Rxq.

NeumPsychnp-.

& Bti.

Psychht.

Copyright0 PrIMed

ELSEVIER

1998. Vol. 22, PP. 576-583 1998 ELsevkr Science Inc.

in the u6A. Au rlghts reserved 0278-5846/98 $19.00 + .OO

PII SO278-S846(98)00027-X

AUDITORY

P300 AND SELF’-REPORTED IMPULSIVE AGGRESSION

JOHN E. GERSTLE, CHARLES W. MATHIAS and MATTHEW S. STANFORD Department of Psychology, University of New Orleans, New Orleans, LA, USA (Final form, February 1998)

Abstract

Gerstle, John E., Charles W. Mathias and Matthew S. Stanford: Auditory P300 and Self-Reported Impulsive Aggression. Prog. Neuropsychopharmacol. & Biol. Psychiat. 1998, 22. pp. 575-583. 01998 Elsevier!ScienceInc. 1. The purpose of the present study was to determine the cognitive psychophysiological correlates of impulsive aggression in a population considered “normal” by societal standards: college students. 2. Auditory event-related potentials were acquired on all subjects during a standard oddball task. The stimuli consisted of a random sequence of two tones, a frequent 1,000 Hz tone and a rare 2,000 Hz tone. Tones were presented in a ratio of 80120. 3. Results of the study demonstrated that impulsive aggressive subjects show significantly lower P300 amplitude at frontal electrode sites when compared to nonaggressive controls. 4. These findings are consistent with the psychophysiological findings in impulsive aggressive incarcerated criminals and support the notion of a specific behavioral syndrome associated with spontaneous aggressive outbursts. Kevwords: aggression, event-related potentials, impulsiveness, P300 Abbreviations: auditory event-related potential (AEP), Barratt Impulsiveness Scale (BIS-1 l), Buss-Durkee Hostility Inventory (BDHI), event-related potential (ERP), impulsive aggressive (IA), nonaggressive control (NC), Peabody Picture Vocabulary Test-Revised (PPVT-R).

Introduction

Impulsive aggression is defined as a hair trigger aggressive response to provocation with loss of behavioral control. This loss of control is not secondary to any medical or psychiatric disorder and, by virtue of the spontaneity of the act, is not planned (Barratt, 1991). Within the clinical and aggression literature, this type of outburst has been given a variety of diagnostic labels: Episodic Dyscontrol Syndrome (Monroe, 1970; Bach-y-Rita et al., 1971), Psychotic Trigger Reaction (Pontius, 1987), Intermittent Explosive Disorder (DSM-IV; American Psychiatric Assoc., 1994) and Impulsive Aggression (Coccaro et al., 1989; Barratt, 1991). While each label has its own proponents and detractors, they share the same basic symptom: overt aggressiveness in at least one situation with a loss of behavioral control. The physiological basis of this loss 575

576

J.E. Gerstle et al.

of behavioral control has long been the focus of intensive study (Moyer, 1968; Garza-Trevino, 1994). The P300 component of the ERP has been used in the study of aggressive behavior. The latency of the P300 has been suggested to measure stimulus evaluation time while the amplitude has been related to efficiency of cognitive processing (Donchin and Coles, 1988; Hillyard and Kutas, 1983). Consistently, studies of the P300 in aggressive populations have shown significantly lower amplitude in aggressive subjects compared to nonaggressive controls (Branchey et al., 1988; Drake et al., 1988; Bauer et al., 1994; Barratt et al., 1997). Lower P300 amplitude has been suggested to involve fewer neural resources and may be associated with less efficient cognitive timctioning (Duncan et al., 1994). While the results of these studies suggest that cognitive psychophysiological differences may be involved in violent behavior, many of the studies cited tend to treat violence as a homogeneous construct combining premeditated and impulsive aggressive individuals into a single group. Further, most of the subjects are either incarcerated prison inmates or neuropsychiatric patients referred for evaluation.

Such research

ignores the large number of individuals in the general population who commit nontrivial acts of violence yet are not currently in the criminal justice or mental health systems. These two sampling problems greatly decrease the usefidness of the results in the development of interventions and treatments for specific forms of violence. The present study was designed to correct these methodological problems by focusing only on impulsive aggressive behavior and by using aggressive subjects sampled from a population considered “normal” by societal standards, college students,

It was hypothesized that the P300 peak amplitude of impulsive

aggressive subjects would be significantly lower compared to nonaggressive matched controls.

Methods

Subjects (N=44) were obtained through screening questionnaires distributed in classes and through selfreferral. Informed consent was obtained from all potential subjects. Potential subjects were then asked to complete a modified version of the Anger Attack Questionnaire (Fava et al., 1991) and the Irritability subscale of the BDHI (Buss and Durkee, 1957). Both of these scales have been found usetbl in obtaining information related to impulsive aggressive outbursts (Stanford et al., 1995). Classification: Imuulsive Aeeressive vs. Nonaggressive To be classified as impulsive aggressive a subject must have met the following criteria: 1) Over the past six months, the subject identified several discrete episodes of failure to resist aggressive impulses that

P300 and

nnpulstve aggression

577

resulted in serious assaultive acts or destruction of property; 2) The degree of aggressiveness expressed during the episodes was grossly out of proportion to any precipitating psychosocial stressors; 3) At least two impulsive aggressive episodes occurred during the previous month; and 4) The subject scored an eight or greater on the Irritabiity subscale of the BDHI. Previous research has demonstrated that subjects who score eight or greater on the Initabiity subscale consistently exhibit chronic aggression control problems (Stanford et al., 1995). The first two criteria used in the present study are also the first two diagnostic criteria for Intermittent Explosive Disorder (DSM-IV, American Psychiatric Assoc., 1994). Subjects reporting that they had not had a temper outburst in the previous six months were classified as nonaggressive controls. Impulsive aggressive subjects were contacted by phone in decreasing order of Irritabiity score and number of aggressive outbursts. Nonaggressive control subjects were randomized and matched to impulsive aggressives by age and sex. All subjects were right handed and no subjects were presently taking any medication. Aggression Interview The focus of the interview was to obtain information concerning the subject’s temper and past aggressive behavior.

Both groups of subjects were questioned about their aggressive behavior and the following

information was obtained: severity of outbursts, frequency of outbursts and most common target of outbursts. Information concerning drug/alcohol use and head injuries was also obtained. A subject was rated positive for marijuana, heroine, LSD or cocaine use if they reported ever using the substance in the past. A subject was rated positive for alcohol use ifthey reported drinking alcohol more than once per week during a six month period. A subject was rated positive for a head injury if he/she had been knocked unconscious for any period of time, had been diagnosed with a concussion by a physician or had suffered trauma to the head severe enough to require medical attention. In addition, all subjects were asked to complete the BIS-11 (Patton et al., 1995), a measure of general impulsiveness and the PPVT-R (Dunn and DUM, 1981). The PPVT-R provides an estimate of verbal IQ. Auditorv Event-Related Potentials Psychophysiological

measures were taken in the morning (0730-1100 hours) to control for diurnal

variations in scalp electrical patterns. Subjects sat in a comfortable chair in a sound and light attenuated room during the procedure. Fifteen scalp electrodes (International lo-20 system) were used in the present study: F3, Fq F4, T3, C3, Cz, C4, T4, TS, P3, Pz, P4, T&01,02.

Impedance for each electrode was less

that 5 Kohms. Electrodes were referenced to linked ears and the subjects were grounded via an electrode on the forehead. All cognitive psychophysiological data were acquired and analyzed using QND software (Neurodata, 1995). Filter bandpass was set at 1 and 35 Hz.

578

J.E. Gerstle et al. The auditory task consisted of one block of 200 trials. The stimuli consisted of a random sequence of two

tones, a frequent

1,000 Hz tone and a rare 2,000 Hz tone.

Tones were presented

in a ratio of 80/20.

Subjects were asked to sit with their eyes closed and silently count to themselves the rare tones.

Tones were

presented at a comfortable level of 70 dl3 through headphones at a rate of one tone every two seconds, with a Sms rise/fall and 50ms sustain.

AEPs were digitized every 4ms for a 1,OOOmsinterval beginning 1OOms

prior to stimulus onset. Statistical Analysis P300 was defined as the largest positive peak in the 250-600ms window. by measuring the peak from the baseline.

Peak amplitude was determined

Peak latency was determined by measuring from stimulus onset

to the peak The 1OOmsrecording prior to stimulus onset was used to determine baseline.

Mixed ANOVAs

with group as the between factor and site (either midline sites [Fz, Cz, and Pz] or lateral sites [left vs right]) as the repeated factor were used to analyze the data.

Results Screening All 22 impulsive aggressive aggression towards others.

subjects (7 males, 15 females) described

Nonaggressive

controls (n=22) reported that they had no problem controlling

their temper and had never had an aggressive outburst. (M=20.25, SD=l.77).

a lifetime history of physical

Subjects (N=44) ranged in age from 18 to 24

As can be seen in Table 1, 15 ofthe impulsive aggressive subjects described the most

common target of their aggression as family members and/or friends while the impulsive aggressive outbursts of the others were more generalized. Nine of the impulsive aggressors (40.9%) and six of the nonaggressive controls

(27.3%)

met criteria for a head injury

0.42

[Yate’s corrected],p

This difference failed to reach statistical significance

> .05). Impulsive aggressive subjects were also more likely than controls to

be involved in drug and alcohol use (Marijuana: IA 50%, NC 22.7%; Heroin: IA 4.5%, NC 0%; Cocaine: IA 13.6%, NC 0%; LSD: IA 22.7%, NC 4.5%; Alcohol: IA 36.4%, NC 18.2%). Of subjects rated positive for marijuana use, 36% of the IA subjects reported weekly use while none of the NC reported using marijuana regularly. BIS-11 and PPVT-R Impulsive aggressive (F(1,42)=15.02,~

subjects scored significantly higher than nonaggressive

controls on the BIS-11

< ,001; IA: M=75.09, SD=13.76; NC: M=60.91, SD=10.26).

No significant difference

was found between groups on the PPVT-R (IA: M=104.91, SD=12.69; NC: M=103.68, SDc13.65).

F F F M F F M F M F F M F M F F M

F F F F M

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

18 19 20 21 22

N/A = No head injury

Sex

Subject

19 21 23 18 23

20 19 22 20 20 18 18 19 20 19 19 22 18 18 19 22 22

Age

Concussion (5 y/o) N/A Closed Head Injury (7 y/o) Unconscious (17 y/o) N/A N/A Unconscious (14 y/o) N/A Concussion (16 y/o) N/A Unconscious (10 y/o) Unconscious (12 y/o) Unconscious (15 y/o) N/A N/A N/A Multiple Concussions (18,21,22 y/o) N/A N/A N/A N/A N/A

Nature of Head Injury (if any)

Family Family Family Family Family

I Friends I Boyfriend

I Friends

/ Boyfriend

I Boy-hiend I Boyfriend

Members / Boyfriend Members Members Members I Friends Members

Family Members Family Members Family Members Generalized Family Members Family Members Generalized Generalized Generalized Generalized Family Members Family Members Friends Generalized Family Members Family Members Generalized

Target of Aggression

Descriptive Information on Impulsive Aggressive Subjects

Table 1

3 6 3 3 10 5 8 4 4 3 5 2 3 4 2 5 5

Number of Impulsive Aggressive Episodes (Month Prior)

580

J.E. Geratle et al.

There were no .significantdiikrences for P300 latencies. Significant within and between groups differences were observed for P300 amplitude.

Repeated measures analysis between midline sites for the rare (.2

probability) and the frequent (.8 probability) tones showed significant within group increases in amplitude proceeding from frontal towards parietal sites (Rare: F(2,84)=53.08,

p < ,001).

The frequent

condition

p < ,001; Frequent: F(2,84)= 21.11,

also showed a significant difference

between

groups (Frequent:

F(1,42)=3.89, p < .OS) with the IA group having lower amplitudes compared to the NC group (Fig 1). A similar trend for the rare tone failed to reach statistical significance. Significant

laterality and between groups differences

frequent condition (Late&y:

F(1,42)=10.89,p

were found at electrode

sites F3 and F4 in the

< .Ol; Croup: F(1,42)=5.45, p < .OS) with lower amplitudes

found at the let-l site compared to the right and lower amplitudes in the IA group compared to the NC group. A significant later&y X group interaction was found for amplitude at electrode sites T5 and T6 in the frequent condition ([IA:T%T6]