Aggression/Hostility/Anger/Agreeableness 103excess of slow wave (theta) activity. Volavka (1995) pointed
out that these results could be due to a variety of factors in-
cluding developmental retardation, brain injuries, decreased
arousal level, cortical disinhibition, or genetic factors. Actu-
ally, twin research suggests that most of the activity in spec-
trum parameters of the EEG is genetically determined
(Lykken, 1982).
One limitation of most of these earlier studies was that
only prisoners referred for neuropsychiatric evaluation were
used. A study by Wong, Lumsden, Fenton, and Fenwick
(1994) selected subjects from a population of prisoners who
had all been rated for violent behavior, and 70% had received
EEG assessment. The prisoners were divided into three
groups based on their history of violence. Going from the
lowest to the highest violent groups, the percentages of ab-
normal EEG’s were, respectively, 26%, 24%, and 43%. The
most frequent EEG characteristics differentiating the most
violent from the less violent groups was focalized EEG ab-
normalities, particularly in the temporal lobes. Twenty per-
cent of the most violent patients showed abnormal temporal
lobe readings compared to 2% to 3% in the other two groups.
Computerized tomography (CT) scans confirmed the high in-
cidence of temporal lobe abnormalities in the most violent
group.
The cortical EP has also been used to study cortical arous-
ability. A study comparing detoxified alcoholic patients with
and without histories of aggression found lower amplitudes
of the P300 in the aggressive group (Branchey, Buydens-
Branchey, & Lieber, 1988). Aggressive alcoholics often have
other characteristics, such as impulsivity and alcoholism,
which might have produced the weaker P300 signal. Another
study found that impulsive aggressive subjects screened from
a college student population also showed lower P300 ampli-
tudes at frontal electrode sites (Gerstle, Mathias, & Stanford,
1998). Still another study showed that a drug that reduced
frequency of aggressive acts among prisoners with a history
of impulsive aggression also increased the amplitude of the
P300 in this group (Barratt, Stanford, Felthous, & Kent,
1997). This effect of the drug was not found in a group of
prisoners who had committed premeditated murders. A re-
duced P300, particularly in the frontal lobes, may be sympto-
matic of a weakened inhibition from the frontal lobes and
may account for the impulsive aspect of the aggression.
Visual imaging methods have been used in the study of vi-
olent behavior. Two structural methods are computed tomog-
raphy (CT) and magnetic resonance imaging (MRI). MRI
yields better images for precise assessment of brain structure.
PET is used to assess brain activity in specific areas of brain
including regions not accessible by ordinary EEG methods.
Mills and Raine (1994) reviewed 15 studies of structural
brain imaging (MRI, CT) and 5 studies using PET and re-
gional CBF. Subject groups were violent prisoners, convicted
murderers, pedophiles, incest offenders, property offenders,
and, in some studies, normal controls. Property offenders
were regarded as controls for violent offenders. Sexual of-
fenders were not necessarily violent. Nine of the 15 studies
using CT or MRI showed some type of structural abnormal-
ity, about evenly divided between frontal and temporal or
frontotemporal deficits. Frontal abnormalities characterized
the violent offenders and frontotemporal the sexual offend-
ers, according to the authors of the review. However, most
studies used small samples. The two studies of violent offend-
ers using large samples (Ns of 128 and 148) found no particu-
lar localization of abnormalities (Elliot, 1982; Merikangus,
1981). The only study using MRI with any kind ofN(another
had only 2 cases) found evidence of temporal lobe lesions in
5 of 14 violent patients (Tonkonogy, 1990). The large study by
Wong et al. (1994), not included in the review, found that 55%
of the most violent group had abnormal CT findings, and 75%
of these were temporal lobe findings. Contrary to the hypoth-
esis of Mills and Raine, temporal lobe lesions alone seem to
be characteristic of violent patients. More MRI studies are
needed to clarify the issue of localization.
The temporal lobe overlays the amygdala and has connec-
tions with it. Animal lesion and stimulation studies have
found sites in the amygdala that inhibit and others that trigger
aggression. Total amygdalectomies in monkeys produce a
drop in the dominance hierarchy and an inability to defend
against aggression from other monkeys. The comparative
data suggest loci for aggression in the amygdala.
Mills and Raine reviewed five PET studies, but of these
only one had a near-adequate number of subjects (3 had less
than 10) and another compared child molesters with controls.
The one study remaining compared 22 murderers with 22
normals and found selective prefrontal dysfunction in the
group of murderers (Raine et al., 1994). Temporal lobe dam-
age and functional hypofrontality are not unique to violent
offenders but are also found in patients with schizophrenia.Cardiovascular Arousal and ArousabilityNumerous studies show that persons who score high on
hostility scales show greater anger and cardiovascular
arousal, especially blood pressure, in response to stress or
perceived attack than do low hostile persons. As an example,
a recent study found that among participants who were delib-
erately harassed in an experiment, the high hostile group who
was harassed showed enhanced and prolonged blood pres-
sure, heart rate, forearm blood flow and vascular resistance,
and increased norepinephrine, testosterone, and cortisol