism (NJ. (ii) Special Hospitals Assess Personality and Socialisation (S burn, 1982), comprising the followmg SC (L); Extraversion (Ex: a measure of both social and behavioural extraversion); Anxiety (A); Psychopathic ate (Pd) from the MMPI; Introversion (I), a sure of sociai withdrawal; Depres); Tension CT); Aggression (Ag); Impulsiv(Im), a measure of behavioural extraversion; ostility (Ho), a measure of covert suspicion and
193
mistrust. (iii) Eysenck Impulsiveness Questionnaire (15), comprising scales for Impulsivity (Imp), Venturesomeness (Vent), and Empathy (Emp) (Eysenck and Eysenck 1978). (iv) Eysenck Personality Inventory (EPI). The Extraversion (E) scale from tke EPI yields sub-scales for social and avioural extraversion. The tests were always administered in the same order. Following electrode application and completion of questionnaires, subjects were transferred to a comfortable chair in the experimental room for CNV recordings. Subjects sat facing a back projection screen with a central red spot. They were instructed to focus on this spot during the CNV sessions zmd to try not to blink during trigir . .U.U” CNV was recorded during ‘verbal’ and ‘spa- : tial’ match/mismatch tasks (see below), the order of these being varied between subjects. Following each CNV recording session, zubjects completed a task related questionnaire in whieii they were asked to rate, on a scale from 1 to 10, task difficulty along with several other task related variables. Following this, the subject was debriefed and his eiectrodes were removed. Prior to the CNV recording subjects were given the Annett handedness questionnaire and tests of immediate recall on the Rey Qsterreith and Wechsler Logical Memory. In the afternoon following the CNV recording (which always took place in the morning), psychometric testing was carried out. This comprised delayed retention for the Wechsler Logical Memory and Rey-Osterreith, the Benton Visual Memory Test, four WAX Sub tests (Information, Vocabulary, Block Design and Object Assembly). W Surnpfe 2. All subjects were asked to refrain from smoking for 2 h before the experiment. On arrival they were asked about recent alcohol, caffeine and other drug intake. Silver/silver chloride electrodes were attached to the scalp with collodion. Electrode placements were Cz, C3, C4 and Fpz referred to linked mastoids, the frontal channel being included as a check for eye movement and blink artifact. Subjects were then transferred to a chair facing a back-projection screen for CNV recording. (iii) Smnpb 3. The procedure was the same as
for sample 2, except that subjects underwent recordings on three occasions, corresponding to different phases of the menstrual cycle (menstrual, premenstrual and mid-cycle). Recordings were carried out at the same time of day on each occasion. Data from left-handed subjects were excluded from further analysis. Match /M&natch task. CNV was recorded usinr tvo versions of a cognitive ‘match/mismatch’ task in which subjects were presented on each trial with a warning stimulus, Sl, which indicated that on occurence of S2 following an interval of 3.5 s, the subject should respond by pressing one of two buttons, according to whether S2 constituted a ‘match’ or a ‘mismatch’ with Sl. In both tasks Sl and S2 comprised slides projected on a back projection screen, placed 1.5 m. in front of the subject. Subjects held one button in each hand, and the allocation of ‘match’ and ‘mismatch’ buttons to left and right hands was randomised. In the verbal task the slides consisted of drawings of either living things or man-made objects. If both Sl and 52 fell into the same category (i.e., both man-made or both living things), the subject was to signal a ‘match’. If Sl and S2 fell into different categories, then a ‘mismatch’ was to be signalled. In the spatial task Sl and S2 comprised abstract geometric designs. A ‘match’ occurred when both designs were identical, while any difference between Sl and S2 constituted a ‘mismatch’. The duration of Sl was 1 s. S2 was presented for 6 s or until the subject responded by pressing the ‘match’ or ‘mismatch’ button. The inter-trial interval varied within the range 5-15 s. ‘Match’ and ‘mismatch’ trials were presented in a pseudo-random sequence. The recording, which generally lasted lo-15 min, ended when 16 artefact-free trials had been accumulated to form an average. CNVs for ‘match’ and fmismatch’ trials werl, averaged together since, as ‘match’/‘mismatch’ could not be predicted at Sl, no CNV differences would be expected. Acquisition and processing of EEG data. The signals were acquired using an 8 channel SLE EEG machine (for sample 1) and a Grass Neurodata Acquisition System (for samples 2 and 3). The low pass filter was set at 0.01 Hz and the high pass filter at 30 Hz. The gain was 100 PV to
194
give a 1 V output. The signals were processed using microprocessor based software, using the Data Acquisition Neurophysiology system (DAN) (;“&fi;istei et a:_ lrY.F?\. 17”J,. The sampling rate was 256 samples per see for each channel. All the parameters of the experimental stimuli were under software control. Digitised EEG signals were displayed on a monitor so that trials with eye-movement or blink artefacts (causing a deviation of > f 50 PV) could be rejected on-line. 16 trials were included in each average. The averaged CNV for each channel under each condition was stored. In addition to a point by point plot, the DAN software gave amplitude values for each of 12 consecutive points
on the CNV waveform. Each of these :2 CNV amplitude values represented the ms epoch relative to a 1 s pre Si baseline. The first epoch began 500 msecs. after S1 onset; the twelfth terminated at S2 onset. An integral measure which summated across the 12 points was also taken as a global measure of CNV amplitude. Averaged CNV waveforms from a representative subject from sample 3 are shown in Fig. 1. A Laterality Index (left-right) was computed for each CNV point, as well as for the integral, such that positive values indicated iateralisation of the CNV in the left hemisphere, while negative values indicated CNV lateralisation in the right hemisphere.
I
Fig. 1. CNVs
(averagesof
16 trials) recorded from a sample 3 subject in verbal (VER) and spatial (SPA) versions of the match/mismatch
task.
195
RESULTS
A. Sample1
5-
CNV lateral@ The CNV Iaterality measure is shown for each of the three samples in Fig. 2. It may be seen that re consistent differences in laterality between tasks, with the verbal task being lateralised in the left hemisphere. For sample 1, points 3-12 inclusive yielded significant between-task differences (P < 0.01, paired t-test), However, the females (sample 31 showed less clear between-task differences in iaterality compared with the males at central derivations (C3 and C4). For the female sample, there were clearly significant between-task differences at points l-5 inclusive (P < 0.01) and marginally significant (P < 0.05) differences at points 6, 7 and 9. Furthermore, females show a left hemisphere CNV asymmetry for both tasks, while males show a right hemisphere asymmetry in the spatial task (this was clearer in sample 2 than in sample I). Relationshipsbetween CiVV asymmetryand person ality Correlations between personality variables and CNV laterality are shown in Table I. It may be seen that Sy from the EPI, a measure of social extraversion, correlated negatively with laterality in both tasks, i.e., was associated with greater right than left hemisphere CNVs. Sy also correlated negatively with CNV amplitude (integral value) at C3 and Cz; r = -0.60 and -0.53, respectively. Measures of behavioural extraversion (Ex from the SHAPS and In;) correlated negatively with CNV laterality in the verbal task only; that is, more impulsive subjects showed greater right than left CNVs in the verbal task. Considering the effects of condition (verbal versus spatial task) on the CNV amplitudes of each placement, the Impulsiveness (Iml scale from the IS correlated highly with task differences in amplitude of the last CNV quarter, such that low impulsives showed high CNV amplitudes for the verbal compared to the spatial task at Cz, 0.74 (P < 0.01) and C4 0.81 (P < 0.01). The effect was smaller for the whole CNV integral (and significant only at Cr. and C4). There was a similar
UV
6. Samnle2
PV CNVpoints
C. Sampie 3
UV
I -5
CNV ppints
I
Fig. 2. CNV laterality (C3-C4) in sample 1 (A: II = 12 males); in sample 2 (B: n = 6 males) and in sample 3 (C: n = 45 females). Values for sample 3 were averaged across three different phases of the menstrual cycle (premenstrual, menstrual and mid-cycle). Positive laterality values indicate a left hemisphere asymmetry (L > R), negative values a right hemisphere asymmetry (R > L).
196 TABLE I Sample
I (n = 12): Correlations
between CW
laterality
(C3
to be greater than right for both tasks in subjects who found the spatial task difficult.
minus C4) and personality / subjectire measures
DISCUSSION
Task
Scale
spatial
verbal
sy(EPl)
-0.83
**
-0.62
Ex (SHAPS)
-0.71
**
ll.S.
Im (15)
-0.50
*
ll.S.
0.63 **
0.61 **
Task difficulty
*
* * P < 0.01. * P < 0.05.
correlation for the impulsivity subscale of the EPI E scale. Reiationships between CNV ad mance
cognitive perfor-
Correlations between CNV amplitude (integral measure) and cognitive performance measures are shown in Table II. Correlations were generally in the expected direction, that is, visuo-spatial memory correlated with CNV in the spatiail task. while verbal memory and verbal I.Q. correlated with CNV in the verbal task. Relationships between CNV asymmetry ami subjective measures
CNV laterality for both tasks correlated highly and significantly (P < 0.01) with the reported difthere was a tendency for left-hemisphere TABLE
II
Sample
I
tn = 12): Correlations
between CNV
CNVs
amphde
in
verbal and spatial tasks and cognitive performance measures Measrrre
Electrode
Ruy-Osterreith
c3
CZ
CJ
0.72 **“I
(I.75 **‘“I
11.60 *(”
immediate
memo9 Wechsler Logical Memo9 (immediate Verbal **
memory)
I.Q. (raw score)
0.59 *‘l”)
0.65 *(“‘)
0.58 *((t’
n.s.
0.63 ~(“1
0.5~ *(“!)
P
191
INTPSY 00406
Rick Howard a, Peter Fenwick b, Diana Brown ’ and Royan Norton d * Department of Psychological Medicine, University of C&ago, Dunedin (New Zealand), London (UK).
b Institute of &chiatry, University of Lortdon, ’ British Airways, London (UK) and d School of Psychology, The Queen’s University of Belfast, Belfast (Northern
Ireland)
(Accepted 19 May 1992)
Key words:
Lateral@; Contingent Negative Variation (CNV); Cognitive performance; Personality; Sex difference
This study used verbal and spatial variants of a cognitive ‘match/mismatch’ paradigm to explore relationships between functional brain asymmetries and individual differences in cognitive performance, personality and gender. Contingent negative variation (CNV) elicited in the ‘match/mismatch’ paradigm was recorded from central (Cz, C3 and C4) derivations in two male and one female samples. Results indicated gender differences in both the degree and direction of CNV asymmetries. Males showed a left hemisphere asymmetry in the verbal task and a right hemisphere asymmetry in the spatial task, with significant laterality (C3-C4) differences between tasks across the foreperiod. Females showed a left hemisphere asymmetry in both tasks with laterality differences between tasks confined to the early part of the foreperiod. CNV amplitude in the verbal task correlated with verbal memory performance and verbal I.Q. CNV amplitude in the spatial task correlated with visuospatial memory performance. Social extraversion was associated with greater left-hemisphere asymmetry in both !e~ks, while behavioural extraversion was associated with left-hemisphere asymmetry in the verbal task only, and with smaller verbal than spatial CNVs.
INTRODUCTION Previous studies have found that slow brain potential shifts such as the contingent negative variation. (CNV: Walter et al., 1964) and the stimulus preceding negativity (SPN: Brunia and Damen, 1988) have a task-specific topography. Thus, the SPN was reported by Roesler (1990) to show a left-hemisphere asymmetry in semantic tasks, but a right-hemisphere asymmetry in nonsemantic tasks. Similarly, the CNV elicited in a numerical ta-,k has been found to lateralise in the left hemisphere, while the CNV in a pattern matching task lateralised in the right hemisphere
Correspondence to: R. Howard, University of Otago Medical School. Department of Psychological Medicine, P.O. Box 913, Dunedin. New Zealand.
CBirbaumer et al., 1981). Pfurtscheller (1990) described lateralisation of event-related desynchronisation evoked during the fore-period in anticipation of a verbal imperative stimulus to be classified as old or new. Together these results suggest anticipatory states of cortical mobilisation in preparation for semantic and spatial information processing. They further offer the possibility that asymmetries in slow brain potential shifts might be usefully apphed to the study of individual differences. In a series of experiments reported here the authors have explored the use or verbal and spatial versions of a cognitive match/mismatch paradigm to examine the relationship between functional brain asymmetries and individual differences in cognitive performance, personality and gender. It would clearly be predicted, first, that CNVs elicited in verbal/semantic and spatial tasks re-
192
spectively would be related to individual differences in cognitive performance in verbal and visuospatial tasks. Second, given recent evidence that ‘psychopathy’ is related to deficits in semantic processing associated with left hemisphere function (Hare and McPherson, 1984), particularly in tasks which place heavy demands on semantic processes and memory (Hare and Jutai, 1988), one would expect CNV asymmetry in a verbal task to correlate with measures of the dimension, labelled ‘lmpulsivity/Bel1igerence’ by Blackburn (1974, 1980) and ‘P-Imp-USS’ by Zuckerman (1989), which describes the core traits of psychopathy. Finally, evidence has been accumulating which support the idea of sex differences in functional cerebral asymmetry, with investigators reporting either greater hemispheric specialisation for men than for women or sex differences in the direction of asymmetry (Day, 1977; Heister, 1984; McGlone, 1980). In order to explore these relationships, CNV was measured in verbal and spatial versions of a match/mismatch task in two male samples and one female sample. In sample 1, relationships were explored between CNV laterality and cognitive performance and personality. In sample 2 an attempt was made to replicate the laterality findings from sample 1. In sample 3, CNV laterality data were acquired in female subjects, in order to verify whether the laterality findings in males to explore any possible sex differences in laterality.
METHODS Subjects
(i) Sample 1. Subjects (n = 12: were recruited through an advertisement in a London weekly entertainment guide. They were all right-handed males with no previous history of head injury, seizures, drug abuse or serious illnesses. They were drug-free and reported no visual or hearing problems. Subjects gave their informed consent and were paid for participation. The mean age was 25.3 years, with a range from 20 to 36 years, and a standard deviation of 5.4 years. (ii) Sample 2. Subjects in sample 2 were all
male smokers (n = 8) who had been recruited into a study of the effects of icotine on lateralised cortical function (Norton and Howard, unpublished data). Subjects gave their informed consent and were paid for their participation. Their mean age was 25 years, with a standard deviation of 6.9 years. Two subjects were excluded on the basis of handedness, one being left-handed, the other ambidextrous, leaving six subjects whose data were analysed. All subjects were asked to refrain from smoking for 2 h before the experiment. Prior to recording subjects were asked about recent alcohol, caffeine and other drug intake. (iii) Sample 3. Sample 3 comprised 45 women with a mean age of 33.7 years (S.D. 7.8 years), recruited from the local community for a study of event-related potential changes during the menstrual cycle (Howard et al., 1992). All gave their informed consent to participate.
PROCEDURE (i) Sample I. On their arrival at 9.30 a.m. subjects were questioned about recent intake of alcohol, caffeine, nicotine and other drugs. Subjects who had taken any of these drugs were rejected. Silver/silver chloride electrodes were then attached to the scalp with collodion. Electrode placements were CZ, C3, and C4 referred to the tip of the nose. Eye movements were recorded from a midfrontal electrode referred to the outer canthus of the left eye. During electrode application, the subjects filied ia four personality inventories as follows: (i) Eysenck Personality Questionnaire (EPQ), comprising scales for Extraversion (El, Psychoticism (P) and Neuroticism (NJ. (ii) Special Hospitals Assess Personality and Socialisation (S burn, 1982), comprising the followmg SC (L); Extraversion (Ex: a measure of both social and behavioural extraversion); Anxiety (A); Psychopathic ate (Pd) from the MMPI; Introversion (I), a sure of sociai withdrawal; Depres); Tension CT); Aggression (Ag); Impulsiv(Im), a measure of behavioural extraversion; ostility (Ho), a measure of covert suspicion and
193
mistrust. (iii) Eysenck Impulsiveness Questionnaire (15), comprising scales for Impulsivity (Imp), Venturesomeness (Vent), and Empathy (Emp) (Eysenck and Eysenck 1978). (iv) Eysenck Personality Inventory (EPI). The Extraversion (E) scale from tke EPI yields sub-scales for social and avioural extraversion. The tests were always administered in the same order. Following electrode application and completion of questionnaires, subjects were transferred to a comfortable chair in the experimental room for CNV recordings. Subjects sat facing a back projection screen with a central red spot. They were instructed to focus on this spot during the CNV sessions zmd to try not to blink during trigir . .U.U” CNV was recorded during ‘verbal’ and ‘spa- : tial’ match/mismatch tasks (see below), the order of these being varied between subjects. Following each CNV recording session, zubjects completed a task related questionnaire in whieii they were asked to rate, on a scale from 1 to 10, task difficulty along with several other task related variables. Following this, the subject was debriefed and his eiectrodes were removed. Prior to the CNV recording subjects were given the Annett handedness questionnaire and tests of immediate recall on the Rey Qsterreith and Wechsler Logical Memory. In the afternoon following the CNV recording (which always took place in the morning), psychometric testing was carried out. This comprised delayed retention for the Wechsler Logical Memory and Rey-Osterreith, the Benton Visual Memory Test, four WAX Sub tests (Information, Vocabulary, Block Design and Object Assembly). W Surnpfe 2. All subjects were asked to refrain from smoking for 2 h before the experiment. On arrival they were asked about recent alcohol, caffeine and other drug intake. Silver/silver chloride electrodes were attached to the scalp with collodion. Electrode placements were Cz, C3, C4 and Fpz referred to linked mastoids, the frontal channel being included as a check for eye movement and blink artifact. Subjects were then transferred to a chair facing a back-projection screen for CNV recording. (iii) Smnpb 3. The procedure was the same as
for sample 2, except that subjects underwent recordings on three occasions, corresponding to different phases of the menstrual cycle (menstrual, premenstrual and mid-cycle). Recordings were carried out at the same time of day on each occasion. Data from left-handed subjects were excluded from further analysis. Match /M&natch task. CNV was recorded usinr tvo versions of a cognitive ‘match/mismatch’ task in which subjects were presented on each trial with a warning stimulus, Sl, which indicated that on occurence of S2 following an interval of 3.5 s, the subject should respond by pressing one of two buttons, according to whether S2 constituted a ‘match’ or a ‘mismatch’ with Sl. In both tasks Sl and S2 comprised slides projected on a back projection screen, placed 1.5 m. in front of the subject. Subjects held one button in each hand, and the allocation of ‘match’ and ‘mismatch’ buttons to left and right hands was randomised. In the verbal task the slides consisted of drawings of either living things or man-made objects. If both Sl and 52 fell into the same category (i.e., both man-made or both living things), the subject was to signal a ‘match’. If Sl and S2 fell into different categories, then a ‘mismatch’ was to be signalled. In the spatial task Sl and S2 comprised abstract geometric designs. A ‘match’ occurred when both designs were identical, while any difference between Sl and S2 constituted a ‘mismatch’. The duration of Sl was 1 s. S2 was presented for 6 s or until the subject responded by pressing the ‘match’ or ‘mismatch’ button. The inter-trial interval varied within the range 5-15 s. ‘Match’ and ‘mismatch’ trials were presented in a pseudo-random sequence. The recording, which generally lasted lo-15 min, ended when 16 artefact-free trials had been accumulated to form an average. CNVs for ‘match’ and fmismatch’ trials werl, averaged together since, as ‘match’/‘mismatch’ could not be predicted at Sl, no CNV differences would be expected. Acquisition and processing of EEG data. The signals were acquired using an 8 channel SLE EEG machine (for sample 1) and a Grass Neurodata Acquisition System (for samples 2 and 3). The low pass filter was set at 0.01 Hz and the high pass filter at 30 Hz. The gain was 100 PV to
194
give a 1 V output. The signals were processed using microprocessor based software, using the Data Acquisition Neurophysiology system (DAN) (;“&fi;istei et a:_ lrY.F?\. 17”J,. The sampling rate was 256 samples per see for each channel. All the parameters of the experimental stimuli were under software control. Digitised EEG signals were displayed on a monitor so that trials with eye-movement or blink artefacts (causing a deviation of > f 50 PV) could be rejected on-line. 16 trials were included in each average. The averaged CNV for each channel under each condition was stored. In addition to a point by point plot, the DAN software gave amplitude values for each of 12 consecutive points
on the CNV waveform. Each of these :2 CNV amplitude values represented the ms epoch relative to a 1 s pre Si baseline. The first epoch began 500 msecs. after S1 onset; the twelfth terminated at S2 onset. An integral measure which summated across the 12 points was also taken as a global measure of CNV amplitude. Averaged CNV waveforms from a representative subject from sample 3 are shown in Fig. 1. A Laterality Index (left-right) was computed for each CNV point, as well as for the integral, such that positive values indicated iateralisation of the CNV in the left hemisphere, while negative values indicated CNV lateralisation in the right hemisphere.
I
Fig. 1. CNVs
(averagesof
16 trials) recorded from a sample 3 subject in verbal (VER) and spatial (SPA) versions of the match/mismatch
task.
195
RESULTS
A. Sample1
5-
CNV lateral@ The CNV Iaterality measure is shown for each of the three samples in Fig. 2. It may be seen that re consistent differences in laterality between tasks, with the verbal task being lateralised in the left hemisphere. For sample 1, points 3-12 inclusive yielded significant between-task differences (P < 0.01, paired t-test), However, the females (sample 31 showed less clear between-task differences in iaterality compared with the males at central derivations (C3 and C4). For the female sample, there were clearly significant between-task differences at points l-5 inclusive (P < 0.01) and marginally significant (P < 0.05) differences at points 6, 7 and 9. Furthermore, females show a left hemisphere CNV asymmetry for both tasks, while males show a right hemisphere asymmetry in the spatial task (this was clearer in sample 2 than in sample I). Relationshipsbetween CiVV asymmetryand person ality Correlations between personality variables and CNV laterality are shown in Table I. It may be seen that Sy from the EPI, a measure of social extraversion, correlated negatively with laterality in both tasks, i.e., was associated with greater right than left hemisphere CNVs. Sy also correlated negatively with CNV amplitude (integral value) at C3 and Cz; r = -0.60 and -0.53, respectively. Measures of behavioural extraversion (Ex from the SHAPS and In;) correlated negatively with CNV laterality in the verbal task only; that is, more impulsive subjects showed greater right than left CNVs in the verbal task. Considering the effects of condition (verbal versus spatial task) on the CNV amplitudes of each placement, the Impulsiveness (Iml scale from the IS correlated highly with task differences in amplitude of the last CNV quarter, such that low impulsives showed high CNV amplitudes for the verbal compared to the spatial task at Cz, 0.74 (P < 0.01) and C4 0.81 (P < 0.01). The effect was smaller for the whole CNV integral (and significant only at Cr. and C4). There was a similar
UV
6. Samnle2
PV CNVpoints
C. Sampie 3
UV
I -5
CNV ppints
I
Fig. 2. CNV laterality (C3-C4) in sample 1 (A: II = 12 males); in sample 2 (B: n = 6 males) and in sample 3 (C: n = 45 females). Values for sample 3 were averaged across three different phases of the menstrual cycle (premenstrual, menstrual and mid-cycle). Positive laterality values indicate a left hemisphere asymmetry (L > R), negative values a right hemisphere asymmetry (R > L).
196 TABLE I Sample
I (n = 12): Correlations
between CW
laterality
(C3
to be greater than right for both tasks in subjects who found the spatial task difficult.
minus C4) and personality / subjectire measures
DISCUSSION
Task
Scale
spatial
verbal
sy(EPl)
-0.83
**
-0.62
Ex (SHAPS)
-0.71
**
ll.S.
Im (15)
-0.50
*
ll.S.
0.63 **
0.61 **
Task difficulty
*
* * P < 0.01. * P < 0.05.
correlation for the impulsivity subscale of the EPI E scale. Reiationships between CNV ad mance
cognitive perfor-
Correlations between CNV amplitude (integral measure) and cognitive performance measures are shown in Table II. Correlations were generally in the expected direction, that is, visuo-spatial memory correlated with CNV in the spatiail task. while verbal memory and verbal I.Q. correlated with CNV in the verbal task. Relationships between CNV asymmetry ami subjective measures
CNV laterality for both tasks correlated highly and significantly (P < 0.01) with the reported difthere was a tendency for left-hemisphere TABLE
II
Sample
I
tn = 12): Correlations
between CNV
CNVs
amphde
in
verbal and spatial tasks and cognitive performance measures Measrrre
Electrode
Ruy-Osterreith
c3
CZ
CJ
0.72 **“I
(I.75 **‘“I
11.60 *(”
immediate
memo9 Wechsler Logical Memo9 (immediate Verbal **
memory)
I.Q. (raw score)
0.59 *‘l”)
0.65 *(“‘)
0.58 *((t’
n.s.
0.63 ~(“1
0.5~ *(“!)
P
