Extraversion/Sociability 89alpha or beta bands but were found only in the most relaxed
condition suggests that the weak correlation may have been
due to impulsive extraverts’ getting drowsy or actually falling
asleep. Regardless of interpretation, the low level of relation-
ship between personality and arousal in this study could ex-
plain the inconsistency of previous studies testing the
hypothesis: They simply did not have enough power to detect
the relationship with any reliability.
Consistent with Eysenck’s model was the finding that
while performing six tasks extraverts tended to perform
worse than introverts at higher levels of alpha (indicating
lower levels of arousal). Only the alpha band, however, sup-
ported the hypothesis of better performance of introverts
at lower levels of arousal. Brain imaging using positron-
emission tomography (PET) and cerebral blood flow (CBF)
have an advantage over EEG because they assess subcortical
as well as cortical activation and analyze activity in particu-
lar structures or brain loci. The problem with studies using
these new techniques is that because of the expense, low num-
bers of subjects are used and many brain areas are analyzed,
increasing the possibilities of both Type I and Type II errors.
Replication across studies is one solution to the problem.
Mathew, Weinman, and Barr (1984) found negative corre-
lations between E and CBF indices of activation in all corti-
cal areas in both hemispheres, supporting Eysenck’s
hypothesis of higher cortical arousal in introverts than in ex-
traverts. All of their participants were female. Stenberg,
Wendt, and Risberg (1993) also found an overall negative
correlation(r=−.37), but this was a function of the high
correlation among the female participants; the correlation
among the males was close to zero. As with the EEG data,
confirmation of the hypothesis was more common in female
than in male samples.
Some studies have found hemispheric differences in the
relationships between E and activation, but these have not
been consistent (Johnson et al., 1999; Stenberg et al., 1993).
Studies of subcortical areas of brain have also yielded little in
the way of consistent findings except for one: E is associated
positively with activation of the anterior cingulate area
(Ebmeier et al., 1994; Haier, Sokolski, Katz, & Buchsbaum,
1987; Johnson et al., 1999). The cingulum is the major path-
way between the frontal cortex and the limbic system and has
been theoretically associated with neuroticism and anxiety
rather than E (Zuckerman, 1991).
The results in the two brain imaging studies described, un-
like the EEG studies, tend to support Eysenck’s hypothesis of
a relationship (albeit a weak one) between E and cortical
arousal. There is no clue in his theory, however, why the find-
ing is supported more in females than in males or why sub-
cortical differences in the cingulum, the executive structure
of the limbic brain, should be associated with extraversion. In
Eysenck’s model limbic arousability is associated with neu-
roticism, and any association with E would be with introver-
sion rather than extraversion.
General arousal may be too broad a construct to be associ-
ated with personality. Arousal is highly dependent on diurnal
variation and general stimulation levels. Arousal as a trait
would represent the state of the nervous system at a given
time under a given set of conditions. In contrast, arousability
is the typical immediate reaction of some part of the nervous
system to a stimulus with specified characteristics. Eysenck’s
(1967) optimal level of stimulation model says that introverts
are more arousable at low to moderate intensities of stimula-
tion, but at higher intensities extraverts are more responsive.
Introverts have strong reactive inhibition mechanisms that
dampen response to high intensities. Strelau (1987), in a
model based on neo-Pavlovian theories, states that persons
with strong nervous systems are relatively insensitive to
stimuli at lower intensities but can process and react to stim-
uli at higher intensities. For weak nervous system types the
opposite is true: They are highly sensitive to low intensities
but show inhibition of response at high intensities.Cortical ArousabilityCortical arousability is usually assessed with the cortical
evoked potential (EP). A brief stimulus, such as a tone or
flash of light, is presented a number of times, and the EEG is
digitized at a fixed rate, that is time locked to stimulus deliv-
ery time and averaged across trials for a given participant.
This process averages out the “noise” and produces a clear
waveform representing the typical reaction of that subject to
the specific stimulus over a 500-ms period. Although laten-
cies of response vary somewhat for individuals, for most one
can identify particular peaks of positivity and negativity. For
instance, a peak of positive potential at about 100 ms after the
stimulus (P1) represents the first impact of the intensity char-
acteristics of stimuli on the cortical centers. Earlier peaks
represent stimulus processing at subcortical centers. The peak
at 300 ms after the stimulus (P3) is influenced by novelty, sur-
prise, or unexpectedness of the stimulus and thus represents a
higher level of cortical processing in that the stimulus must be
compared with previous stimuli.
Stelmack (1990) reviewed the relationship between E and
cortical EPs. As might be expected, the results depend on the
characteristics of the stimuli used to evoke the EPs as well as
the reactor’s age and personality characteristics. For instance,
Stelmack said that introverts have greater amplitude EPs in
response to low-frequency tones, but there are no differences
between introverts and extraverts for high-frequency tones.