Neuroticism/Anxiety/Harm Avoidance 95on reactivity of the limbic system, which regulates the ANS,
but he did not distinguish particular pathways, structures, or
neurotransmitters within that system that were involved in N.
Some theories did not even make a distinction between corti-
cal and autonomic arousal in emotions. Eysenck felt that
there was some correlation between the two kinds of arousal
because of collaterals between the limbic and ascending
reticulocortical system. Gray (1982) and others, extrapolat-
ing from experimental studies of animals, delineated specific
limbic systems involved in anxiety and the neurotransmitters
involved in these systems. Neuroimaging studies have at-
tempted to extend these brain models to humans.
Autonomic Arousal
Large-scale studies of the relationship between cardiovascu-
lar measures, either in resting levels of activity or reactivity
to stressful experimental situations, and Measures of N failed
to reveal any significant relationships (Fahrenberg, 1987;
Myrtek, 1984). On the assumption that high cardiovascular
activity put high-N subjects at risk for cardiovascular disease,
Almada et al. (1991) investigated the relation between mea-
sures of N and subsequent health history in nearly 2,000 men.
N was not associated with systolic BP or serum cholesterol
but was associated with cigarette smoking and alcohol con-
sumption. When tobacco and alcohol consumption were held
constant there was no relationship between N and cardiovas-
cular disease. Similar studies have failed to find any relation-
ships between electrodermal activity and N or trait anxiety
(Fahrenberg, 1987; Hodges, 1976; Naveteur & Baque, 1987).
Given the fact that many anxiety disorders do show ele-
vated heart rate and electrodermal reactivity, how can we
explain the lack of correlation with N? The answer may lie in
the difference between generalized anxiety disorder (GAD)
and panic disorder (PD), agoraphobia (Ag), and obsessive-
compulsive disorder (OCD). Whereas the latter (PD, Ag,
OCD) show elevated basal HRs and frequent spontaneous
SCRs, GAD patients show little evidence of this kind of
autonomic arousal (Zuckerman, 1991). Their anxiety is
expressed cognitively (worry) and in symptoms of muscle
tension such as fatigue. In contrast, PD, Ag, and OCD pa-
tients complain of autonomic symptoms, such as accelerated
heart rate, even when they are not experiencing an actual
panic attack (Zuckerman, 1999). Most persons who are high
on N probably represent subclinical GAD disorder rather
than the other types of anxiety disorders.
Brain Arousal
Studies of general cortical arousal using the EEG have
historically focused on E, but some of these studies found
interactions with N. These effects were inconsistent; some
found higher and some reported lower arousal for high-N
persons. Application of PET methods has not shown any as-
sociation of general cortical or limbic arousal with N in situ-
ations that were not emotionally provoking (Fischer et al.,
1997; Haier et al., 1987). Similar results are seen in anxiety
patients; but when anxiety is provoked in patients by present-
ing them with feared stimuli, increased activity is seen in
areas like the orbitofrontal cortex, insular cortex, temporal
cortex, and anterior cingulate (Breier et al., 1992; Rauch et
al., 1995). These studies identify an anxiety pathway in hu-
mans (orbitofrontal-frontal to cingulate to temporal lobe and
amygdala) already established in animals, but they do not
show a preexisting sensitivity of this pathway in normals
scoring high in N. Another study of anxiety patients in non-
stimulated conditions, which did use normal controls, found
that whole brain blood flow did not distinguish anxiety
patients from normals but did find a negative correlation be-
tween a depression scale and caudate activation. The previ-
ously mentioned study by Canli et al. (2001) found that in a
small sample of normal women N correlated with increased
brain activation to negative pictures (relative to activation by
positive pictures) in left-middle frontal and temporal gyri and
reduced activation in the right-middle frontal gyrus. Taken
together, the clinical studies and this last study of normals
suggests that whole brain activation does not vary with N-
Anx, but given negative emotional provocation there may be
a reactive disposition in frontal cortex of high-N persons that
activates a pathway through the orbitofrontal cortex around
the cingulum to the temporal lobe and amygdala.
Davis (1986) argued that the central nucleus of the amyg-
dala is a major center where the input of fear-provoking stim-
uli is organized and where output to various intermediate
nuclei organizes the entire range of behavioral, autonomic,
and neurotransmitter reactions involved in panic or fear. A
recent MRI study (van Elst, Woermann, Lemieux, & Trimble,
1999) found an enlargement of left and right amygdala vol-
umes in epileptic patients with dysthymia (a chronic kind of
neurotic depression). Amygdala volume within the group did
not correlate with trait or state anxiety but did correlate posi-
tively with a depression inventory. Because anxiety and
depression are usually highly correlated and both correlate
highly with N, it is not clear why depression alone was re-
lated to amygdala volume.MonoaminesMuch of the recent exploration of the role of the monoamines
in N-Anx have been based on Cloninger’s (1987) biosocial
model of personality and therefore used his scale of Harm