The relationship between anxiety and memory biases for material that has been selectively processed in a prior task.

Despite evidence for a broad range of processing biases for threat-related material associated with anxiety, there is relatively little research which demonstrates such biases in memory in anxious individuals. The present paper reports two methodologies which demonstrate selective processing of threat-related stimuli in anxious subjects. These are a homophone spelling task and an anagram solution task. Both paradigms require subjects to somehow 'generate' the threat-related material. Despite such selective processing, subsequent recall and recognition memory tests for the same material revealed no evidence of a memory bias. The clinical and theoretical implications of the data are discussed.