Memory bias for anxiety information in patients with panic disorder

Abstract Based on Foa and Kozak's (1986) information processing theory of fear, we hypothesised that panic-disordered (PD) patients, in contrast to normal controls (NC), would exhibit a memory bias for anxiety-related information, and that physiological arousal would enhance this bias. PD subjects met DSM-III-R criteria for panic disorder. Self-report measures of mood, heart rate, and spontaneous skin conductance fluctuations were taken at baseline and at recall. After baseline, subjects rated the self-descriptiveness of anxiety (e.g. nervous) and nonanxiety (e.g. POLITE) words. Half of the subjects in each group then performed a 5-minute exercise step-task (i.e. high arousal condition); the remaining subjects relaxed for 5 minutes (i.e. low arousal condition). After this manipulation, subjects were asked to recall the rated words. Consistent with our prediction, PD subjects recalled more anxiety than nonanxiety words, whereas NC subjects recalled more nonanxiety than anxiety words. This memory bias was nonsignificantly (P < 0.11) enhanced in PD subjects in the high arousal condition. The results were not interpretable in terms of a response bias for reporting anxiety words in general. Nor was the rated self-descriptiveness of the words the principal determinant of recall.