64 COGNITIVE THEORY AND RESEARCH ON ANXIETY
match the current concerns or worries of the patient, and when threat intensity is moder-
ate to severe. In addition facilitated attention to threat may be enhanced by an impaired
disengagement from highly threatening stimuli in anxious individuals (e.g., Koster et
al., 2006). Attentional avoidance of threat clearly plays an important role in defining
perceptual bias in anxiety but it may be less prominent in high trait anxiety (Mogg et
al., 2004). Finally, attentional bias is probably not unique to anxiety, with depression,
for example, characterized by an attentional bias for negative information (e.g., Gotlib,
Krasnoperova, et al., 2004; Mathews et al., 1996).
Stimulus Identification Tasks
Stimulus identification paradigms involve a search for threatening or nonthreatening
words within a matrix of random words or measurement of latency to identify words
presented at participants’ threshold of awareness. In a number of studies panic patients
had enhanced identification of threat stimuli (Lundh et al., 1999; Pauli et al., 1997;
see Lim & Kim, 2005, for negative findings) and social phobia individuals had facili-
tated identification of angry faces (Gilboa- Schechtman, Foa, & Amir, 1999). However,
studies of generalized anxiety have been more complicated, with some showing facili-
tated detection of threat (Mathews & MacLeod, 1986; Foa & McNally, 1986) and
others indicating that the problem might be increased distraction by threatening stimuli
(Mathews, May, Mogg, & Eysenck, 1990; Rinck, Becker, Kellerman, & Roth, 2003).
Summary
There is strong empirical support for the first hypothesis of the cognitive model. Despite
some inconsistencies across studies, there is still substantial evidence from a variety
of experimental methodologies that anxiety is characterized by a hypervigilance for
threatening stimuli and that this attentional bias is absent in low anxiety states. How-
ever, it is also clear that a number of qualifications must be added to this statement.
Attentional threat bias is more evident in the immediate or early stages of processing
when conscious awareness is reduced, when threat stimuli match the specific anxiety-
relevant concerns of the individual, and when threat intensity has reached a moderate to
high level. Figure 3.2 provides a schematic illustration of how exposure duration, mean-
ing, and threat value determine the role of selective attentional processing for threat in
anxiety (see Mogg & Bradley, 1998, 2004, for further elaboration).
Hypervigilance for threat will be absent when mildly threatening and impersonal
stimuli (e.g., general threat words) are presented at long exposure intervals. At the other
extreme, all individuals will exhibit heighten vigilance when stimuli are extremely
threatening, highly personal, and preconscious or automatic. That is, anyone will attend
to stimuli evaluated as posing a significant threat. However, it is the moderately threat-
ening, personally specific stimuli presented at brief, preconscious exposure intervals
that will result in the exaggerated attentional threat bias that characterizes the anxi-
ety disorders. Moderately threatening stimuli are considered threatening by vulnerable
individuals but nonthreatening to those with low anxiety (Mogg & Bradley, 1998).
However, selective attention to threat (i.e., facilitation effects) must be understood as
an interplay with avoidant (i.e., inhibitory) attentional processes, which in turn depends