Use of Visual Evoked-Potential Studies

and EEG Data to Classify Aggressive,

Explosive Behavior of Youths
Donald R. Bars, Ph.D.
F. La Marr Heyrend, M.D.
C. Dene Simpson, Ph.D.
James C. Munger, Ph.D.

Objective: Data from electroencephalograms (EEGs), including data them achieve success in the world
from visual evoked-potential studies, were analyzed to assess their as- around them. In a study by Cadoret
sociation with a specific type of explosive behavior in children and ado- and associates (3), a significant amount
lescents. Methods: Data for 326 children and adolescents treated in a of the variability in the aggressivity of
psychiatric clinic were examined. Eighty-two percent exhibited behav- adopted children, as well as in conduct
ior consistent with intermittent explosive disorder, although diagnosis disorder and adult antisocial behavior,
was not an inclusion criterion for the study. The presence of explosive was accounted for by environmental
behaviors was indicated by reports from the legal system, schools, par- effects and genetic-environmental in-
ents, health care workers, and psychiatric intake interviews. A quanti- teractions. Empirical evidence of
tative EEG and a series of pattern-reversal evoked-potential studies these interactions is important to both
were administered to each patient. In these studies, children are shown basic science and applied social sci-
a rapidly reversing checkerboard pattern or rapid flashes of light, and ences for intervention and prevention.
their brain waves (evoked potentials) are measured. Results: Logistic re- As Elliott (4) noted, “As with all be-
gression indicated that patients who exhibited explosive behaviors were havior, aggression is the result of inter-
significantly more likely to produce high-amplitude P100 wave forms in actions between the brain (including
the evoked-potential studies than patients who did not exhibit explosive psychological experience) and envi-
behaviors. Forty-six percent of those with explosive behaviors met the ronmental challenges—a biopsychoso-
clinically defined electrophysiological criteria for the high-amplitude cial phenomenon.”
P100 wave forms. Conclusions: The use of visual evoked-potential stud- Although environmental factors are
ies helped us classify a large subset of youths who exhibited out-of-con- important, a review of the literature
trol explosive behaviors. The findings suggest that a subgroup of indi- reveals that many aggressive behaviors
viduals exhibiting explosive behaviors may have a predisposition for vi- have a biological basis, and when such
olent or explosive behavior that is an innate characteristic of their cen- behaviors are not identified and ad-
tral nervous system. An understanding of the etiology of explosive be- dressed appropriately, positive treat-
haviors permits the use of more appropriate intervention and treatment ment outcomes are extremely difficult
strategies. (Psychiatric Services 51:81–86, 2000) to obtain (5). Matthews and his col-
leagues (5) have labeled some patients
“neurobehavioral” because they be-

W
ith each violent outburst by by Stevens and colleagues (1). lieve that underlying abnormalities in
an adolescent in our society, The deterioration of family struc- brain function are the primary cause of
early identification of chil- ture and the noticeable moral decline their explosive behavioral outbursts.
dren who exhibit explosive, aggressive, in society (2) have only increased the Studies evaluating the electrophysi-
and uncontrollable outbursts becomes need to accurately identify individuals ological correlates of aggressive, explo-
more critical. This difficult population with explosive and aggressive tenden- sive, and assaultive behaviors have re-
was identified more than 30 years ago cies and to develop methods to help lied primarily on standard electroen-
cephalogram (EEG) data (6–11). In
two previous studies, we evaluated the
Dr. Bars is vice-president of research and development and Dr. Heyrend is president relationship between pattern reversal,
and medical director of Behavioral Management Centers, 411 North Allumbaugh Street, visual evoked potentials, and explosive
Boise, Idaho 83704 (e-mail, dbars@hotmail.com). Dr. Simpson is in private practice in behaviors. An evoked potential is a
Boise. Dr. Munger is chair of the biology department at Boise State University. regular pattern of electrical activity
PSYCHIATRIC SERVICES ♦ January 2001 Vol. 52 No. 1 81
recorded from neural tissue that is Methods episodic dyscontrol syndrome was
evoked by a controlled stimulus. In An evaluation of electrophysiological, recorded in the clinical file. A total of
pattern-reversal visual evoked-poten- behavioral, and medication data was 267 of the 326 children (82 percent)
tial studies, a subject is shown a rapid- conducted for all children through 18 met these criteria. In our 1991 ex-
ly alternating pattern, and the subject’s years of age who successfully complet- ploratory study, only 26 percent of the
EEG data are examined for amplitude ed a standard protocol of computerized sample exhibited explosive behaviors.
and latency of the wave forms, the sym- electroencephalographic and evoked- The difference reflects an increase in
metry of the responses, and any abnor- potential evaluations during 1995 and the proportion of explosive individuals
mal activity. Such abnormalities often 1996. Before coming for evaluation, among our clinical population because
indicate difficulties with behavior and many of these patients had been placed of our clinic’s growing reputation for
learning. The etiology of explosive be- in one or more residential treatment working with explosive, out-of-control
haviors may be different in children facilities and had been seen by several individuals.
who exhibit such abnormalities. therapists and doctors, yet their uncon- To account for any effect medication
In our first study of 278 children and trollable behaviors continued. might have had on the electrophysio-
adolescents, 71 (26 percent) exhibited Of the 454 children and adolescents logical data, a dichotomous variable,
explosive behaviors, and 237 (85 per- evaluated, complete data were avail- medications, was created, which indi-
cent) exhibited at least one type of able for 326—105 females (32 per- cated whether or not an individual was
brain electrical abnormality (12). The cent) and 221 males (68 percent)— taking medication at the time of testing.
other study compared six explosive in- who were included in this study. Chil- Informed consent was given by the
dividuals (four males and two females) dren excluded from the study were participants or their parents or
who were matched by age, sex, and those who had only a sleep-deprived guardians after we reviewed the pro-
handedness with six children who did EEG—that is, those for whom evalua- cedure and answered their questions.
not have a psychiatric diagnosis (13). tion of evoked potentials was not com-
The mean±SD age of the children was pleted or those who did not correctly Electrophysiological procedure
13.5±2.06 years. In the pattern-rever- count the number of target stimuli giv- Each patient underwent a series of
sal evoked-potential studies, the chil- en during the auditory-event-related evoked-potential studies and a quanti-
dren with explosive behaviors were potential studies. We did not consider tative EEG in accordance with guide-
found to have significantly higher N75 the amplitude of wave forms or any lines of the American Electroen-
and P100 wave forms than the controls. specific behavioral characteristics for cephalographic Society (14,15). We
This study investigated whether a the children not included in the study. used the Brain Atlas III, a product of
relationship exists between quantita- The mean±SD age of the 326 chil- the Bio-Logic Corporation in Chicago.
tive EEG data, visual evoked poten- dren was 13.27±2.92 years; the mean Electrode placements were in accor-
tials, and explosive behavior of chil- age of those with explosive behaviors dance with the international 10-20 sys-
dren and adolescents referred to our was 13.03±2.95 years, and the age of tem, using an Electro-Cap, with 16 ac-
clinic. During the exploratory studies those without explosive behaviors was tive electrodes. A monopolar montage
(12,13), one variable in the visual 14.32±2.52. Explosive behavior was with forehead ground with linked-ear
evoked-potential series—the P100 defined as that with apparent limbic reference was used. Electrode imped-
wave form—consistently showed a sig- system involvement. Typically it oc- ance was maintained at less than 2
nificant relationship with explosive be- curs illogically, in situations that would kohms, and the impedance between
havior. Therefore, in the study report- not be expected to prompt an out- homologous sites was maintained with-
ed here we hypothesized that the am- burst. Such children control them- in 1 kohm. The gain was set at 30,000,
plitude of the P100 wave form in the selves in many situations in which a the low-pass filter at 100 Hz, the high-
pattern-reversal visual evoked-poten- strong emotional response would be pass filter at 1 Hz, and the 60-Hz notch
tial studies would be significantly relat- understandable; however, an inconse- filter was set to the “in” position.
ed to explosive, aggressive, and out-of- quential situation causes an “explo- Patients were comfortably seated in
control behaviors. sion” of emotions. The outburst lasts at a padded reclining chair in a small,
The data we used are a subset of an least 20 minutes and sometimes goes sound-attenuated room. A channel-by-
extensive database gathered over two on for a day or more. When the chil- channel calibration was performed be-
years involving electrophysiological dren are approached, the behavior is fore and after each recording session.
parameters of various psychiatric be- escalated, and the length of the out- The electrophysiological test series
haviors. It should be emphasized that burst is extended. Usually they are consisted of four visual evoked-poten-
our research deals with behaviors and apologetic after the outburst, and they tial studies, including pattern reversal
not diagnostic categories, because ex- sometimes deny remembering what (both eyes, left eye, and right eye) and
plosive behaviors cut across a variety of occurred. flash (both eyes); three auditory-ev-
diagnoses. The results of this study In our study, an individual was con- ent-related potential studies (the odd-
provide possible evidence of a specific sidered to be explosive if any mention ball paradigm at three different speeds);
class of explosive disorders and may of explosive rage, out-of-control anger, two brainstem auditory evoked poten-
lead to more effective treatment based out-of-control aggression, verbal or tials; and 20 minutes (after two min-
on neurophysiology and scientific physical attacks on another individual, utes of hyperventilation) of computer-
principles. intermittent explosive disorder, or ized electroencephalography. Digital
82 PSYCHIATRIC SERVICES ♦ January 2001 Vol. 52 No. 1
EEG data were evaluated, and arti- wave form at CZ and PZ; and the eyes- Table 1
fact-free data were used to create open FFT data recorded from occipital Two logistic regression models pre-
eyes-open and eyes-closed (resting) lobe electrodes, O1 and O2, the frontal dicting explosive behavior among 326
fast Fourier transformed files. We an- lobe electrodes, F3 and F4, two central children and adolescents
alyzed data on visual evoked potentials locations, CZ and PZ, and the temporal
and fast Fourier transformed files. lobes, T5 and T6 (16–18). Model and Odds
All clinical data except explosivity, variable p ratio
Visual evoked potentials sex, age, and medications were ex- Model 11
Data from the pattern-reversal visual cluded from the final analysis be- Delta F4 .001 3.354
evoked-potential evaluation were re- cause although the concordance de- Sex .183 .664
corded for each individual in accor- creased slightly when clinical data Age .823 .948
dance with guidelines of the American were removed, we decided that the Medications .597 1.177
Model 22
Electroencephalographic Society (14, small decrease in concordance was OP100 <.001 1.256
15). The checkerboard pattern-rever- less important than the inclusion of Delta F4 .007 2.931
sal paradigm used 19 millimeter, variables that are subjectively based Sex .022 .469
black-and-white alternating squares and that may vary dramatically be- Age .149 1.127
displayed on a model TC1115 RCA tween professionals. Medications .585 1.190
monitor positioned at eye level, 76 Logistic regression analysis was used 1 Model 1 included only the EEG data.
centimeters in front of the patient, to determine which electrophysiologi- –2LOG L=22.745, df=4, p<.001; concordant
subtending a visual angle of 23 de- cal variables were significantly associat- pairs=68.6 percent; discordant pairs=31 per-
grees. The pattern reversed every .59 ed with the presence of explosive be- cent
2 Model 2 included the EEG data and the
seconds for a total of 1.7 stimuli per haviors. Stepwise analyses were com- evoked-potentials variables. –2LOG L=
second. A 256 ms epoch was used with pleted on various combinations of EEG 44.518, df=5, p<.001; concordant pairs=75.8
a 5 ms prestimulus time. The flash par- data, data from the evoked-potential percent; discordant pairs=23.9 percent
adigm used a 512 ms epoch with 10 ms evaluations, and clinical information,
of prestimulus time. The intensity of using a cutoff probability value of .25 to
stimulus from the checkerboard pat- produce the preliminary model. variables, which are shown in Table 1.
tern-reversal was 12.69 candelas per Interactions between sex, age, med- They were the OP100 wave form—
square meter (cd/m2) and that of the ications, and the electrophysiological that is, the maximum P100 amplitude
flash 19.26 cd/m2. variables identified in the preliminary at the O1 and O2 electrodes—from
The patient was instructed to visual- model were assessed individually. All the pattern-reversal evoked-potential
ly fixate on a red dot centered on the interaction effects with p values of .15 studies; delta F4—that is, the delta
RCA monitor, not to speak, and to re- or less were added to the preliminary absolute power in the right frontal
main relaxed with as little movement model. We then sequentially assessed lobe from the EEG data; sex; age; and
as possible throughout the recording variables that had p values in excess of medications.
time. Artifacts were detected and re- .05, starting with the least significant Concordance in the models was de-
moved using the Bio-Logic online arti- variables. A confounding variable was termined as follows: All possible pair-
fact rejection program. For each pa- defined as one that changed the pa- ings of individuals who had explosive
tient, 200 artifact-free trials were aver- rameter estimate of a significant vari- behavior and who did not have explo-
aged together to produce the final able by 20 percent or more. sive behavior were created. A pair was
wave form. To gain a basic understanding of the defined as concordant if the individual
relationship of evoked potentials to un- who had explosive behavior was also
Procedure derlying brain electrical activity, corre- the individual predicted by the logistic
Patients’ clinical files were reviewed lations were determined between am- regression model to be the one more
for information about sex, age, medica- plitudes of the P100 wave forms in the likely to have explosive behavior on the
tions, head injuries, loss of conscious- pattern-reversal evoked-potential stud- basis of physiological predictor vari-
ness, explosive behaviors, narcissism, ies and the EEG data from the occipi- ables. A pair was defined as discordant
rumination, symptoms of bipolar disor- tal electrodes (O1 and O2). if the model incorrectly predicted that
der, or impulsivity. These clinical vari- the individual who did not have explo-
ables were evaluated for association Results sive behavior was more likely to be the
with several electrophysiological vari- Stepwise logistic regression analysis of individual who had explosive behavior.
ables. They included the maximum the EEG data alone and the EEG data Higher amplitudes of the OP100
amplitudes of the pattern-reversal plus the evoked-potential variables in- wave form were significantly associat-
evoked-potential N75 and P100 wave dicated that the concordance percent- ed with explosive behaviors, and the
forms at O1 and O2; the flash evoked age increased with the addition of amplitude was somewhat higher for
potential P100 wave form at O1 and evoked-potential data and that the ad- the female explosive individuals. As
O2 and P200 at F3 and F4; the audito- dition of these data significantly im- Table 2 shows, males were more likely
ry-even-related evoked potential P50 proved the analysis, as shown by the than females to exhibit explosive be-
wave form at F3 and F4, the N200 –2LOG L chi square statistic in Table 1. haviors, and the explosive group was
wave form at F3, F4, and CZ; the P300 The final model consisted of five slightly younger. Male explosive indi-
PSYCHIATRIC SERVICES ♦ January 2001 Vol. 52 No. 1 83
Table 2 analyses was performed in which the
Mean±SD ages and EEG values for 326 children and adolescents, by sex and by subjects taking medication at the time
whether or not they exhibited explosive behavior of testing were analyzed separately
from those who were not.
Age OP1001 Delta F42 A correlation analysis was used to
evaluate whether the underlying brain
Variable N Mean SD Mean SD Mean SD
electrophysiology might be the driving
Females 105 13.96 2.67 9.91 4.23 65.44 32.57 force of the amplitude of the P100
Males 221 12.93 2.98 9.58 5.07 82.70 46.40 wave form. An increase in the EEG
Explosive 267 13.03 2.95 10.31 4.96 81.67 44.52 absolute power was significantly relat-
Nonexplosive 59 14.32 2.52 6.80 2.60 56.64 28.68 ed (p<.001) to an increase in P100 am-
Explosive females 80 13.75 2.86 10.68 4.33 68.60 34.60
Nonexplosive females 25 14.64 1.85 7.46 2.73 55.33 22.72
plitude in the occipital lobe at the O1
Explosive males 187 12.72 2.95 10.15 5.20 87.27 47.13 and O2 electrodes (O1; delta, beta,
Nonexplosive males 34 14.89 2.93 6.48 2.46 57.60 32.69 and at O2; delta, theta, alpha, and
beta). However, the two were not
1 Maximum amplitude, in microvolts, between the O1 and O2 electrodes. The normal range is 5 to highly correlated; the highest correla-
8 µV.
2 Absolute power in microvolts squared. The normal range is 30 to 45 µV2. tion occurred between the delta activ-
ity and the P100 amplitude, recorded
at the O2 electrode; the correlation ac-
counted for only about 20 percent of
viduals were the youngest overall. In both regression models shown in the variance in the amplitude of the
However, these differences were not Table 1, the absolute power of delta F4 P100 wave form.
statistically significant. was significantly higher in the explo-
Explosive individuals also had high- sive group than in the nonexplosive Discussion
er delta absolute power at the F4 elec- group. In the model that included the The purpose of this study was to evalu-
trode; the most intense delta F4 activ- EEG data plus the evoked-potential ate whether EEG data predicted be-
ity was found in males. Using a 95 per- variables, the OP100 was significantly haviors exhibited by a group of children
cent reliability level, our clinical crite- correlated with explosive behavior. and adolescents referred to a psychi-
ria require a P100 of 10 microvolts or The medication variable was not sig- atric clinic for evaluation of a wide vari-
greater to indicate a predisposition to nificantly associated with explosive be- ety of psychiatric disorders, including
explosive behavior (Figure 1). havior in either model. A further set of explosive, aggressive, and out-of-con-
trol behaviors. These individuals were
referred by psychiatrists, psychologists,
social workers, educators, and parents.
Figure 1 They had been treated with little suc-
Absolute power of delta at the F4 electrode and OP100 values (maximum P100 cess, given various medications, and
amplitude at the O1 and O2 electrodes) for 326 children and adolescents, by given a profusion of diagnoses.
whether or not they exhibited explosive behaviors1 Our study showed that individuals
with a high-amplitude P100 wave form
300
in evoked-potential evaluations and
high delta band absolute power in the
250 right frontal lobe were more likely to
Not explosive
I II exhibit explosive behaviors.
Explosive
It was not surprising to find that
Delta F4 (µV )

200
2

most explosive individuals had high

frontal delta band absolute power
150 (quadrants I and II in Figure 1); most
experts in neurophysiology would ex-
pect such a result (Nuwer M, personal
100
IV communication, 1998). Convit and as-
sociates (19) found a positive relation-
50 ship between the level of violence and
III the amount of delta activity in psychi-
atric patients with violent behaviors.
0 Fishbein and colleagues (20) reported
0 5 10 15 20 25 30 35
that aggressive subjects in their study
OP100 (µV) had more delta and less alpha activity
1 The four quadrants are based on breakpoints of 135 µV2 of delta F4 and OP100 values of 10 µV.
in a spontaneous EEG, which they
The breakpoints are set such that 95 percent of the individuals in quadrants I and II exhibited ex- note has been observed in psy-
plosive behaviors. chopaths and criminals.
84 PSYCHIATRIC SERVICES ♦ January 2001 Vol. 52 No. 1
 What is new in this study is the find- trolled by the limbic system (23). If the the visual evoked potential. Another
ing that individuals with higher P100 intensity of the stimuli is excessive, study concluded that sodium valproate
wave form amplitudes in pattern-re- they may cause a kindling effect as has no effect on simple reaction time
versal evoked-potential studies were seen in seizure disorders. Andy and or the visual evoked potential in nor-
more likely to be explosive. Indeed, we Velamati (24) concluded that “repeat- mal subjects, although it may cause a
believe that the P100, which occurs edly recurring limbic system seizures slight increase in slow-wave sleep (30).
within the obligatory portion of the through superkindling mechanisms Among patients with seizure disor-
brain’s electrophysiological response can eventually render the limbic-basal ders, no pronounced influence of the
to sensory stimulation (approximately ganglia-preoptico-hypothalamic ag- disorder itself on the parameters and
within the first 200 milliseconds after gressive system hyper-responsive to wave form of the normal visual evoked
stimulus), is a biological signature and both recurring seizures and to extero- potential using pattern reversal was
represents an individual’s unique bio- ceptive stimuli with resulting aggres- demonstrated if the patients were not
logical receptive process underlying a sive behavior with or without an ac- taking anticonvulsant drugs; the find-
predisposition to process incoming companying seizure.” ings for patients under complete
stimuli in a given manner. These patients appear to respond seizure control with anticonvulsants
Although not all explosive individu- positively to various combinations of were similar (31).
als in our study had a high P100 wave anticonvulsant, antidepressant, an- The question arises as to whether
form (Figure 1), the results suggest tipsychotic, and stimulant medica- the increased amplitude of the P100
that the higher the amplitude, the more tions. Of these, the anticonvulsants ap- wave form can be explained by the
probable the explosive behaviors. pear to deliver the most consistent augmenting-reducing (A-R) phenom-
Overall, these patients constitute a dis- benefits (5,18,25–27). Because explo- enon, also known as stimulus intensity
tinct psychiatric population in need of sive behavior appears to be a problem modulation. The A-R refers to the
the unique medical treatment provid- with emotional expression or control, central nervous system’s modulation of
ed by the psychiatric community—not defects in limbic system function have responses to sensory stimuli of differ-
identification and incarceration as so- been proposed by Tancredi and ent intensities. In 1976 Buchsbaum
ciopaths by the justice system. Further Volkow (26) and by Matthews and as- (32) suggested that A-R is related to
studies are needed to confirm our re- sociates (5). The positive effect of anti- central control factors rather than pe-
sults with children and adolescents, to convulsants may be related to the large ripheral mechanisms. Several authors
investigate adult populations, to evalu- number of gamma-aminobutyric acid have suggested that this central mech-
ate the predictability of the behavior (GABA) receptors located in the anism is designed to protect the cortex
before it is actually present, and to amygdala, a critical component of the from overstimulation (32-36).
evaluate the relationship between the limbic system; several of the anticon- Blenner and Yingling (21) reported
wave form and various medication reg- vulsants are thought to exert their ef- that their visual evoked-potential data
imens. fects via the GABAergic systems (17). were in agreement with those of Sten-
We wish to emphasize that we do This study found that results for pa- berg and colleagues (37), who found a
not hypothesize that the high P100 tients who were taking medication at weak augmenting pattern at both FZ
amplitudes are the cause of explosive the time of the evaluation were not sig- and CZ but a reducing pattern at OZ.
behavior. Instead, as described above, nificantly different from the results of Overall they found in the visual system
it is best to conceive of high P100 am- those who were not. The effect of med- a slight augmentation at the vertex, but
plitudes as a marker for explosive be- ication on visual evoked potentials, a reduction at the occiput. In our study
havior. Descending inhibitory path- both checkerboard and flash, has been the effect of the A-R response in our
ways parallel the ascending sensory evaluated in several studies involving patients could not be evaluated be-
systems and modulate responsiveness normal subjects and individuals with cause the stimulus intensity remained
at a very early level of processing, ei- seizure disorders. The majority of stud- constant.
ther at the receptors themselves or ies found no effect on the amplitude of This study found a group of patients
soon after, and in any event before the the wave forms. Although the effect of exhibiting explosive behaviors who
cortical events responsible for the gen- valproic acid may be nonspecific, a had low-amplitude OP100 wave forms
eration of scalp evoked potentials (21). wide variety of anticonvulsant and in pattern-reversal evoked-potential
As a visual stimulus moves past the sedative agents have been reported to evaluations and low absolute-power
retina, it proceeds down two distinct lower the amplitude of flash evoked delta (Figure 1, quadrant IV). These
paths. The first allows it to be trans- potentials (28). Faught and Lee (29) patients represent other subsets of in-
ported to the primary receptive area in reported a trend toward higher ampli- dividuals with explosive behaviors of
the occipital lobes (Brodmann’s area tudes among photosensitive patients different etiologies, perhaps those
[17]). The hypothalamus also receives when pattern reversal was used as the with temporal lobe syndrome or
one-way afferent connections directly visual stimulus, but the differences frontal inhibition problems. This find-
from the retina (22). This is the path- were not statistically significant. ing is consistent with the literature in-
way of interest because the hypothala- No significant treatment effect was dicating that several different roads
mus participates in autonomic or be- found for simple reaction time, nor can lead to explosive, aggressive be-
havioral expressions of emotions as was there any significant effect on ei- haviors. Research in our laboratory is
part of the survival mechanism con- ther the latency or the amplitude of focusing on electrophysiologically de-
PSYCHIATRIC SERVICES ♦ January 2001 Vol. 52 No. 1 85
 with an electroencephalographic focus in ficity of evoked potential augmenting/re-
lineating these other groups. However, children without seizures. Neurology 11: ducing. Electroencephalography and Clini-
the importance of the amplitude of the 337–344, 1961 cal Neurophysiology 88:131–142, 1993
P100 wave form in classifying and pre-
7. Williams D: Neural factors related to habit- 22. Kupfermann I: Hypothalamus and limbic
dicting explosive behaviors should not ual aggression. Brain 92:503–520, 1969 system: peptidergic neurons, homeostasis,
be overlooked. and emotional behavior, in Principles of
8. Fields WS, Sweet WH (eds): Neural Bases Neural Science, 3rd ed. Edited by Kandel
The study had several limitations. of Violence and Aggression. St Louis, ER, Schwartz J, Jessell TM. New York, El-
First, the results are generalizable only Green, 1975 sevier, 1991
to individuals with psychiatric prob- 9. Mednick SA, Pollock V, Volavka JH, et al: 23. Laddies JE: Emotion, in Handbook of
lems severe enough for them to be Biology and violence, in Criminal Violence. Physiology: The Nervous System: Higher
identified by their parents, school sys- Edited by Wolfgang ME, Weiner NA. Bev- Cortical Functions of the Brain. Edited by
erly Hills, Calif, Sage, 1982 Plum F, Mount Castle V. Bethesda, Md,
tem personnel, or health care profes-
10. Scarpa A, Raine A: Psychophysiology of American Physiological Society, 1987
sionals. Second, evoked potentials
anger and violent behavior. Psychiatric 24. Andy OJ, Velamati S: Limbic system
were recorded from scalp electrodes, Clinics of North America 20:375–394, 1997 seizures and aggressive behavior (superkin-
and therefore the subcortical involve- dling effects). Pavlovian Journal of Biologi-
11. Pillmann F, Rohde A, Ullrich S, et al: Vio-
ment identified can only be hypothe- lence, criminal behavior, and the EEG: sig- cal Science 13:251–264, 1978
sized. Third, we did not evaluate all nificance of left hemispheric focal abnor- 25. Maletzky BM: Treatable violence. Medical
psychophysiological response systems malities. Journal of Neuropsychiatry and Times 100(10):74–79, 1972
Clinical Neuroscience 11:454–457, 1999
(for example, the electrodermal or car- 26. Tancredi LR, Volkow N: Neural substrates
diovascular) and their relationships to 12. Bars DF, Heyrend FL, Simpson CD: The of violent behavior: implications for law and
use of evoked potential studies in the iden- public policy. International Journal of Law
explosive behaviors. tification of explosive adolescents. Present- and Psychiatry 11:13–49, 1988
ed at the annual meeting of the American
Conclusions Educational Research Association, Chica- 27. Amen DG, Stubblefield M, Carmichael
go, Apr 1991 BA, et al: Brain SPECT findings and ag-
Our findings suggest that individuals gressiveness. Annals of Clinical Psychiatry
who exhibit explosive behaviors may 13. Bars DF, Heyrend FL, Simpson CD, et al: 8:129–137, 1996
The clinical use of computerized electroen-
have a predisposition for violent or ex- cephalography and evoked potentials in the 28. Newmark ME, Penry JK: Photosensitivity
plosive behavior that is an innate char- treatment of assaultive/aggressive adoles- and Epilepsy: A Review. New York, Raven,
acteristic of their central nervous sys- cents in a residential treatment facility. Pre- 1979
sented at the annual meeting of the Ameri-
tems. The use of noninvasive visual can Association of Children’s Residential 29. Faught E, Lee SI: Pattern-reversal visual
evoked potentials and quantitative Centers, Nashville, Tenn, Oct 19–23, 1994 evoked potentials in photosensitive epilep-
sy. Electroencephalography and Clinical
EEG data appears to allow accurate 14. American Electroencephalographic Soci- Neurophysiology 59:125–133, 1984
identification of this substantial sub- ety: American electroencephalographic so-
ciety guidelines in electroencephalography, 30. Harding GF, Alford CA, Powell TE: The
group of patients who have explosive effect of sodium valproate on sleep, reac-
evoked potentials, and polysomnography.
disorders that appear to be biologically Journal of Clinical Neurophysiology 11:1– tion times, and visual evoked potential in
based. Once we identify such patients, 147, 1994 normal subjects. Epilepsia 26:597–601,
1985
we can direct them into the most ap- 15. American Electroencephalographic Soci-
propriate treatment regimens. ♦ ety: Evoked Potentials. Bloomfield, Conn, 31. Martinovic Z, Ristanovic D, Dokic-Ris-
American Electroencephalographic Soci- tanovic D, et al: Pattern-reversal visual
ety, 1992 evoked potentials recorded in children with
Acknowledgment generalized epilepsy. Clinical Electroen-
16. Lesèvre N, Rèmond A: Selected applica- cephalographer 21:233–243, 1990
The authors thank Marc Nuwer, M.D.,
tions of a topographic approach to event-
Ph.D., for reviewing the manuscript. related potentials, in Topographic Mapping 32. Buchsbaum M: Self-regulation of stimulus
of Brain Electrical Activity. Edited by intensity: augmenting/reducing and the av-
References Duffy FH. Boston, Butterworths, 1986 erage evoked response, in Consciousness
and Self-regulation. Edited by Schwartz
1. Stevens JR, Sachdev K, Milstein V: Behav- 17. Borsook D: Effect of anticonvulsants on GE, Shapiro D. New York, Plenum, 1976
ior disorders of childhood and the elec- circadian rhythms: implications for major
troencephalogram. Archives of Neurology affective disorder and primary sleep disor- 33. Silverman I, Buchsbaum M, Henkin R:
18:160–177, 1968 ders, in Use of Anticonvulsants in Psychia- Stimulus sensitivity and stimulus intensity
try. Edited by McElroy SL, Pope HG. control. Perceptual and Motor Skills 28:71–
2. Grinfeld MJ: New killing fields: will the
Clifton, NJ, Oxford Health Care, 1988 78, 1969
campus shootings stop? Psychiatric Times,
July 1998, pp 1–3 18. Donovan J, Susser ES, Nunes EV, et al: Di- 34. Zuckerman M, Murtaugh T, Siegel J: Sen-
3. Cadoret RJ, Yates WR, Troughton E, et al: valproex treatment of disruptive adoles- sation seeking and cortical augmenting-re-
Genetic-environmental interaction in the cents: a report of 10 cases. Journal of Clini- ducing. Psychophysiology 11:535–542,
genesis of aggressivity and conduct disor- cal Psychiatry 58:12–15, 1997 1974
ders. Archives of General Psychiatry 52: 35. Knorring L: The experience of pain in de-
19. Convit A, Czobor P, Volavka J: Lateralized
916–924, 1995 pressed patients: a clinical and experimen-
abnormality in the EEG of persistently vio-
4. Elliott FA: Neuroanatomy and neurology of lent psychiatric inpatients. Biological Psy- tal study. Neuropsychobiology 1:155–165,
aggression. Psychiatric Annals 17:385–388, chiatry 30:363–370, 1991 1975
1987
20. Fishbein DH, Herning RI, Pickworth WB, 36. Lukas J, Siegel J: Cortical mechanisms that
5. Matthews D, Williamson B, Seals J, et al: et al: EEG and brainstem auditory evoked augment or reduce evoked potentials in
Treatment planning for violent juveniles. response potentials in adult male drug cats. Science 198:73–75, 1977
Presented at the annual meeting of the Na- abusers with self-reported histories of ag-
tional Association of Private Psychiatric gressive behavior. Biological Psychiatry 26: 37. Stenberg G, Rosén I, Risberg J: Personality
Hospitals, Fort Lauderdale, Fla, Jan 1993 595–611, 1989 and augmenting/reducing in visual and au-
ditory evoked potentials. Personal and Indi-
6. Green JB: Association of behavior disorder 21. Blenner JL, Yingling CD: Modality speci- vidual Differences 9:571–579, 1988

86 PSYCHIATRIC SERVICES ♦ January 2001 Vol. 52 No. 1