ESBRA–NORDMANN 1998 AWARD LECTURE: VISUAL P3 AS A POTENTIAL VULNERABILITY MARKER OF ALCOHOLISM: EVIDENCE FROM THE AMSTERDAM STUDY OF CHILDREN OF ALCOHOLICS

Odin Van Der Stelt

Department of Child and Adolescent Psychiatry, Academic Medical Center, University of Amsterdam, P.O. Box 22700, 1100 DE Amsterdam, The Netherlands

Received 16 October 1998; )

ABSTRACT

Recent data from the Amsterdam Study of Children of Alcoholics add to the evidence for considering the P300 or P3 component of the event-related potential (ERP) as a potential vulnerability marker of alcoholism. In this study, multi-channel ERPs were recorded from 7- to 18-year-old children of alcoholics (COAs) and age- and sex-matched low-risk controls using several experimental paradigms, including a visual novelty oddball task and a visual selective attention task. The results indicated that differences between COAs and controls in the visual P3 amplitude: (1) can be elicited both actively by task-relevant target stimuli and passively by irrelevant novel stimuli; (2) are a function of both the attentional relevance and the target properties of the eliciting stimulus; (3) are mediated by multiple brain generators, rather than by a single generator; (4) originate from a difference in the strength, rather than in the spatial configuration, of the underlying brain generators; (5) cannot be accounted for by differences in visual attention-related earlier occurring ERP components; and (6) can be moderated by current behavioural and emotional problems, general intellectual ability, and socio-economic background. These findings support the notion that a relatively small visual P3 amplitude in COAs reflects heritable biases in attention and information processing that are related to their increased vulnerability to alcoholism.

INTRODUCTION

A major thrust in contemporary research on alcoholism is the search for behavioural and biological markers that reflect a genetic vulnerability to the development of this disorder (Devor and Cloninger, 1989Go). A trait marker, genetic risk factor, or vulnerability marker for an illness may be defined as ‘a heritable trait, associated with a causative pathophysiologic factor in an inherited disease’ (Gershon and Goldin, 1986Go). Accordingly, research on vulnerability markers may provide basic insight into the pathophysiological mechanism that is genetically transmitted in alcoholism, even though the exact mode of transmission is unknown and the specific gene(s) involved have not been identified. Also, this research holds promise for contributing to the development of more effective and more efficient methods for the prevention and treatment of alcoholism.

Over the past two decades, a wide variety of psychological, psychophysiological, and biochemical characteristics have been assessed in human subjects, either with or without alcohol challenges, in an attempt to identify a phenotypic marker of alcoholism vulnerability (for reviews, see Begleiter and Porjesz, 1988; Schuckit, 1989, 1994; Pihl et al., 1990; Sher, 1991; Couzigou et al., 1993). One of the phenotypic traits assessed involves the non-invasive recording of brain electrical activity as indexed by the P300 or P3 component of the event-related potential (ERP).

The P3 refers to a late ERP component, occurring with a latency of 300 ms or more after stimulus presentation, which can be detected using signal-averaging techniques in the scalp-recorded electroencephalogram (EEG). The P3 is usually elicited in an auditory or visual ‘oddball’ paradigm when a subject detects an infrequent but task-relevant stimulus event (target) randomly interspersed among more frequent standard stimuli. In contrast to early occurring sensory-evoked or exogenous ERP components, such as P1 or N1, the P3 is relatively insensitive to the physical properties of the stimulus. The P3 generation is determined by the psychological context of the eliciting stimulus, being dependent on instruction and the nature of the stimulus and task given to the subject. Thus, the P3 is considered a cognitive-related or endogenous ERP component, because its elicitation is dependent on the active cognitive processing of stimulus information on the part of the subject (Donchin et al., 1978Go; Picton and Hillyard, 1988Go). Although the functional role of the P3 in human information processing remains to be fully understood, this ERP component is most often associated with attentional and memory processes (e.g. Polich, 1991), processes that involve multiple regions of the brain (Picton, 1992Go; Johnson, 1993Go).

The P3 should meet four criteria in order to be considered as a vulnerability marker of alcoholism. The P3 should be: (1) associated with alcoholism, or with a given subtype, in the general population; (2) heritable; (3) state-independent; and (4) associated with alcoholism within pedigrees. The rationale for these requirements is to rule out false positive or false negative conclusions on the validity of the P3 as a vulnerability marker of alcoholism due to certain conditions, particularly secondary effects of alcoholism or treatment, population stratification, or genetic heterogeneity, which can obscure the association of a genetic risk factor with a clinical syndrome (Gershon and Goldin, 1986Go; Goldin et al., 1986Go).

Available evidence indicates that the P3 may satisfy each of the four criteria for the identification of a vulnerability marker of alcoholism (for reviews, see Porjesz et al., 1996; Van der Stelt, 1997, 1998). More specifically, the P3 size or amplitude recorded at baseline, without an alcohol challenge, has been found to be: (1) smaller in alcoholics than in non-alcoholic unrelated controls (Porjesz and Begleiter, 1983Go, 1995Go; Begleiter and Porjesz, 1988Go; Pfefferbaum et al., 1991Go; Cohen et al., 1995Go; Porjesz et al., 1996Go); (2) mediated by genetic factors, rather than by environmental factors (Aston and Hill, 1990Go; O'Connor et al., 1994Go; Porjesz et al., 1996Go; Katsanis et al., 1997Go); (3) smaller in individuals at high risk for alcoholism than in low-risk controls (Begleiter et al., 1984Go; O'Connor et al., 1987Go; Whipple et al., 1988Go, 1991Go; Porjesz and Begleiter, 1990Go; Hill and Steinhauer, 1993Go; Steinhauer and Hill, 1993Go; Polich et al., 1994Go; Benegal et al., 1995Go; Hill et al., 1995aGo; Porjesz et al., 1996Go; Ramachandran et al., 1996Go; Cohen et al., 1997Go; Sharma et al., 1997Go; Van der Stelt et al., 1998aGo,bGo); (4) associated with alcoholism within densely affected families of alcoholics (Porjesz et al., 1996Go).

The most compelling current evidence for considering the P3 as a potential vulnerability marker of alcoholism has come from investigations of populations at high and low risk for alcoholism. These studies have shown that a reduction in the P3 amplitude, initially observed in alcoholics (Porjesz and Begleiter, 1983Go), can also be observed in non-alcoholic young adult individuals who by reason of a family history of alcoholism are at high risk for the future development of alcoholism (Elmasian et al., 1982Go; O'Connor et al., 1987Go; Porjesz and Begleiter, 1990Go; Benegal et al., 1995Go; Porjesz et al., 1996Go; Ramachandran et al., 1996Go; Cohen et al., 1997Go). Moreover, a reduced P3 amplitude has been found to characterize young children of alcoholics with no or only minimal drinking histories (Begleiter et al., 1984Go; Whipple et al., 1988Go, 1991Go; Hill and Steinhauer, 1993Go; Steinhauer and Hill, 1993Go; Hill et al., 1995aGo; Sharma et al., 1997Go; Van der Stelt et al., 1998aGo,bGo), and has been shown also to predict prospectively later substance use (Berman et al., 1993Go) as well as alcoholism (Hill et al., 1995bGo). These results strongly suggest that a reduced P3 amplitude might be a genetically mediated individual characteristic that precedes the development of alcoholism, rather than being acquired as a consequence of alcohol use. That is, the P3 amplitude might provide a trait or vulnerability marker, rather than a state marker, of alcoholism.

In this paper, data from the Amsterdam Study of Children of Alcoholics (Gunning et al., 1994Go; Van der Stelt et al., 1994Go) will be reviewed, which add to the evidence for considering the P3 as a putative vulnerability marker of alcoholism. In this study, multi-channel ERPs were recorded from 7- to 18-year-old children of alcoholics (COAs) and age- and sex-matched low-risk controls using several experimental paradigms. The major aims of the study were: (1) a detailed assessment of the P3 as well as of several other, earlier occurring endogenous ERP components in COAs and controls; (2) identification of ‘moderator’ variables, that is, variables that can affect the direction and/or strength of the family history–ERP relationship under study. The study participants, methods, and results have been described in detail elsewhere (Van der Stelt, 1997Go; Van der Stelt et al., 1997Go, 1998aGo, Van der Stelt et al., bGo), and are now summarized below.

SUBJECTS AND METHODS

Subjects
Participants were children at high risk (n = 50; 24 boys and 26 girls) and low risk (n = 50; 24 boys and 26 girls) for alcoholism. At the time of enrolment in the study (1993–1996), the children were aged 7.5 to 18.5 years. They were subdivided into two age groups: 7- to 12-year-olds (mean age ± SD = 9.7 ± 1.6 years) and 13- to 18-year-olds (mean age = 15.7 ± 1.8 years). The high-risk children within each age group were matched to the low-risk controls by age, sex, and handedness. The sample of high-risk children consisted of COAs. They were recruited through one of their biological parents, who was undergoing treatment for alcohol dependence or abuse according to DSM-III-R criteria (American Psychiatric Association, 1987Go). Most of the alcoholic parents (76.5%) had diagnosed psychiatric co-morbidity, mainly depression and anxiety disorders. To control for possible confounding fetal alcohol effects, the alcoholic mothers were interviewed about their use of alcohol and other substances during pregnancy. The COAs also underwent a physical examination to screen for neurological abnormalities and dysmorphic signs characteristic of fetal alcohol effects (Clarren et al., 1987Go). The low-risk children were recruited from announcements in local schools and via newspaper advertisements. They were included only if none of their first- and second-degree biological relatives had a personal history of alcoholism or other substance abuse, as assessed by a face-to-face parental interview. All high- and low-risk children included were free of medical problems, had no history of neurological disorder or head trauma, and had no history of alcoholism or abuse of other drugs or substances. The children were noted to have no or only minimal drinking histories. They were drawn predominantly from families of Caucasian racial background. The research protocol was approved by the Ethics Committee of the Academic Medical Center of the University of Amsterdam.

The COAs were observed to manifest significantly more externalizing problems (e.g. aggression) and more internalizing problems (e.g. anxiety/ depression, somatic complaints) than controls, as assessed by parental ratings using the Child Behavior Checklist (CBCL; Achenbach, 1991), though the range of group means did not exceed one standard deviation. Additionally, COAs obtained normal, but significantly lower, scores on the Standard Progressive Matrices (SPM; Raven et al., 1992) than controls, a psychometric test for problem solving consistently associated with estimates of general intelligence (Raven et al., 1993Go; Gustafsson, 1994Go). Finally, although both groups of children were drawn from parents mainly of the middle socio-economic stratum, COA families obtained significantly lower socio-economic status (SES) scores than control families. In the light of the group differences in externalizing and internalizing problem scores, SPM performance, and SES, the effects of these variables on the performance and ERP measures were assessed.

Experimental procedure
Participants were tested individually in a single experimental session, utilizing a passive (non-task) auditory paradigm (Van der Stelt et al., 1997Go), a visual novelty oddball task (Van der Stelt et al., 1998aGo), and a visual selective attention task (Van der Stelt et al., 1998bGo). In the auditory paradigm, ERPs were elicited while the child was reading a self-chosen book by infrequent tone pips differing in either frequency or duration from standard tones. This paradigm was used to assess in a subset of the children the mismatch negativity (MMN), an early occurring endogenous ERP component associated with automatic, pre-attentive discriminative acoustic processes at the level of the sensory (auditory) cortex. As reported previously (Van der Stelt et al., 1997Go), the results revealed no significant differences between COAs and controls in the MMN peak latency, amplitude, and scalp topography, and these data will not be discussed further here.

In the visual novelty oddball task, the children were exposed to three types of stimuli, referred to as non-target, target, and novel stimuli. These stimuli were delivered successively in a random order at regular intervals of 1300 ms for 100 ms on a monitor. Target and non-target stimuli were either the letter ‘X’ or ‘O’. Novel stimuli consisted of complex, colourful, abstract patterns that were different on each trial. Initially, the children received one block of practice trials, including non-target (88%) and target (12%) stimuli. They were instructed to respond with a right-finger button-press response only to the target stimuli. Next, the children were presented with four blocks of experimental trials, including non-target (76%), target (12%), and novel (12%) stimuli. Participants were not informed that the novel stimuli would be presented. This paradigm basically served to evaluate the P3 elicited actively by the task-relevant, target stimuli, as well as probing the P3 elicited passively by the irrelevant but ‘intrinsically’ salient, attention-capturing novel stimuli.

Finally, in the visual selective attention task, the children were presented red and blue circles, referred to as relevant (or attended) and irrelevant (or unattended) stimuli. Half of the circles had a small gap, the gap being distributed randomly over four possible locations (0°, 90°, 180°, and 270°); these stimuli served as non-targets. The other half of the circles had no gaps; these stimuli served as targets. The children were instructed to ignore the irrelevant (e.g. red) stimuli, and to focus attention on the relevant (e.g. blue) stimuli in order to detect the presence of target stimuli. The colour-selection cue (red or blue) was counterbalanced over subjects. Targets were presented among both relevant and irrelevant stimuli. Accordingly, the children received four types of stimuli, relevant target (25%), relevant non-target (25%), irrelevant target (25%), and irrelevant non-target (25%) stimuli, with only the relevant target stimuli requiring a button-press response. The stimuli were presented successively in a random order, with interstimulus interval varying randomly between 1300 and 1900 ms, for a duration of 100 ms. This paradigm was used to assess the P3 as well as three other, earlier occurring visual attention-related ERP components, previously identified as frontal selection positivity, selection negativity, and N2b (Van der Stelt et al., 1998cGo). In this and the novelty oddball task, the children were instructed to respond as quickly as possible while maintaining a low error rate. They were given small monetary rewards for accurate performance.

Electroencephalographic recording
The EEG was recorded from 29 electrodes located at frontopolar (Fp1, Fpz, and Fp2), anterior frontal (AF7 and AF8), frontal (F7, F3, Fz, F4, and F8), frontocentral (FC5 and FC6), central (T7, C3, Cz, C4, and T8), centroparietal (CP5, CP1, CP2, and CP6), parietal (P7, P3, Pz, P4, and P8), and occipital (O1, Oz, and O2) scalp locations, according to the American Electroencephalographic Society (1991) guidelines for standard electrode position nomenclature. The right mastoid served as reference and AFz as ground. Additionally, electro-oculograms (EOGs) were recorded and utilized for off-line removal of ocular EEG artifacts. EEG and EOG were amplified with a filter bandpass of 0.16–40 Hz and stored on computer disk for off-line processing and analysis.

Data analysis
Behavioural performance measures consisted of the button-press response time to the target stimuli (reaction time), the number of targets missed (misses), and the number of incorrect responses (false alarms). The data were subjected to separate univariate analyses of variance (ANOVAs), including family history and age group as between-subject factors.

The analyses of the ERP data were conducted only on trials with correct behavioural responses. For each child, ERPs were first averaged across trials separately for each type of stimulus at each scalp location. Next, the latency of each ERP component at a selected midline location was extracted by determining the maximal amplitude within a predefined time window. Amplitudes were quantified by computing the mean voltage over the component's latency range, using prestimulus samples as baseline. The data obtained were entered as dependent variables in univariate or multivariate analyses of variance (MANOVAs), including family history and age as between-subject factors, and stimulus type and scalp location as within-subject factors.

To assess differences between COAs and controls in the ERP component amplitudes across specific scalp regions, a series of ‘regional’ MANOVAs were performed on the data from the frontopolar (Fp1, Fpz, and Fp2), frontal (F7, F3, Fz, F4, and F8), central (T7, C3, Cz, C4, and T8), centroparietal (CP5, CP1, CP2, and CP6), parietal (P7, P3, Pz, P4, and P8), occipital (O1, Oz, and O2), and midline (Fpz, Fz, Cz, Pz, and Oz) locations. Significant interactions between family history and location, if present in these analyses on the raw, non-normalized amplitude data, were followed by a profile analysis using normalized data (McCarthy and Wood, 1985Go; Faux and McCarley, 1990Go) to assess differences between COAs and controls in the scalp topographies of the ERP components (for a more detailed description of the topographic analysis procedure, see Van der Stelt et al., 1998a). Normalization served to eliminate absolute amplitude differences between groups, so that in the MANOVA on the normalized amplitude data, only shape or profile differences between scalp distributions were assessed.

Additionally, using the Brain Electric Source Analysis (BESA) software, topographic maps of the scalp potential data (re-referenced to the average reference) and associated scalp current density (SCD) distributions were obtained to supplement the topographic profile analysis. The maps of SCD, being the second spatial derivative of the scalp potential, made possible the identification of scalp areas where local radial current density emerges or converges into the scalp (Hjorth, 1975Go; Nunez, 1981Go). Basically, the application of SCD to the scalp potential data served as a spatial high-pass filter to enhance underlying local brain activity (both in depth and tangentially) while minimizing global and distant contributions (Perrin et al., 1988Go; Nunez and Pilgreen, 1991Go).

Finally, because of pre-existing differences between COAs and controls in externalizing problem behaviour, internalizing problem behaviour, SPM performance, and SES, multiple regression was used to assess and, eventually to control for, the effects of these variables on the behavioural and electrophysiological measures. Also, the impact of the child's sex, as well as its interaction with family history and/or age group, were assessed. Additionally, within-group analyses were performed to assess possible important sources of heterogeneity among COAs with respect to their familial background, involving comparisons between: (1) COAs with a unigenerational family history (i.e. alcoholism in only one or both biological parents) and COAs with a multigenerational family history of alcoholism (i.e. alcoholism in one or both biological parents and in at least one grandparent); (2) COAs with a low-density family history (i.e. alcoholism in only one biological parent) and COAs with a high-density family history (i.e. alcoholism in one biological parent and in at least one other first- or second-degree biological relative); (3) COAs whose parent had alcoholism as the primary disorder, either with or without co-morbid psychopathology, and COAs whose parents had alcoholism secondary to psychopathology; the primary/secondary distinction was made on a temporal basis, in which the disorder with the earlier onset was considered primary; and (4) COAs whose parent had alcoholism alone, without co-morbid psychopathology, and COAs whose parent had alcoholism with co-morbid psychopathology. Finally, to assess potential fetal alcohol effects, comparisons were made between: (1) COAs with a history of maternal alcoholism and COAs with a history of paternal alcoholism, and (2) COAs whose alcoholic mother reported using alcohol during pregnancy and COAs whose alcoholic mother denied drinking during pregnancy.

RESULTS

Behavioural performance
The results on reaction time, misses, and false alarms for the novelty oddball task and the selective attention task are summarized in Table 1Go. There were no significant differences observed between COAs and controls in any of the behavioural measures. Significant behavioural differences were noted as a function of age group, regardless of family history risk status. Overall, the 7- to 12-year-old children displayed a longer reaction time and higher error rates than did the 13- to 18-year-old children in both tasks. Finally, irrespective of family history and age group, the children showed a longer reaction time and made more errors in the selective attention task than in the novelty oddball task.


View this table:
[in this window]
[in a new window]
 
Table 1. Comparison of children of alcoholics (COAs) and controls for the novelty oddball and selective attention tasks
 
P3 amplitude
Overall, COAs were found to produce a smaller P3 amplitude to correctly detected target stimuli over the parietal (P7, P3, Pz, P4, and P8) and occipital (O1, Oz, and O2) scalp locations than controls in both the novelty oddball task and the selective attention task. Although the P3 amplitude differences between COAs and controls tended to be larger for the 7- to 12-year-old age group than for the 13- to 18-year-old group, the observed family history–P3 amplitude relationship did not vary significantly as a function of age group.

These results are illustrated in Figs 1 and 2GoGo. Figure 1Go depicts for each of the family history groups the grand average ERPs from midline parietal (Pz) and occipital (Oz) scalp sites elicited by target stimuli in the novelty oddball task, whereas Fig. 2Go presents the corresponding ERPs recorded in the selective attention task. The group mean values of the P3 amplitude measured at these scalp locations in each task are given in Table 2Go. In the novelty oddball task, the P3 amplitude averaged across age groups was decreased in COAs, relative to controls, at Pz by 17% [22.7 ± 7.7 µV vs 27.3 ± 7.6 µV; F(1, 98) = 8.77, P < 0.004] and at Oz by 30% [14.9 ± 7.3 µV vs 21.4 ± 7.5 µV; F(1, 98) = 19.06, P < 0.001]. In the selective attention task, COAs as compared with controls showed a reduction in the P3 amplitude at Pz by 13% [20.0 ± 7.1 µV vs 23.1 ± 6.9 µV; F(1, 98) = 4.91, P < 0.030] and at Oz by 31% [12.3 ± 7.2 µV vs 17.8 ± 8.2 µV; F(1, 98) = 12.75, P < 0.001]. Thus, the differences between COAs and controls observed in the target P3 amplitude were fairly consistent across experimental paradigms.



View larger version (20K):
[in this window]
[in a new window]
 
Fig. 1. Comparison of grand-average event-related potentials (ERPs) between children of alcoholics (COAs) and controls elicited by target stimuli in the visual novelty oddball task.

Grand-average ERPs for midline parietal (Pz) and occipital (Oz) scalp locations are superimposed for COAs and controls. Left column shows the ERPs for the 7- to 12-year-old age group; right column shows the ERPs for the 13- to 18-year-old age group. Adapted here from Alcohol, 15, Van der Stelt et al., P3 scalp topography to target and novel visual stimuli in children of alcoholics, pp. 119–136, Copyright (1998), with permission from Elsevier Science.

 


View larger version (20K):
[in this window]
[in a new window]
 
Fig. 2. Comparison of grand-average event-related potentials (ERPs) between children of alcoholics (COAs) and controls elicited by target stimuli in the visual selective attention task.

Grand-average ERPs for midline parietal (Pz) and occipital (Oz) scalp locations are superimposed for COAs and controls. Left column shows the ERPs for the 7- to 12-year-old age group; right column shows the ERPs for the 13- to 18-year-old age group. Adapted here from Alcoholism: Clinical and Experimental Research, 22, Van der Stelt et al., event-related potentials during visual selective attention in children of alcoholics, pp. 1877–1889, Copyright (1998), with permission from The Research Society on Alcoholism.

 

View this table:
[in this window]
[in a new window]
 
Table 2. Comparison of children of alcoholics (COAs) and controls for P3 amplitude to target stimuli in the novelty oddball and selective attention tasks
 
In the novelty oddball task, systematic differences between COAs and controls in the P3 amplitude were detected in response both to the task-relevant, target stimuli and to the irrelevant, novel stimuli. However, in contrast to the target stimuli, the group amplitude differences to novel stimuli were found only over the occipital, and not the parietal, scalp locations. This result is illustrated in Fig. 3Go, which shows for each of the family history groups the ERPs from Pz and Oz elicited by the novel stimuli. Table 3Go presents the group mean P3 amplitudes at these scalp locations. Whereas COAs and controls showed no significant overall differences in the P3 amplitude to novel stimuli at Pz [15.7 ± 9.3 µV vs 16.6 ± 7.2 µV; F(1, 98) = 0.33, P > 0.568], the novelty P3 amplitude at Oz was decreased by about 33% in COAs as compared with controls [10.2 ± 6.8 µV vs 15.3 ± 7.0 µV; F(1, 98) = 13.49, P < 0.001].



View larger version (20K):
[in this window]
[in a new window]
 
Fig. 3. Comparison of grand-average event-related potentials (ERPs) between children of alcoholics (COAs) and controls elicited by novel stimuli in the visual novelty oddball task.

Grand-average ERPs for midline parietal (Pz) and occipital (Oz) scalp locations are superimposed for COAs and controls. Left column shows the ERPs for the 7- to 12-year-old age group; right column shows the ERPs for the 13- to 18-year-old age group. Adapted here from Alcohol, 15, Van der Stelt et al., P3 scalp topography to target and novel visual stimuli in children of alcoholics, pp. 119–136, Copyright (1998), with permission from Elsevier Science.

 

View this table:
[in this window]
[in a new window]
 
Table 3. Comparison of children of alcoholics (COAs) and controls for P3 amplitude to novel stimuli
 
In the selective attention task, convincing experimental evidence, not confounded by probability effects or response requirements, was obtained that selective (i.e. attentional) factors determined, at least in part, the P3 amplitude decrements seen in COAs. The evidence was two-fold: (1) the P3 amplitude was found to be highly sensitive to the experimental manipulation of selective attention, that is, to the instructional set or colour-selection cue given to the child, consistent with previous observations in this task (Van der Stelt et al., 1998cGo) and with the P3 literature (see, e.g. Picton and Hillyard, 1988; Polich, 1991; Picton, 1992); (2) the increase in the P3 amplitude induced by the experimental manipulation of attention was found to be significantly smaller in COAs than in controls. For instance, both COAs and controls manifested a larger P3 amplitude at Pz to relevant (attended) non-target stimuli than to irrelevant (unattended) non-target stimuli, but the attention-related increase in the P3 amplitude at this scalp location was markedly smaller in COAs than in controls [7.0 ± 6.0 µV vs 9.7 ± 4.0 µV; F(1, 98) = 7.35, P < 0.009].

P3 scalp topography
In both tasks, the COAs and controls showed marked differences in the P3 raw amplitudes to target stimuli over the parietal and occipital scalp, whereas no significant group amplitude differences were seen over the central, frontal, and frontopolar, scalp locations. This finding raised the possibility that significant differences existed between COAs and controls in the P3 scalp topography. However, a profile analysis using normalized amplitude data (to eliminate group differences in the raw amplitudes) revealed no significant differences between COAs and controls in the topographic profile of the P3 amplitudes across the midline or coronal scalp locations. This implied that the observed P3 amplitude decrements in COAs originated from a difference in the strength, rather than in the spatial configuration, of the underlying brain generators. By comparison, reliable differences in the P3 scalp topography across the midline locations were found in both tasks as a function of age group. Accordingly, in the younger age group, both COAs and controls manifested a similar Pz > Oz > Cz > Fz > Fpz amplitude distribution of the target P3, whereas in the older age group, both family history groups showed a comparable Pz > Cz > Oz > Fz > Fpz amplitude distribution.

The results of the topographic profile analysis were supported by the topographic maps of the P3 voltage and corresponding SCD distributions, as shown in Fig. 4Go. The scalp potential maps (upper row) indicated in each age group a basic similarity in the target P3 scalp topography for COAs and controls. However, the scalp potential data alone did not allow inferences about the nature of the putative underlying neural generators, because a given scalp potential field can be generated by any of an infinite number of different current sources in the brain (Nunez, 1981Go). For that reason, a mathematical transformation known as the surface Laplacian or current source density (SCD) was applied to the scalp potential data, which may allow an interpretation in terms of the underlying brain generators (Nunez, 1981Go; Nunez and Pilgreen, 1991Go). The obtained SCD maps (lower row) indicated that a similar cortical activation pattern, involving both the anterior and the posterior association cortex, contributed to the scalp-recorded visual P3 in COAs and controls. Accordingly, despite differences in the amplitude or intensity of the visual P3 generation, the spatial configuration of the neural generators underlying this ERP component seemed to be identical in COAs and low-risk controls.



View larger version (62K):
[in this window]
[in a new window]
 
Fig. 4. Scalp potential and current density maps for the grand-average P3 elicited by target stimuli for children of alcoholics (COAs) and controls.

Scalp potential (SP) and scalp current density (SCD) maps for the grand-average P3 elicited by target stimuli are given for each of the family history groups. The maps for the 7- to 12-year-old age group were computed at 474 ms post-stimulus; for the 13- to 18-year-old age groups, they were computed at 404 ms. Contours represent voltage increments of 1 µV for the potential maps and arbitrary but equal current units (µV/cm2) for the current density maps. Unshaded areas indicate positive voltages or outward current flow; shaded areas indicate negative voltages or inward current flow. The data were obtained in the visual novelty oddball task. Reprinted here from Alcohol, 15, Van der Stelt et al., P3 scalp topography to target and novel visual stimuli in children of alcoholics, pp. 119–136, Copyright (1998), with permission from Elsevier Science.

 
The presence of distinct multiple scalp regions of radial current flow visible in the SCD maps provided tentative evidence that a distributed, rather than a localized, cortical source distribution mediated the scalp-recorded visual P3 in COAs and controls. This notion was further substantiated by the results of a specific analysis of the overall P3 amplitude to attended target stimuli in the selective attention task. The results of that analysis revealed, not only that the portions of the P3 amplitude specifically related to (i.e. elicited by) the attention and target manipulations differed in scalp topography, indicating that each originated from a distinct configuration of brain generators (Johnson, 1993Go), but also that these specific portions of P3 amplitude were both significantly smaller in COAs than in controls. These results also indicate that multiple brain generators, rather than a single generator, determined the P3 amplitude decrements seen in COAs.

P3 peak latency
The measured P3 peak latencies from the Pz electrode in response to target stimuli as a function of family history, age group, and task are presented in Table 4Go. The P3 peak latency did not vary significantly as a function of family history in either of the tasks. Independently of family history, systematic differences were noted as a function of age group. Overall, the 7- to 12-year-old children manifested in both tasks a longer P3 peak latency to target stimuli than did the 13- to 18-year-old children. Also, irrespective of family history and age group, the target P3 peak latency was longer in the selective attention task than in the novelty oddball task. In accordance with the behavioural data, the latter result signified that the detection of the target stimuli was more difficult and, consequently, took longer in the selective attention task than in the novelty oddball task.


View this table:
[in this window]
[in a new window]
 
Table 4. Comparison of children of alcoholics (COAs) and controls for the P3 peak latency to target stimuli in the novelty oddball and selective attention tasks
 
Frontal selection positivity, selection negativity, and N2b
In the selective attention task, the COAs and controls showed significant differences in the P3 amplitude to attended target stimuli. By contrast, no significant group differences were seen in this task in the latencies, amplitudes, and scalp topographies of three other, earlier occurring visual attention-related ERP components, previously identified as frontal selection positivity, selection negativity, and N2b (Van der Stelt et al., 1998cGo).

Additional analyses
Extra analyses revealed that the P3 amplitude correlated significantly with externalizing problem scores, SPM performance, internalizing problem scores, and SES, although the observed correlations were not invariant across scalp locations and task paradigms. Because COAs differed from controls on each of these four variables, multiple regression analyses were performed to determine whether family history risk status had incremental predictive validity, that is, whether it added significantly to predicting the P3 amplitude over and above what could be predicted from knowing scores on the other variables. The results indicated that considerable redundancy existed in the relationship to the P3 amplitude among family history and the four other predictors, perhaps indicating that these variables all share to some extent a common conceptual meaning (Meehl, 1970Go). Nonetheless, it was also found that family history risk status made a significantly non-zero unique contribution to the prediction of the P3 amplitude, particularly over the occipital scalp, where the differences between COAs and controls were largest. Thus, initial differences between COAs and controls in externalizing and internalizing problems, SPM performance, and SES could not fully account for the observed family history–P3 amplitude relationship. Though not mediating the relationship, these pre-existing differences between the groups of children did act to moderate, to enhance the strength of the observed family history–P3 amplitude relationship.

Sex of the children was found to have no significant effect on the P3 amplitude. Nor were there significant interaction effects observed among sex, family history, and/or age group on the P3 amplitude. Thus, both sons and daughters of alcoholics were characterized by a reduction in the P3 amplitude, as compared with controls.

Additionally, the P3 amplitude did not vary significantly as a function of the sex of the alcoholic parent. A reduction in the P3 amplitude was found to characterize the children of both male and female alcoholics. Although high-risk research with children of female alcoholics is complicated due to potential adverse effects of maternal alcohol consumption during pregnancy on the developing fetus, the children with an alcoholic mother could not be distinguished from the children with an alcoholic father, nor from the control children, on the basis of the P3 peak latency, amplitudes and latencies of the other ERP components, and behavioural performance. Thus, the observed P3 amplitude decrements seen in the children of the female alcoholics are unlikely to be confounded by fetal alcohol effects.

Finally, possible heterogeneity within the high-risk group was assessed. Table 5Go illustrates the results of these within-group comparisons for the P3 amplitude data obtained in the selective attention task. No significant P3 differences among COAs were seen as a function of a unigenerational vs multigenerational family history of alcoholism, low-density vs high-density family history of alcoholism, or the presence vs absence of co-morbid psychopathology in the alcoholic parent. It was noted that COAs, whose parent had alcoholism as the primary disorder, manifested a smaller P3 amplitude than COAs whose parent had alcoholism secondary to psychopathology, but these significant subgroup P3 amplitude differences were present only in the selective attention task and only over the occipital scalp.


View this table:
[in this window]
[in a new window]
 
Table 5. P3 amplitude to target stimuli in the selective attention task as a function of parental family history and associated psychopathology
 

DISCUSSION

The present results indicate that differences between COAs and low-risk controls in the visual P3 amplitude: (1) can be elicited both actively by task-relevant target stimuli and passively by irrelevant novel stimuli; (2) are a function of both the attentional relevance and the target properties of the eliciting stimulus; (3) are mediated by multiple brain generators, rather than by a single generator; (4) originate from a difference in the strength, rather than in the spatial configuration, of the underlying brain generators; (5) cannot be accounted for by differences in visual attention-related earlier occurring ERP components; and (6) can be moderated by current behavioural and emotional problems, general intellectual ability, and socio-economic background. These findings corroborate results from previous high-risk studies (Elmasian et al., 1982Go; Begleiter et al., 1984Go; O'Connor et al., 1987Go; Whipple et al., 1988Go, 1991Go; Porjesz and Begleiter, 1990Go; Hill and Steinhauer, 1993Go; Steinhauer and Hill, 1993Go; Polich et al., 1994Go; Benegal et al., 1995Go; Hill et al., 1995aGo; Porjesz et al., 1996Go; Ramachandran et al., 1996Go; Cohen et al., 1997Go; Sharma et al., 1997Go), and support the notion that a small-amplitude visual P3 in COAs reflects heritable biases in attention and information processing that are related to their increased vulnerability to alcoholism.

In general, two types of genes have been implicated as potential candidates for the major loci conferring risk for alcoholism, namely neurochemical and alcohol-metabolizing genes (Mullan, 1989Go; Cloninger and Begleiter, 1990Go). The results from studies of the P3 in alcoholics and their relatives favour the notion that genetically determined variation in neurochemical mechanisms mediates a person's vulnerability to alcoholism. Although more specific, pharmacogenetic factors (e.g. decreased intensity of response to alcohol, see Schuckit, 1994) could also be involved, the P3 results indicate that alcoholism vulnerability is mediated by individual differences in neurochemical mechanisms that can be detected at baseline, that is, without the administration of alcohol. Moreover, these data indicate that such individual differences exist prior to the first exposure to alcohol. Accordingly, the P3 results support ‘adaptive’, rather than ‘exposure’, theories of alcohol and drug abuse. Whereas exposure theories emphasize the role of drugs as rewarding stimuli (or positive reinforcers) in risk for abuse, adaptive theories view substance use as an adaptive mechanism (or negative reinforcer) for coping with some pre-existing deficiency or suboptimality (Alexander and Hadaway, 1982Go; Bardo et al., 1996Go). Thus, a relatively small P3 amplitude often seen in individuals at high risk may index some dysfunctional brain and cognitive state in which addictive and perhaps other risk-taking behaviours are easily elicited as an attempt to counteract this deficiency.

The notion that genetic differences among individuals in higher brain and cognitive function, as indexed by the P3 amplitude, determine in part the risk for substance abuse seems to be further supported by the results from a recent study from our laboratory (Ratsma et al., in preparation). In this study, a highly significant linear relationship was observed in a sample of non-alcoholic young adults with a family history of alcoholism between the visual P3 amplitude and the personality trait Sensation-seeking, particularly the subtrait Disinhibition: the lower the visual P3 amplitude, the more the person was inclined to seek sensation and to find release through social disinhibition. Given the association between sensation-seeking characteristics and substance abuse risk (Cloninger, 1987Go; Cloninger et al., 1988Go; Zuckerman, 1994Go), these data imply, indeed, that those individuals who showed the smallest P3 amplitude are the ones most vulnerable for developing a substance use disorder. These data also suggest that a relatively small P3 amplitude may not be seen as a real aberration, but rather as the extreme of a continuum that places an individual, at least with respect to this characteristic, at the higher end of a continuum of addiction risk.

Additionally, adaptive theories of addiction seem to be substantiated by the present data. First, significant linear relationships were noted between the P3 amplitude and externalizing and internalizing problem scores. Thus, the lower the visual P3 amplitude, the more likely the child was to manifest pre-existing behavioural and emotional problems. If it is true that the degree to which an individual manifests some pre-existing psychopathology determines in part the risk for substance abuse, this observation implies, again, that those children who showed the smallest P3 amplitude are the ones most at risk for addiction. Second, the differences between COAs and controls in the target P3 amplitude were found over both the parietal and occipital scalp, but the group differences in the novelty P3 amplitude were found only over the occipital, and not the parietal, scalp locations. This result suggests that the presentation of novel stimuli, as opposed to target stimuli, had some beneficial, partly normalizing, effect on the responses of COAs, so that they no longer differed from controls in the P3 amplitude at the parietal scalp locations. This partly normalizing effect of stimulus novelty on the P3 amplitude appears to support the idea that prevention efforts might benefit by providing individuals at high risk with novel experiences that potentially compete with alcohol and drug-taking behaviour (e.g. Bardo et al., 1996).

The P3 seems to meet each of the four criteria for the identification of a vulnerability marker of alcoholism (for reviews, see Porjesz et al., 1996; Van der Stelt, 1998), as discussed in the Introduction. However, before this conclusion can be fully accepted, more information is needed on the heritability or molecular genetic basis of the P3, on its predictive validity with respect to alcoholism, and on its relationship with alcoholism within pedigrees. Also, in order to assess its aetiological significance, researchers should attempt to modify the incidence of alcoholism by altering the P3 amplitude. Specifically, if the P3 indeed indicates a causative pathophysiologic factor in alcoholism, then psychological or pharmacological interventions that succeed in increasing the P3 amplitude in individuals at high risk should to some extent protect against the onset of alcoholism. In contrast, if the P3 represents merely a correlate or a byproduct of the underlying pathophysiological mechanism, then intervention should not make a difference in outcome.

Alternatively, researchers interested in the aetiological significance of the P3 could attempt to elucidate a pathophysiological link between this ERP component and alcoholism risk. Recently, an association has been reported in children at high risk for alcoholism between the dopamine D2 receptor locus (DRD2) and the P3 peak latency (Noble et al., 1994Go) and amplitude (Hill et al., 1998Go), which perhaps indicates that the link between low P3 amplitude and increased alcoholism risk is mediated by alterations in brain dopaminergic function. However, the significance of these findings awaits elucidation, because our recent study (Ratsma et al., in preparation) could not detect a significant DRD2–P3 association in a sample of non-alcoholic young adults with a family history of alcoholism.

Finally, it is unlikely that the P3 reflects a vulnerability that is specific to alcoholism, because a reduction in its amplitude has been found to characterize a number of different substance use and neuropsychiatric disorders (Friedman, 1990Go; Picton, 1992Go; Van der Stelt, 1998Go). Accordingly, the P3 should be viewed as a putative pathophysiological marker, rather than a diagnostic marker of alcoholism (Nurnberger, 1992Go). That is, the P3 amplitude potentially indicates a pathophysiological factor or biological vulnerability that may be involved, to a greater or lesser extent, in multiple psychopathological conditions. Conceivably, the actual behavioural outcome of the vulnerability in any particular individual may depend on other genetic and environmental factors specific to the individual.

REFERENCES

Achenbach, T. M. (1991) Integrative Guide for the 1991 CBCL/4-18, YSR, and TRF Profiles. University of Vermont, Department of Psychiatry, Burlington, VT.

Alexander, B. K. and Hadaway, P. F. (1982) Opiate addiction: the case for an adaptive orientation. Psychological Bulletin 92, 367–381.[ISI][Medline]

American Electroencephalographic Society (1991) Guidelines for standard electrode position nomenclature. Journal of Clinical Neurophysiology 8, 200–202.[ISI][Medline]

American Psychiatric Association (1987) Diagnostic and Statistical Manual of Mental Disorders, 3rd edn, revised. American Psychiatric Association, Washington, DC.

Aston, C. E. and Hill, S. Y. (1990) A segregation analysis of the P300 component of the event-related brain potential. American Journal of Human Genetics 47 (Suppl.), A127.

Bardo, M. T., Donohew, R. L. and Harrington, N. G. (1996) Psychobiology of novelty seeking and drug seeking behaviour. Behavioral Brain Research 77, 23–43.[ISI][Medline]

Begleiter, H. and Porjesz, B. (1988) Potential biological markers in individuals at high risk for developing alcoholism. Alcoholism: Clinical and Experimental Research 12, 488–493.[ISI][Medline]

Begleiter, H., Porjesz, B., Bihari, B. and Kissin, B. (1984) Event-related brain potentials in boys at risk for alcoholism. Science 225, 1493–1496.[ISI][Medline]

Benegal, V., Jain, S., Subbukrishna, D. K. and Channabasavanna, S. M. (1995) P300 amplitudes vary inversely with continuum of risk in first degree male relatives of alcoholics. Psychiatric Genetics 5, 149–156.[ISI][Medline]

Berman, S. M., Whipple, S. C., Fitch, R. J. and Noble, E. P. (1993) P3 in young boys as a predictor of adolescent substance use. Alcohol 10, 69–76.[ISI][Medline]

Clarren, S. K., Sampson, P. D., Larsen, J., Donnell, D. J., Barr, H. M., Bookstein, F. L., Martin, D. C. and Streissguth, A. P. (1987) Facial effects of fetal alcohol exposure: assessment by photographs and morphometric analysis. American Journal of Medical Genetics 26, 651–666.[ISI][Medline]

Cloninger, C. R. (1987) Neurogenetic adaptive mechanisms in alcoholism. Science 236, 410–416.[ISI][Medline]

Cloninger, C. R. and Begleiter, H. eds (1990) Banbury Report 33: Genetics and Biology of Alcoholism. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

Cloninger, C. R., Sigvaardsson, S. and Bohman, M. (1988) Childhood personality predicts alcohol abuse in young adults. Alcoholism: Clinical and Experimental Research 12, 494–505.[ISI][Medline]

Cohen, H. L., Wang, W., Porjesz, B. and Begleiter H. (1995) Auditory P300 in young alcoholics: regional response characteristics. Alcoholism: Clinical and Experimental Research 19469–475.

Cohen, H. L., Porjesz, B., Begleiter, H. and Wang, W. (1997) Neuroelectric correlates of response production and inibition in individuals at risk to develop alcoholism. Biological Psychiatry 42, 57–67.[ISI][Medline]

Couzigou, P., Begleiter, H., Kiianmaa, K. and Agarwal, D. P. (1993) Genetics and alcohol. In Health Issues Related to Alcohol Consumption, Verschuren, P. M. ed., pp. 282–329. International Life Sciences Institute Press, Washington, DC.

Devor, E. J. and Cloninger, C. R. (1989) Genetics of alcoholism. Annual Review of Genetics 23, 19–36.[ISI][Medline]

Donchin, E., Ritter, W. and McCallum, W. C. (1978) Cognitive psychophysiology: the endogenous components of the ERP. In Event-related Brain Potentials in Man, Callaway, E., Tueting, T. and Koslow, S. H. eds, pp. 349–411. Academic Press, New York.

Elmasian, R., Neville, H., Woods, D., Schuckit, M. and Bloom, F. (1982) Event-related brain potentials are different in individuals at high and low risk for developing alcoholism. Proceedings of the National Academy of Sciences of the USA 79, 7900–7903.[Abstract]

Faux, S. F. and McCarley, R. W. (1990) Analysis of scalp voltage asymmetries using Hotelling's T2 methodology. Brain Topography 2, 237–245.[Medline]

Friedman, D. (1990) Event-related potentials in populations at genetic risk: a methodological review. In Event-related Brain Potentials: Basic Issues and Applications, Rohrbaugh, J. W., Parasuraman, R. and Johnson, R., Jr. eds, pp. 310–332. Oxford University Press, New York.

Gershon, E. S. and Goldin, L. R. (1986) Clinical methods in psychiatric genetics. I. Robustness of genetic marker investigative strategies. Acta Psychiatrica Scandinavica 74, 113–118.[ISI][Medline]

Goldin, L. R., Nurnberger, J. I. and Gershon, E. S. (1986) Clinical methods in psychiatric genetics. II. The high risk approach. Acta Psychiatrica Scandinavica 74, 119–128.[Medline]

Gunning, W. B., Pattiselanno, S. E., Van der Stelt, O. and Wiers, R. W. (1994) Children of alcoholics. Predictors for psychopathology and addiction. Acta Paediatrica Suppl. 404, 7–8.[Medline]

Gustafsson, J-E. (1994) General intelligence. In Encyclopedia of Human Intelligence, Sternberg, R. J. ed., pp. 469–475. Macmillan, New York.

Hill, S. Y. and Steinhauer, S. R. (1993) Assessment of prepubertal and postpubertal boys and girls at risk for developing alcoholism with P300 from a visual discrimination task. Journal of Studies on Alcohol 54, 350–358.[ISI][Medline]

Hill, S. Y., Muka, D., Steinhauer, S. and Locke, J. (1995a) P300 amplitude decrements in children from families of alcoholic female probands. Biological Psychiatry 38, 622–632.[ISI][Medline]

Hill, S. Y., Steinhauer, S., Lowers, L. and Locke, J. (1995b) Eight-year longitudinal follow-up of P300 and clinical outcome in children from high-risk for alcoholism families. Biological Psychiatry 37, 823–827.[ISI][Medline]

Hill, S. Y., Locke, J., Zezza, N., Kaplan, B., Neiswanger, K., Steinhauer, S., Wipprecht, G. and Zu, J. (1998) Genetic association between reduced P300 amplitude and the DRD2 dopamine receptor A1 allele in children at high risk for alcoholism. Biological Psychiatry 43, 40–51.[ISI][Medline]

Hjorth, B. (1975) An on-line transformation of EEG scalp potentials into orthogonal source derivations. Electroencephalography and Clinical Neurophysiology 39, 526–530.[ISI][Medline]

Johnson, R., Jr (1993) On the neural generators of the P300 component of the event-related potential. Psychophysiology 30, 90–97.[ISI][Medline]

Katsanis, J., Iacono, W. G., McGue, M. K. and Carlson, S. R. (1997) P300 event-related potential heritability in monozygotic and dizygotic twins. Psychophysiology 34, 47–58.[ISI][Medline]

McCarthy, G. and Wood, C. C. (1985) Scalp distributions of event-related potentials: an ambiguity associated with analysis of variance models. Electroencephalography and Clinical Neurophysiology 62, 203–208.[ISI][Medline]

Meehl, P. E. (1970) Nuisance variables and the ex post facto design. In Minnesota Studies in the Philosophy of Science, Vol. IV, Analyses of Theories and Methods of Physics and Psychology, Radner, M. and Winokur, S. eds, pp. 373–402. University of Minneapolis Press, Minneapolis, MN.

Mullan, M. (1989) Alcoholism and the ‘New Genetics’. British Journal of Addiction 84, 1433–1440.[ISI][Medline]

Noble, E. P., Berman, S. M., Ozkaragoz, T. Z. and Ritchie, T. (1994) Prolonged P300 latency in children with the D2 dopamine receptor A1 allele. American Journal of Human Genetics 54, 658–668.[ISI][Medline]

Nunez, P. L. (1981) Electrical Fields of the Brain: The Neurophysics of EEG. Oxford University Press, New York.

Nunez, P. L. and Pilgreen, K. L. (1991) The spline-Laplacian in clinical neurophysiology: a method to improve EEG spatial resolution. Journal of Clinical Neurophysiology 8, 397–413.[ISI][Medline]

Nurnberger, J. L. (1992) Should a biological marker be sensitive and specific? Acta Psychiatrica Scandinavica 86, 1–4.[ISI][Medline]

O'Connor, S., Hesselbrock, V., Tasman, A. and De Palma, N. (1987) P3 amplitudes in two distinct tasks are decreased in young men with a history of paternal alcoholism. Alcohol 4, 323–330.[ISI][Medline]

O'Connor, S., Morzorati, S., Christian, J. C. and Li, T.-K. (1994) Heritable features of the auditory oddball event-related potential: peaks, latencies, morphology and topography. Electroencephalography and Clinical Neurophysiology 92, 115–125.[ISI][Medline]

Perrin, F., Pernier, J. and Bertrand, O. (1988) Scalp current density fields: concept and properties. Electroencephalography and Clinical Neurophysiology 69, 385–389.[ISI][Medline]

Pfefferbaum, A., Ford, J. M., White P. M. and Mathalon, D. (1991) Event-related potentials in alcoholic men: P3 amplitude reflects family history but not alcohol consumption. Alcoholism: Clinical and Experimental Research 15839–850.[ISI][Medline]

Picton, T. W. (1992) The P300 wave of the human event-related potential. Journal of Clinical Neurophysiology 9, 456–479.[ISI][Medline]

Picton, T. W. and Hillyard, S. A. (1988) Endogenous event-related potentials. In Handbook of Electroencephalography and Clinical Neurophysiology, Vol. 3, Human Event-related Potentials, Picton, T. W. ed., pp. 361–426. Elsevier, Amsterdam.

Pihl, R. O., Peterson, P. and Finn, P. (1990) Inherited predisposition to alcoholism: characteristics of sons of male alcoholics. Journal of Abnormal Psychology 99, 291–301.[ISI][Medline]

Polich, J. (1991) P300 in clinical applications: meaning, method, and measurement. American Journal of EEG Technology 31, 201–231.

Polich, J., Pollock, V. E. and Bloom, F. E. (1994) Meta-analysis of P300 amplitude from males at risk from alcoholism. Psychological Bulletin 115, 55–73.[ISI][Medline]

Porjesz, B. and Begleiter, H. (1983) Brain dysfunction and alcohol. In The Biology of Alcoholism, Vol. 7, The Pathogenesis of Alcoholism: Biological Factors, Kissin, B. and Begleiter, H. eds, pp. 415–483. Plenum Press, New York.

Porjesz, B. and Begleiter, H. (1990) Event-related potentials in individuals at risk for alcoholism. Alcohol 7, 465–469.[ISI][Medline]

Porjesz, B. and Begleiter, H. (1995) Event-related potentials and cognitive function in alcoholism. Alcohol Health and Research World 19, 108–112.

Porjesz, B., Begleiter, H., Litke, A., Bauer, L. O., Kuperman, S., O'Connor, S. J. and Rohrbaugh, J. (1996) Visual P3 as a potential phenotypic marker for alcoholism: evidence from the COGA national project. In Recent Advances in Event-related Brain Potential Research, Ogura, C., Koga, Y. and Shimokochi, M. eds, pp. 539–549. Elsevier, Amsterdam.

Ramachandran, G., Porjesz, B. and Begleiter, H. (1996) A simple auditory oddball task in young adult males at high risk for alcoholism. Alcoholism: Clinical and Experimental Research 20, 9–15.[ISI][Medline]

Raven, J. C., Court, J. H. and Raven, J. (1992) Manual for Raven's Progressive Matrices and Vocabulary Scales, Section 3, Standard Progressive Matrices. Oxford Psychologists Press, Oxford.

Raven, J., Raven, J. C. and Court, J. H. (1993) Manual for Raven's Progressive Matrices and Vocabulary Scales, Section 1, General Overview. Oxford Psychologists Press, Oxford.

Schuckit, M. A. (1989) Biomedical and genetic markers of alcoholism. In Alcoholism: Biomedical and Genetic Aspects, Goedde, H. W. and Agarwal, D. P. eds, pp. 290–302. Pergamon Press, New York.

Schuckit, M. A. (1994) A clinical model of genetic influences in alcohol dependence. Journal of Studies on Alcohol 55, 5–17.[ISI][Medline]

Sharma, A., Malhotra, S., Raghunathan, M. and Malhotra, A. (1997) A study of event-related potential-P3 characteristics in children of alcoholics. Addiction Biology 2, 431–438.[ISI]

Sher, K. J. (1991) Children of Alcoholics: A Critical Appraisal of Theory and Research. University of Chicago Press, Chicago, IL.

Steinhauer, S. R. and Hill, S. Y. (1993) Auditory event-related potentials in children at high risk for alcoholism. Journal of Studies on Alcohol 54, 408–421.[ISI][Medline]

Van der Stelt, O. (1997) Children of Alcoholics: Attention, Information Processing, and Event-related Brain Potentials. Doctoral Dissertation, University of Amsterdam, Amsterdam.

Van der Stelt, O. (1998) The P3 component of the human event-related potential as a vulnerability marker of alcoholism. Alcohol Research 3, 6–9.

Van der Stelt, O., Gunning, W. B., Snel, J., Zeef, E. and Kok, A. (1994) Children of alcoholics: attention, information processing and event-related brain potentials. Acta Paediatrica (Suppl. 404), 4–6.

Van der Stelt, O., Gunning, W. B., Snel, J. and Kok, A. (1997) No electrocortical evidence of automatic mismatch dysfunction in children of alcoholics. Alcoholism: Clinical and Experimental Research 21, 569–575.[ISI][Medline]

Van der Stelt, O., Geesken, R., Gunning, W. B., Snel, J. and Kok, A. (1998a) P3 scalp topography to target and novel visual stimuli in children of alcoholics. Alcohol 15, 119–136.[ISI][Medline]

Van der Stelt, O., Gunning, W. B., Snel, J. and Kok, A. (1998b) Event-related potentials during visual selective attention in children of alcoholics. Alcoholism: Clinical and Experimental Research 22, 1877–1889.[ISI][Medline]

Van der Stelt, O., Kok, A., Smulders, F. T. Y., Snel, J. and Gunning, W. B. (1998c) Cerebral event-related potentials associated with selective attention to color: developmental changes from childhood to adulthood. Psychophysiology 35, 227–239.[ISI][Medline]

Whipple, S. C., Parker, E. S. and Noble, E. P. (1988) An atypical neurocognitive profile in alcoholic fathers and their sons. Journal of Studies on Alcohol 49, 240–244.[ISI][Medline]

Whipple, S. C., Berman, S. M. and Noble, E. P. (1991) Event-related potentials in alcoholic fathers and their sons. Alcohol 8, 321–327.[ISI][Medline]

Zuckerman, M. (1994) Behavioral Expressions and Biosocial Bases of Sensation Seeking. Cambridge University Press, New York.