Event-related potential negative shift in sons of polysubstance- and alcohol-use disorder fathers

PSYCHIATRY RESEARCH ELSEVIER

Psychiatry Research 73 (1997) 133-146

Event-related potential negative shift in sons of polysubstance- and alcohol-use disorder fathers Janet Brigham*, Howard B. Moss’, E. Lenn Murrelle2, Levent Kirisci, James S. Spinelli3 Center for Education and Drug Abuse Research, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA

Received 1 August 1996;revised 20 July 1997;accepted 19 September 1997

Abstract

Previous research has considered event-related potentials (ERPs) in relation to liability for alcohol and other substance use. This study explored ERPs in preadolescent boys at elevated risk for substance use due to paternal history of substance abuse or dependence. Sons (age 10-12) of fathers with an alcohol-use disorder (ALC, n = 29) were matched by age, IQ, education and parental alcohol use with sons of fathers with a polysubstance abuse or dependence diagnosis (POLY, n = 37). These two groups were matched with a low-risk comparison group (LOW, n = 29) of boys whose fathers had no substance-use disorder diagnosis. No boy in the study met criteria for a substance-use disorder. ERPs were collected from midline (Fz, Cz, Pz) and parietal (P3, P4) electrode leads during an auditory oddball task. ERPs of boys from the ALC and POLY groups showed a slow negative shift prominent at Cz and Pz. This negative shift, evident by 100 ms post-stimulus and lasting for the duration of the lOOO-ms recording period, overlapped temporally with Nl, N2 and P3 amplitude differences distinguishing the ALC and POLY groups from the LOW group. The ALC and POLY groups differed from each other in N2 amplitude at Cz, which was larger for ALC subjects. These findings offer a possible alternative explanation for previously observed amplitude anomalies noted in children at risk for substance-use disorders and suggest new avenues of inquiry. 0 1997 Elsevier Science Ireland Ltd. Keywords:

Risk factors; Alcoholism; Auditory event-related

potentials; Substance use; Biological vulnerability

*Corresponding author, Health Sciences and Policy Program,

SRI International, 333 Ravenswood Avenue, Menlo Park, CA-94025, USA, e-mail: [email protected] ‘Present address: Western Psychiatric Institute and Clinic, University of Pittsburgh. *Present address: Virginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, Richmond, VA, USA. 3Present address: CDS Technologies, Pittsburgh, PA, USA.

01651781/97/$17.00 0 1997 Elsevier PII SO165-1781(97)00121-2

Science

Ireland

Ltd. All rights reserved.

134

J. Brigham et al. /Psychiatry Research 73 (1997) 133-146

children at elevated risk for developing a substance-use disorder reflect possible cognitive dysfunctions? Although that question cannot be answered directly within the scope of this article, the known relationship between electrophysiological phenomena and information processing highlights the need to evaluate findings in regard to possible cognitive dysfunction.

1. Introduction

The possibility that an event-related potential (ERP) component or other electrophysiological measure could serve as a biological indication of susceptibility to substance-use disorders has been explored in numerous investigations. Most such research has focused on the offspring of alcoholic fathers. Toward this end, the ERP component identified as P3 or P300 (third positive peak, occurring approximately 300 ms post-stimulus) has been examined in dozens of studies, frequently showing a lower amplitude in sons of alcoholics than in sons of control fathers (e.g., Elmasian et al., 1982; Begleiter et al., 1984). Polich et al. (19941, in a meta-analysis and review, concluded that although this component has not been demonstrated to be a specific biological marker, it may provide an index of alcoholism vulnerability. Several research groups (Herning et al., 1989; Berman et al., 1993; Brigham et al., 199.5) have extended the alcohol-research paradigm to substance use, in an effort to identify elements distinguishing risk for alcoholism from risk for substance use. Berman et al. (1993) found that although P3 amplitude significantly predicted adolescent substance use, the effect size was small. They concluded that weak utility of P3 as an indication of vulnerability required that it be used in conjunction with other measures. The largest prior ERP study of children at risk for substance use, Brigham et al. (19951, included evidence of a diminished P3 amplitude in sons of fathers with substance-use disorders. Several questions have remained unexamined or only partially examined regarding ERP research in at-risk populations: .

.

.

Do offspring of alcohol-use-disorder fathers differ from offspring of fathers diagnosed with other substance-use disorders (the latter fathers commonly being comorbid for both alcohol- and other substance-use disorder)? Additionally, how do boys from families affected by alcohol and other substance-use disorders compare with offspring of parents with no substance-use disorder? Do electrophysiological variations detected in

The two and the

goals of this study were to address the first issues within the limits of this ERP paradigm, to suggest possible approaches for examining third issue effectively.

2. Methods 2.1. Subject ascertainment and diagnosis

The present study examined ERP components in three groups of lo- to 12-year-old boys whose biological fathers met diagnostic criteria for one of the following categories: 1. A history of alcohol abuse or dependence, but no other substance-use disorder diagnosis except nicotine (AL0 2. A history of alcohol-related substance-use disorder and at least one other substance-use abuse or dependence diagnosis, except nicotine, thus indicating polysubstance use (POLY). 3. No parental history of other substance-use disorder diagnosis, thus resulting in the sons’ lower risk for substance-use disorder (LOW). Ninety-five male subjects, aged lo-12 years, participated in the study. They were ascertained based on their traits (e.g., no substance-use disorder diagnosis, demographic characteristics) and on their parents’ substance-use histories (Table 1). Subjects were selected for three groups based on parental substance-use disorder diagnosis. The groups were matched by group means and standard deviations on criteria of age, intelligence test scores, and years of education (see diagnostic procedure, below). Subjects were drawn, based on

135

J. Brigham et al. /Psychiatry Research 73 (1997) 133-146

Table 1 Subject characteristics Parental substance use

Characteristic

Alcohol-only abuse or dependence (n = 29) SD.

Total (n = 95) Polysubstance abuse or dependence (n = 37) Mean

SD.

No abuse or dependence (n = 29) Mean

Mean

S.D.

S.D.

SES

,

43.00

12.66

39.95

10.72

44.97

12.94

42.41

12.04

Age

Father Mother Son

43.17 40.61 11.76

6.20 4.18 1.06

39.35 35.95 11.57

6.20 4.59 1.01

39.07 37.71 11.46

4.96 4.63 0.84

40.43 37.88 11.59

6.04 4.82 0.97

Education

Son

4.62

1.01

4.43

0.96

4.52

1.15

4.52

1.03

Verbal IQ

Father Mother Son

104.72 98.82 105.24

13.61 15.3 15.53

104.35 96.89 100.22

18.68 14.94 11.30

115.79 104.11 101.34

21.2 16.13 11.30

107.96 99.63 102.09

18.16 15.47 12.73

Performance IQ

Father Mother Son

107.55 97.86 105.24

16.48 15.94 15.53

102.51 96.79 100.22

17.26 15.42 11.3

110.62 100.41 101.34

18.50 17.03 11.3

106.53 98.19 102.09

17.50 16.14 12.73

Full-Scale IQ

Father Mother Son

106.38 98.61 107.86

13.01 14.71 16.23

103.54 96.56 102.65

16.36 14.88 9.71

115.45 102.22 103.55

19.02 14.78 12.74

108.04 98.87 104.52

16.85 14.76 12.91

Count

Percent

Count

Percent

Count

Percent

Count

Percent

Ethnic&y

Handedness

Caucasian AfricanAmerican Other

25 4

86 14

26 8

70 11

23 6

79 21

74 18

78 19

0

0

3

6

0

0

3

3

Left Right

6 23

21 79

5 31

14 84

4 25

14 86

15 79

16 83

-

those matching criteria, from a subject pool of some 600 boys tested as part of a longitudinal study of biological and social risk for substanceuse at the Center for Education and Drug Abuse Research in Pittsburgh, PA. The groups of boys whose parents had alcohol-use diagnoses were matched to sons of fathers with polysubstance-use diagnoses by range and group means of the above-specified matching criteria and by paternal and maternal alcohol-use disorder diagnoses. These two groups were then matched by age, IQ and education to a group of boys whose parents did not meet criteria for any substance-use diagnosis.

Diagnostic interviews, administered by trained graduate-level clinicians, were discussed in a consensus conference with two clinical psychologists or psychiatrists, the interviewers and the evaluation coordinator. The consensus team reviewed all available information gathered in the assessment protocol, including psychiatric treatment records and teacher reports, if available. Psychiatric diagnoses were then determined by the best-estimate diagnostic procedure (Leckman et al., 1982). Numerical identifiers were used so that psychiatric diagnoses of children were conducted ‘blind’ with respect to the psychiatric status of parents.

J. Brigham et al. /Psychiaq

136

Research 73 (1997) 133-146

2.1.1. Diagnoses of parents Parental diagnoses were based on DSM-III-R (American Psychiatric Association, 1987) criteria, utilizing data collected with several instruments. Substance-use disorder diagnoses were based on DSM-III-R, using data collected by the CEDAR Substance Use Disorder Interview. This structured interview, designed specifically for the CEDAR study, utilized the Structured Clinical Interview for Diagnosis @CID) (Spitzer et al., 1987) substance use disorders section and the Lifetime Alcohol Use Interview (Skinner, 1982a). Screening information was gathered on 40 classes of psychoactive substances with the Drug and Alcohol Checklist (Skinner, 1982b), and detailed information was gathered on the four most frequently used substances, including alcohol. For each drug, data collection was organized into time periods (i.e. phases), typically months or years, during which drug-use frequency was relatively consistent. The first assessment period covered the period from first use to the first period when a substantial increase occurred. Subsequent phases were considered separately when the subject indicated that a substantial increase or decrease in quantity or frequency of drug consumption occurred with corresponding changes in con-

Table 2 Lifetime history

of psychiatric

disorders

Characteristic

Antisocial Personality Conduct

Parental

Disorder Disorder

sequences. For each time period, the subjects were asked about drug consumption quantity and frequency, the direction in and perceived reasons for the change from the previous time period, drug use consequences, and the typical social contexts and times of day of drug consumption. SCID questions were used to determine drug-use consequences according to DSM-III-R. Other psychiatric diagnoses in the father and mother were made with DSM-III-R, using data collected through several instruments. Conduct Disorder and Antisocial Personality Disorder were assessed by administering the SCID Personality Disorders Questionnaire (Spitzer et al., 1987) and confirming the positive endorsements by interview. The validity of this method has been documented (Nussbaum and Rogers, 1992). Other psychiatric disorders were assessed using the SCID (Spitzer et al., 1987). The groups also did not differ in categories of parental education level, as determined by chisquare analysis. The most common education level for fathers was some college, technical or business school. Mothers most commonly had a high school education. Fathers of boys in the POLY group differed from fathers of ALC and LOW in three non-substance-related diagnostic categories

substance

use

Total (n = 95)

Alcohol-only abuse or dependence (n = 29)

Polysubstance or dependence (n = 37)

Count

Count

Percent

Percent

abuse

No abuse or dependence (n = 29) Count

Mean

SD.

Percent

Father Mother

0

10

10

0

27 0

0

0

0

0

11 0

Son

3

3

8

3

5

5

Attention Deficit Disorder

Father Mother Son

0 0 19

5 0 9

14 0 24

3 0 7

6 0 18

6 0 19

Major Depressive Disorder

Father Mother Son

17 31 3

11 8 3

30 22 8

0 28 0

16 25 4

17 26 4

J. Brighamet al. /PsychiatryResearch 73 (1997)133-146

(see Table 2). Antisocial Personality Disorder and major depression were more common in the POLY fathers than in other groups. Although Attention Deficit Hyperactivity Disorder was more prevalent in POLY fathers and sons than in the other groups, the differences were not statistically significant. To determine their possible effect on ERP component measures, paternal and son attentional disorder diagnoses were considered separately and jointly as grouping factors in preliminary ANOVA, and as predictive factors in preliminary regression analyses. The results were non-significant. 2.1.2. Diagnoses of sons The child and one parent, typically the mother, completed a modified Schedule of Affective Disorders and Schizophrenia for Children interview (Orvaschel et al., 1982) concerning the index child’s psychiatric disorders. Mothers and teachers completed the Child Behavior Checklist (Achenbach and Edelbrook, 1983) and other questionnaires. 2.1.3. Recruitment Boys were ascertained through their fathers, who were recruited through substance-abuse treatment programs, newspaper advertisements and public service announcements. Informed consent to test the children was obtained from their parents, and informed assent was obtained from each child. Subjects were paid for participation. The protocol was approved by the Institutional Review Board of the University of Pittsburgh. 2.1.4. Exclusionary criteria No boy in this study met criteria for a substance-use disorder, had a significant history of substance exposure or experimentation, or evidenced current substance use as determined by biological assays. Boys in both risk groups (ALC and POLY) were screened through physical examination to eliminate any subject whose maternal history suggested in-utero exposure to alcohol or drugs, including those subjects with signs of in-utero teratogenic effects of maternal substance use as manifested by minor physical abnor-

137

malities. Boys were excluded from any group if their vital signs were out of the normal healthy range; if they or their parents had chronic physical disabilities, chronic neurological disease, neurological injury requiring hospitalization, schizophrenia or other psychotic disorders; if their parents could not read at eighth grade level; if they or their parents had a full-scale IQ score less than 80, or had a history of childhood psychosis, chronic physical disability, fetal alcohol/drug exposure or syndrome, or neurological injury requiring hospitalization; or if they had uncorrectable sensory deficits. Cognitive performance was screened in all subjects using the Wechsler Intelligence Scale for Children-Revised (WISC-III). Socio-economic status was assessed with the Hollingshead (1990) procedure (see Table 1). 2.1.5. Inclusionary criteria Boys, ages 10-12, were included in one of three groups, based on parental substance-use diagnosis: The ALC group consisted of boys whose biological father met criteria for a DSM-III-R lifetime diagnosis of Psychoactive Substance Use Disorder (SUD) of alcohol abuse or dependence (Table 3). Four ALC fathers met criteria for both an Alcohol Abuse and an Alcohol Dependence diagnosis. No parent of a boy in the ALC group met criteria for a diagnosis of drug abuse or dependence. The POLY group consisted of boys whose biological father met criteria for a DSM-III-R lifetime diagnosis of alcohol abuse or dependence and for a psychoactive substance-use disorder (SUD) for non-alcohol drug abuse or dependence (other than nicotine). Seven fathers of POLY boys met criteria for both alcohol abuse and dependence diagnoses. The LOW group consisted of boys at low risk for substance-use disorder, based on the fact that their biological parents did not meet criteria for either a psychiatric or a substance-use disorder. The three groups of boys did not differ significantly on any measured demographic or psychiatric variable (see Tables 1 and 2). Parents differed by age, with fathers of ALC subjects being approximately 4 years older than fathers of the

138

.I. Brigham et al. /Psychiatry Research 73 (1997) 133-146

Table 3 Parents’ lifetime psychoactive substance-use disorders Diagnosis

Parental substance-use disorder Alcohol-only abuse or dependence (n = 29)

Total (n = 95)

Polysubstance abuse or dependence (n = 37)

No abuse or dependence (n = 29)

Count

Percent

Count

Percent

Count

Percent

Father Mother

15 4

52 14

18 7

49 19

0 0

33 11

35 12

Alcohol dependence

Father Mother

19 3

66 10

26 8

70 22

0 0

45 11

47 12

Drug abuse

Father Mother

0 0

0 0

17 8

46 22

0

17 8

18 8

Father Mother

0 0

0 0

24 12

65 32

0 0

24 12

2.5 13

Father Mother

0 0

0

11 4

30 11

0

0

11

0

0

0

4

12 4

Father Mother

0 0

23 9

62 24

0

0

0

0

0

23 9

24 9

Father Mother

0

0

0

0

24 16

0

0

0

0

9 6

9 6

Father Mother

0

0

0

0

3 5

0

0

0

0

1 2

1 2

Father Mother

0

0

0

14 3

0

0

0

0 0

5 1

5 1

Father Mother

0

0

0

0

11 8

0

0

0

0

4 3

4 3

Overall substance use Alcohol abuse

Drug dependence

Specific drugs Amphetamine Cannabis Cocaine Hallucinogen Opiate Sedative

0

other two groups, and mothers in the ALC group being 3-5 years older than mothers of the boys in the other two groups. Mothers in the ALC group had approximately 1 year less education than mothers in the other two groups. Fathers in the LOW group scored 9-11 points (less than a standard deviation) higher on the Wechsler Adult Intelligence Scale (Verbal IQ and Full-Scale IQ) than fathers in the other two groups. As stated previously (see Sec. 2.1.11, the groups’ non-significant difference in prevalence of Attention Deficit Disorder did not statistically affect ERP component measures.

Count

Percent

0

2.2. ERP design

The subjects, i.e. the boys in the three groups, performed a counting task during the administration of an auditory oddball ERP paradigm. Subjects were instructed to keep their eyes closed during the testing. ERPs were collected within the rubric of an established 2-day protocol, and thus were all measured at approximately the same time of day. High- and low-pitched tones were presented in a 0.10 probability (24 target : 240 standard trials) presentation. Tones (50 ms in

J. Brighamet al. /PsychiatryResearch 73 (1997) 133-146

duration, 64.5dB SPL) were presented to subjects’ right and left ears via earphones, with one tone presented every 2 s. Tones were presented in the same random sequence to all subjects. Tone pitch was 1000 Hz for low (standard) tones and 2000 Hz for high (target) tones. Subjects were instructed to count the high-pitched tones mentally and were tested in discerning and counting high tones before any data were collected. After the testing, each subject’s tone count was noted. No subjects were excluded for having an excessive number of incorrect counts. As was the case with the larger sample from which these subjects were drawn, all subjects were within three tones of the correct count. Comparisons of hits/misses were not conducted because of the truncated nature of those data in this relatively easy tone-discrimination paradigm. 2.3. Recording procedures Auditory event-related potentials (ERPs) were recorded with Ag/AgCl electrodes from five scalp leads [Fz, Cz, Pz, P3 and P4; Jasper (1958)], referred to linked ears, with forehead ground. Bandpass was 0.01-30 Hz, with digitized samples obtained at 125 Hz for a 1200-ms epoch beginning 200 ms prior to stimulus onset. Electrodes were placed at the outer canthus and supraorbitally to the left eye for a bipolar recording of the electro-ocular (EOG) activity. Impedance for all recording leads was 5 KR or less. Artifacts were removed before averaging by processing all trials through a customized program that eliminated trials according to the following criteria of range and velocity: The maximum eye channel range was 120 PV per trial. The maximum EEG range was 90 ~.LVper trial. The maximum velocity change was 30 PV per two trials, or 1.875 PV per millisecond. The program also calculated a running average and eliminated trials that exceeded a 55-PV difference from the running average. Additionally, all trials not rejected by the automated program were screened visually for artifacts. Average ERPs were calculated from artifact-free trials. Data included in these analyses are from target trials, which were averaged separately from standard trials. The mean number of

139

retained high-tone trials was 21.78. The numbers of retained trials, which did not differ statistically by group, were 22.0 for POLY, 22.1 for ALC and 21.2 for LOW. 2.4. Data analysis 2.4.1. Peak identification

ERP components were identified visually from a graphic display of the numerical output of the averaged waveforms. Peaks were determined by a trained rater blind to the diagnosis of the subjects, to prevent bias. A second blind rater identified peaks on a subset of the recordings; the peaks selected by the two raters were identical for 96% of the peaks identified. Peak amplitudes were calculated as the deviation from the mean voltage during the baseline period. ERP components Nl, P2, N2 and P3 were identified by latency window, polarity, and order. Nl (90-150 ms post-stimulus) and N2 (200-275 ms> were the first and second negative peaks post-stimulus. P2 (130-220 ms> and P3 (250-450 ms) were the second and third positive peaks post-stimulus. Nl and P2 were calculated at Fz and Cz. N2 was calculated at all leads. The P3 component was determined for leads Cz, Pz, P3, and P4. 2.4.2. Analysis of variance Differences in ERP peaks’ amplitudes and latencies were statistically evaluated by multivariate analysis of variance (MANOVA), with main factors SUD (Substance Use Disorder) and Lead (electrode lead, i.e. Fz, Cz, Pz, P3, P4). SUD consisted of three groups, identified as ALC (paternal lifetime diagnosis of alcohol abuse or dependence), POLY (paternal lifetime diagnosis of polysubstance abuse or dependence, including an alcohol-use disorder) and LOW (no parental history of diagnosable substance abuse or dependence). Analyses were conducted separately for amplitude and latency measures, and for midline (Fz, Cz, Pz> and lateralized parietal (P3, P4) groupings. Pairwise group differences (e.g. ALC vs. POLY; ALC vs. LOW; and POLY vs. LOW) were

140

J. Brigham et al. /Psychiatry

Research 73 (1997) 133-146

contrasted using post-hoc simple Bonferroni tests of significance.

3.2. Analysis of variance

2.4.3. Principal components analysis (PCA) PCA was applied to decompose the effects of overlapping average components in averaged data from lead Cz, and to distinguish slow-wave activity from peaks. The PCA was also conducted to determine whether any amplitude differences were mediated by an overlapping component.

3.2.1. Main effects

3. Results 3.1. Data transfomtation

Preliminary analyses indicated that the ERP variables evidenced only slight to moderate skewness and kurtosis, and thus no normalization procedures were conducted.

Table 4 ERP component Component

Fig. 1 presents the ERP grand means for all electrode leads. The mean values and standard deviations are listed in Table 4 (amplitudes). Event-related potentials from boys in the ALC and POLY groups showed a post-stimulus negative shift relative to the amplitude at stimulus onset. The shift continued throughout the l-s recording period and resolved by the time baseline readings were taken 2 s later for the next presentation of a stimulus. At electrode lead Cz, this negative shift coincided closely with stimulus onset. This shift is illustrated in Fig. 1, the grand mean ERPs of the five electrode leads, and Fig. 2, depicting four factors from a PCA at Cz, as explained in Sec. 3.3. MANOVA (F2,276 = 28.33,

amplitudes Site

Parental

substance-use

disorder

Total (n = 95)

Alcohol-only abuse or dependence (n = 29)

Polysubstance abuse or dependence (n = 37)

Mean

S.D.

Mean

S.D.

No abuse or dependence (n = 29) Mean

Mean

SD.

S.D.

Nl amplitude

Fz cz Pz P3 P4

- 10.92 - 9.67 -6.13 - 8.55 - 7.94

6.94 7.60 5.68 6.10 5.70

- 8.21 - 8.67 - 6.93 -8.18 - 8.37

5.79 6.12 4.80 4.90 5.67

-8.14 - 5.93 - 4.64 - 7.46 - 6.29

6.71 4.89 4.64 4.44 5.02

- 9.02 -8.14 -5.99 -8.07 - 7.60

6.50 6.40 5.07 5.13 5.50

P2 amplitude

Fz cz Pz P3 P4

-4.21 2.55 5.03 -0.81 0.11

6.24 7.46 5.77 5.48 6.17

-3.62 1.13 2.99 - 2.45 - 0.77

6.40 7.06 6.63 5.27 6.55

-3.65 2.59 3.87 - 2.82 0.50

6.69 6.76 7.41 5.69 6.56

- 3.81 2.01 3.88 - 2.06 - 0.48

6.38 7.06 6.62 5.47 6.37

N2 amplitude

Fz cz Pz P3 P4

- 18.39 - 10.92 - 4.83 - 10.36 - 10.72

8.06 7.35 6.25 6.37 6.09

- 13.56 - 8.30 - 4.20 - 8.52 - 9.64

9.21 7.43 6.19 6.51 6.92

- 16.29 - 7.33 -2.05 - 8.35 - 10.24

8.92 8.38 7.91 8.51 7.29

- 15.87 -8.81 - 3.74 - 9.03 - 9.24

8.92 7.77 6.80 7.12 6.87

P3 amplitude

Fz cz PZ P3 P4

2.28 12.66 19.07 15.06 15.43

10.17 11.62 6.48 6.32 7.15

2.81 12.15 18.87 13.29 13.22

8.82 9.01 8.04 8.00 7.69

1.11 17.01 22.00 15.34 17.72

9.07 9.20 8.06 9.83 9.41

2.13 13.79 19.89 14.46 15.27

9.25 10.06 7.65 8.13 8.23

Bold type indicates

cells that differ significantly

between

groups.

J. Brigham et al. /Psychiahy Research 73 (1997) 133-146

Fig. 1. Event-related potentials (ERPs) from five electrode lead sites in a O.lO-probability oddball paradigm, for boys whose biological fathers have lifetime diagnosis of (1) polysubstance abuse or dependence (e.g. alcohol-use diagnosis plus at least one additional substance-use diagnosis), (2) alcohol abuse or dependence, or (3) no substance abuse or dependence.

142

J. Brigham et al. /Psychiatty Research 73 (1997) 133-146

‘1

-Factor1 (44.5%)

0.5 -

O-

-0.5

-1 -I 0

200

400

600

800

Latency (milliseconds)

Fig. 2. Four factors resulting from principal component analysis of amplitude measures at lead Cz.

P = 0.00) and post-hoc tests of electrode lead differences indicated that the negative shift was more pronounced in amplitude at Cz and Pz than at Fz. It was more pronounced at Cz than at Pz for all ERP component amplitudes except P2. 3.2,2. Analyses of amplitudes A negative shift across most of the epoch in recordings from the ALC and POLY groups was reflected in the amplitudes of waveform components (see Fig. 1). The ALC and POLY groups had larger midline Nl and N2 component amplitudes and smaller P3 component amplitudes than the LOW group. The ALC group had larger Nl and N2 amplitudes than the LOW group at all midline leads. The POLY group evidenced a similarly enlarged Nl and N2 amplitude along the midline, and a similarly diminished P3 amplitude at Cz and parietal leads. The ALC and POLY groups differed in N2 and P3 amplitudes along the midline, with ALC having a larger N2 amplitude than POLY at all leads, and a smaller P3 amplitude at Cz. MANOVA findings were as follows, with effect sizes reported as y, the absolute value of ([Xi -X,1/s): Nl: ALC and POLY groups had larger Nl amplitudes at midline leads (Fz, Cz and Pz) than the LOW group (model, P = 0.001; variable, F 2,276= 4.49, P = 0.012; post-hoc, P = 0.047 [POLY vs. LOW], P = 0.003 [ALC vs. LOW]). The strongest effect sizes were at Cz (y = 0.58)

for ALC vs. LOW, and at Cz (y = 0.43) and Pz (y = 0.45) for LOW vs. POLY. The strongest effect sizes were at Cz (y = 0.58) for ALC vs. LOW, and at Cz (y = 0.43) and Pz (y = 0.45) for LOW vs. POLY. N2: The ALC group had larger N2 amplitudes at midline leads (Fz, Cz and Pz) than the POLY and LOW groups (model, P = 0.001; variable, F 2,276= 3.76, P = 0.025; post-hoc, P = 0.017 [ALC vs. POLY], P = 0.018 [ALC vs. LOW]). Effect sizes were most powerful for ALC vs. LOW at Cz (y = 0.46) and Pz (y = 0.411, for POLY vs. LOW at Fz (y = 0.31) and Pz (y = -0.321, and for ALC vs. POLY at Fz (y = 0.54). P3: ALC and POLY groups had smaller P3 amplitudes at two midline leads (Cz and Pz> than the LOW group (model, P = .002; variable, F2,184 = 3.82, P = 0.024; post-hoc, P = 0.013 [POLY vs. LOW] and P = 0.014 [POLY vs. ALC]). MANOVA at lateral parietal leads (P3 and P4) reflected a non-significant data trend (P = 0.067) in the same direction as the midline results. Effect sizes were largest at Cz (y = 0.43) and Pz (y= 0.38) for ALC vs. LOW, and also at Cz (y = 0.48) and Pz (y = 0.41) for POLY vs. LOW. 3.2.3, Analyses The factor SUD status, between-group tencies.

of latencies

of Group, representing parental did not result in any significant differences in the boys’ ERP la-

3.2.4. Interaction effects There were no significant Group X Lead interaction effects. 3.3. Principal component analysis PCA conducted on Cz target-tone amplitude readings from 0 ms to 800 ms yielded four factors meeting criteria for significance (eigenvalue > 1.01, accounting for 80% of the total common variance, as depicted in Fig. 2. The first factor, accounting for 44.5% of the common variance, was a slow negativity beginning after 100 ms and resolving by 800 ms. The second factor, accounting for 16.1% of the variance, was a positive peak

.I. Brigham et al. /Psychiafry Research 73 (1997) 133-146

at approximately 300 ms post-stimulus onset, corresponding temporally to P3. All four factors contained a negative peak within the time window associated with ERP components Nl and N2. Separate analyses of variance of factor 1 and factor 2 summary scores were used to test their sensitivity to SUD group effects. ANOVAs on PCA components indicated that the POLY and LOW groups differed significantly for the time period 136-168 ms (t = 2.06 to 2.43, P < 0.05), 232-248 ms (t = 1.04 to 2.02, P < 0.051, and 376-392 ms (t = 1.99 to 2.23, P < 0.05). The PCA differences between ALC and LOW were significant in the 376-392 ms period (t = -2.10 to -2.21, P < 0.05). ALC and POLY differed significantly at 136-168 ms (t = 2.11 to 2.28, P < 0.05). 4. Discussion This study examined ERPs in boys believed to be at risk for an alcohol- or drug-related SUD. Because boys were ascertained on the basis of paternal substance use, the study examined familial effects, which could be environmental as well as biological. The study was not, however, designed to be a study of genetic effects. It identified electrophysiological differences between the two high-risk groups and the low-risk comparison group, as well as between the two high-risk groups. These findings thus extended previous alcoholand drug-abuse research by identifying additional aspects of the ERP that differed among these populations. Analyses indicated that these groups were characterized by distinct electrophysiological profiles, perhaps reflecting variations in information processing specific to young persons at risk for later development of substance abuse or dependence. 4.1. Relationship to other slow negative activity The negative shift noted in these findings differed from published reports of other slow negativities. The phenomenon reported here was similar in morphology to the processing negativity (PN) explored by NZXnen (1982, 1988>, but differed in scalp distribution and latency of onset;

143

PN is more frontal and fronto-central and has a later onset. The negativity identified in this study was prominent at Cz and Pz. This negative shift also had only partial resemblance to the anticipatory and preparatory negativity noted by Chwilla and Brunia (1991) and to the contingent negative variation (CNV) slow potential (McCallum et al., 1988>, both of which are more anterior. It also differed from the more frontally located searchrelated negativity associated with memory load (DeSonneville et al., 1989). Similarly, the mismatch negativity examined by Alho [summarized in Alho (1992)] has a fronto-central distribution more anterior than that of this report’s negative shift. This phenomenon also varied in scalp distribution, onset and duration from a negativity observed in those with obsessive-compulsive disorder (Towey et al., 1994). Despite those differences, it remains possible that factors related to those negativities could be implicated in this phenomenon. Previously cited slow-negativity findings dealing primarily with adults might not be directly comparable to data from children, since topographic patterns of electrocortical activity are known to change markedly across childhood development (e.g. Berman and Friedman, 1995; Taylor and Pourcelot, 1995; Yordanova and Kolev, 1996). Additionally, the observation by Michie et al. (1990) that a slow negativity could reflect arousal levels could relate to responsivity anomalies noted previously in this high-risk population, e.g. cortisol hyporesponsivity associated with behavioral dysregulation (Vanyukov et al., 1993; Moss et al., 1995). The electrophysiological profiles of at-risk children invite speculation about cognitive correlates of risk status. Slow negative waves have been associated with a variety of cognitive processes, including extended working memory (Ruchkin et al., 19881, activation of cognitive systems involved in information storage and transformation [RGsler et al. (1995); parietal and occipital], concept formation and conceptual difficulty (Ruchkin et al., 19SS>, long-term retrieval [Riisler et al. (1993); fronto-central], or semantic complexity (Gunter et al., 1995). Kok (1997) summarized that these many variations of slow waves can be elicited through tasks using differing conceptual operations when

144

J. Brigham et al. /Psychiatry Research 73 (1997) 133-146

task demands are high. As processing demands increase, the amplitude of the slow waves increases. Ruchkin et al. (19881, in a review of slow-wave sources and activities, concluded that negative slow waves indicate difficulty of working memory’s conceptual operations. Kok (1997) clarified negative slow waves as reflecting ‘activities of different components of working memory during controlled processing’. He postulated that different types of these waves represent ‘the mobilisation of different subtypes of central resources (e.g. codes of processing) in specific brain areas’, as modeled by Wickens (1980, 1984). It is also possible, as Kok noted, that negative slow waves indicate activation of different posterior cortical ‘modules’ that are controlled by the same anterior attentional network. 4.2. Relationship to P3 findings The possibility that cognitive anomalies might be reflected in the ERP activity of at-risk groups also raises questions about P3 findings, particularly since the relationship between slow-wave activity and P3 is unclear. Some scientists have suggested that P3 is reduced under high processing demands as a result of overlapping slow-wave negativity (e.g. Wijers et al., 1989). Although principal component analyses have successfully separated overlapping ERP components, the extent to which sustained negative waves influence P3 amplitude is uncertain. Ruchkin et al. (1980) stated that overlap between slow wave activity and P3 may account for the difficulty in establishing an underlying construct for P3. Kok (1997) proposed that this appearance of overlap could be affected by overlapping mental processes (e.g. Meyer et al., 1988) or could result from ‘smearing’ of variable-latency components through the averaging process (Wijers et al., 1989). All of these factors make the role of any single ERP component as an index of vulnerability uncertain. 4.3. High- vs. low-risk group comparability An additional

question beyond the scope of this article relates to the relationship of this observed negative shift to the diminished P3 am-

plitude seen in many studies of offspring of alcoholics. Although many of those studies published prior to 1994 were included in a meta-analysis (Polich et al., 1994), no research published to date has examined the inter-study consistency of all inclusionary, exclusionary and matching criteria used to establish comparability of high- and low-risk groups. Would at least some of those studies have also shown this negative shift, which herein accounts for between-group P3 amplitude differences, had the same subject-selection criteria been used? That remains unexamined. This study by no means established or even explored genetic effects; instead, it examined familial effects that could be influenced by a broad array of environmental factors. The negative shift associated with risk status could be related to factors that are independent of parental substance use per se, e.g. the educational equivalency among schools attended by the three groups, cultural and environmental elements to which boys in the higher-risk groups may have become acclimated, hypervigilance, or environmentally linked attentional factors. 4.4. Alcohol risk vs. polysubstance risk Interestingly, ERPs from the two risk groups were differentiated only by minor variations. Among them was the larger N2 amplitude in ALC, as compared to POLY, particularly at Fz. The extent to which N2 and the negative shift overlapped, with one obscuring the other, was unclear in this paradigm. The PCA decomposing Cz demonstrated a marked negative peak in factor 2 that corresponded temporally with N2 (Fig. 2). Analysis of variance distinguished ALC and POLY between-groups differences related to the time window associated with Nl (136-168 ms>. The question of whether these differences involved information-processing variations associated with the N2 component or whether they resulted from slow-wave activity awaits further inquiry. 4.5. Conclusion Additional exploration using ERP paradigms will be necessary to clarify the differences noted

145

.I. Brigham et al. /Psychiatry Research 73 (1997) 133-146

in this article, perhaps utilizing ERP tasks focused on Nl-related information-processing functions (e.g. allocation of perceptual resources, modulation of attention, or information selection based on stimulus properties). A research paradigm designed to elicit and examine Nl and N2 specifically - as, indeed, this paradigm was not designed to do - could also clarify the extent to which these differences are separate from the negative shift. Similarly, the negative shift and the P3 amplitude differences should be examined through EPR tasks focusing on information processing tasks associated with P3 (e.g. response set, task difficulty, attentional processes, memory and demands on perceptual-central processing resources), as well as tasks examining anticipation. Consistent with a number of previous reports, these results strongly suggest an electrophysiological component to the biopsychosocial factors affecting vulnerability to alcohol and drug abuse and dependence. This research also raises significant questions about the interpretation of results in ERP studies of risk groups. Additionally, these analyses underscore the need for careful use of subject-selection criteria before basing prima facie assumptions about biological vulnerability and cognitive anomalies on electrophysiological findings.

American Psychiatric Association, 1987. Diagnostic and Statistical Manual of Mental Disorders, 3rd ed., revised. American Psychiatric Press, Washington, DC.

Acknowledgements

Hollingshead, A.B., 1990. Four Factor Unpublished manuscript.

This work was conducted at the Center for Education and Drug Abuse Research (a consortium of the University of Pittsburgh and St Francis Medical Center); it was supported by grants from the National Institute on Drug Abuse (P50DA 05605) and the National Institute on Alcohol Abuse and Alcoholism (AA 07453-12). The authors gratefully acknowledge the technical assistance of Gary Doheny and David Mazzarella.

Jasper, H.H., 1958. The ten-twenty electrode system International Federation. Electroencephalography Clinical Neurophysiology 10, 371-375.

References Achenbach, T., Edelbrook, C., 1983. Manual for the Child Behavior Checklist and Revised Child Behavior Profile. TM Achenbach, University of Vermont, Burlington, VT. Alho, K., 1992. Selective attention in auditory processing as reflected by event-related brain potentials. Psychophysiology 29, 247-263.

Begleiter, H., Porjesz, B., Bihari, B., Kissin, B., 1984. Event-related brain potentials in boys at risk for alcoholism. Science 225, 1493-1495. Berman, S., Friedman, D., 1995. The development of selective attention as reflected by event-related brain potentials. Journal of Experimental Child Psychology 59, 1-31. Berman, SM., Whipple, S.C., Fitch, R.J., Noble, E.P., 1993. P3 in young boys as predictor of adolescent substance use. Alcohol 10, 69-76. Brigham, J., Herning, RI., Moss, H.B., 1995. Event-related potentials and alpha synchronization in preadolescent boys at risk for psychoactive substance use. Biological Psychiatry 37,834-846. Chwilla, D.J., Brunia, C.H.M., 1991. Event-related potential correlates of non-motor anticipation. Biological Psychology 32, 125-141. DeSonneville, L.M., Visser, S.L., Njiokiktjien, C., 1989. Late sequelae of a non-optimal neonatal condition in ERPs at the age of 11-13 years. Electroencephalography and Clinical Neurophysiology 72, 491-498. Elmasian, R., Neville, H., Woods, D., Schuckit, M., 1982. Event-related brain potentials are different in individuals at high and low risk for developing alcoholism. Proceedings of the National Academy of Sciences 79,7900-7903. Gunter, T.C., Jackson, J.L., Mulder, G., 1995. Language, memory and aging: an electrophysiological exploration of the N400 during reading of memory demanding sentences. Psychophysiology 32, 2155229. Herning, RI., Hickey, J.E., Pickworth, W.B., Jaffe, J.H., 1989. Auditory event-related potentials in adolescents at risk for drug abuse. Biological Psychiatry 25, 598-609. Index of Social Status. of the and

Kok, A., 1997. Event-related potential (ERP) reflections of mental resources: a review and synthesis. Biological Psychology 45, 19-56. Leckman, J.F., Sholomskas, D., Thompson, W.D., Belanger, A., Weissman, M.M., 1982. Best estimate of lifetime psychiatric diagnosis: a methodological study. Archives of General Psychiatry 39, 879-883. McCallum, W.C., Cooper, R., Pocock, P.V., 1988. Brain slow potential and ERP changes with operator load in a visual tracking task. Electroencephalography and Clinical Neurophysiology 69, 453-468. Meyer, D.E., Osman, A.M., Irwin, D.W., Yantis, S., 1988. Modern mental chronometry. Biological Psychology 26, 3-67. Michie, P.T., Bearpark, H.M., Crawford, J.M., Glue, L.C., 1990. The nature of selective attention effects of auditory event-related potentials. Biological Psychology 30, 219-250.

146

J. Brigham et al. /Psychiatry Research 73 (1997) 133-146

Moss, H.B., Vanyukov, M.M., Martin, C.S., 1995. Salivary cortisol responses and the risk for substance abuse in prepubertal boys. Biological Psychiatry 38, 547-555. Naltanen, R., 1982. Processing negativity: an evoked-potential reflection of selective attention. Psychological Bulletin 92, 605-640. Nlatlnen, R., 1988. Implications of ERP data for psychological theories of attention. Biological Psychology 26,117-163. Nussbaum, D., Rogers, R., 1992. Screening psychiatric patients for Axis II disorders. Canadian Journal of Psychiatry 37, 658-660. Orvaschel, H., Puig-Antich, .I., Chambers, W., Tabrizi, M.A., Johnson, R., 1982. Retrospective assessment of prepubertal major depression with the Kiddie-SADS-E. Journal of the American Academy of Child Adolescence and Psychiatry 21,393-397. Polich, J., Pollock, V.E., Bloom, F.E., 1994. Meta-analysis of P300 amplitude from males at risk for alcoholism. Psychological Bulletin 115, 55-73. Rosier, F., Heil, M., Glowalla, U., 1993. Monitoring retrieval from long-term memory by slow event-related brain potentials. Psychophysiology 30, 170-182. Rosier, F., Heil, M., Bajric, J., Pauls, C.A., Henninghausen, E., 1995. Patterns of cerebral activation while mental images are rotated and changed in size. Psychophysiology 32, 135-154. Ruchkin, D.S., Sutton, D.S., Kietzman, M.L., Silver, K., 1980. Slow wave and P300 in signal detection. Electroencephalography and Clinical Neurophysiology 50, 35-47. Ruchkin, D.S., Johnson, R., Jr., Mahaffey, D., Sutton, S., 1988. Towards a functional categorization of slow waves. Psychophysiology 25, 339-353. Skinner, H.A., 1982a. Alcohol dependence syndrome: measurement and validation. Journal of Abnormal Psychology 91, 199-209.

Skinner, H.A., 1982b. Development and Validation of a Lifetime Alcohol Consumption Assessment Procedure. Final Report, Health and Welfare Grant #6606-1693-42. Spitzer, R., Williams, B., Gibbon, M., 1987. Instruction Manual for the Structured Clinical Interview for DSM-III-R (SCID, 4/l/87 revision). Biometrics Research Department, New York State Psychiatric Institute, New York, N.Y. Taylor, M.J., Pourcelot, L., 1995. Cognitive evoked potentials in children: normal and abnormal development. Neurophysiologie Clinique 25, 130-145. Towey, J.P., Tenke, C.E., Bruder, G.E., Leite, P., Friedman, D., Liebowitz, M., Hollander, E., 1994. Brain event related potential correlates of overfocused attention in obsessivecompulsive disorder. Psychophysiology 31, 535-543. Vanyukov, M.M., Moss, H.B., Plail, J.A., Blackson, T., Mezzich, A.C., Tarter, R.E., 1993. Antisocial symptoms in preadolescent boys and in their parents: associations with cortisol. Psychiatry Research 46, 9-17. Wickens, C.D., 1980. The structure of attentional resources. In: Nickerson, R., Pew, R. (Eds.), Attention and performance, vol. 8. Lawrence Erlbaum Associates, Hillsdale, NJ. Wickens, C.D., 1984. Processing resources in attention. In: Parasuraman, R., Davies, R. (Eds.1, Varieties of attention. Academic Press, New York, NY. Wijers, A.A., Otten, L.J., Feenstra, S., Mulder, G., Mulder, L.J.M., 1989. Brain potentials during selective attention, memory search and mental rotation. Psychophysiology 25, 33-60. Yordanova, J.Y., Kolev, V-N., 1996. Developmental changes in the alpha response system. Electroencephalography and Clinical Neurophysiology 99, 527-538.