Attention Versus Memory Biases

Biases discovered on the lexical decision and valence identification tasks have, thus far, been proposed to reflect biases in attention. A number of theorists (e.g., MacLeod & Matthews, 1991) suggest that depressive information processing biases occur at a ``later'' stage of processing (e.g., memory recall) rather than at an early stage, as would be characteristic of biased attention. This phenomenon may be theoretically juxtaposed to biases found at an earlier stage of processing, e.g., in people who are anxious with respect to threatening stimuli. Importantly, the explanations for ``attention biases'' advanced thus far have relied on both differential attention to negative information, and differential association with negative information in memory; these processes are proposed to interact. The hypothesized ruminative nature of recurrent association with the negative aspects of a stimulus in memory may be thought of as a ``memory bias'', but ultimately, because these biases affect how information which is present in the environment is processed they may also be considered attentional biases. That is, rumination is proposed to affect which elements of the environment are selectively focused upon, and thus, may be considered as affecting attention, as defined above. A computational model of information processing in depression is considered later in the thesis and allows comparison of different amounts of rumination vs. nonruminative attentional biases occurring in the proposed cognitive-affective feedback loop. This technique begins to allow determination of the relative strength of attentional vs. memory factors in creating biases found on the lexical decision and valence identification tasks.