Implicit and Explicit Memory for Threat Words in High- and Low-Anxiety-Sensitive Participants

This study examined memory for anxiety andthreat words in high-anxiety-sensitive (HAS; n = 38) andlow-anxiety-sensitive (LAS; n = 36) participants. Basedon Foa and Kozak's (1986) information processing theory of fear, it was hypothesized that HASparticipants would remember anxiety and threat-relatedinformation better than LAS participants and thatphysiological arousal would enhance this difference. Support for the first hypothesis was limited.Anxiety sensitivity did affect explicit memory but onlyfor general threat words, not for anxiety words. HASparticipants recalled significantly more threat words than neutral, positive, or anxiety wordswhereas LAS participants recalled significantly morepositive words than neutral or anxiety words. HASparticipants also recalled significantly more threat words than LAS participants. Thus, HASparticipants exhibited an explicit memory bias forthreat words, but, unexpectedly, not for anxiety words.No implicit memory differences were found. Inconsistent with the second hypothesis, the arousalmanipulation (i.e., hyperventilation) had no effect.Because anxiety sensitivity is a cognitive risk factorfor panic disorder, these results suggest that memory biases for threat may exist in patients priorto the onset of the disorder, perhaps even before thefirst unexpected panic attack.