A modified version of the attentional deployment task developed by Mac
Leod, Mathews and Tata (1986) [Journal of Abnormal Psychology, 95, 15-
20] was used to examine two issues: first, whether there was any evide
nce of attentional bias in depressed subjects, rather than in anxious
subjects alone; and second, whether attentional effects would occur in
the location of stimuli that could not be identified. Subjects were p
resented with pairs of words, one above the other, and the extent to w
hich attention favored threatening rather than neutral words was asses
sed from the latency to detect a dot in the same location of one them.
These detection latencies showed that depressed, but not anxious subj
ects, were selectively attentive to socially threatening words. There
was also evidence for attentional effects in the anxious subjects favo
ring physically threatening words. Furthermore, panic disorder patient
s were preferentially attentive to the location of physically-threaten
ing stimuli that could not be accurately identified. Overall, the resu
lts provide further evidence that emotionally disturbed subjects tend
to orient attention towards personally-relevant emotional stimuli. How
ever, the previous hypothesis that this attentional bias occurs only i
n anxiety, and not in depression, was not supported. Copyright (C) 199
6 Elsevier Science Ltd