Abstract
The representation of task-related information determines the speed of attentional selection and distractor suppression. However, it remains unclear how individually-unique neural representation of stimulus similarity affects attentional processes. We investigated this problem by examining the relationship between brain regions that encode unique categorical representations of face identity and performance in an independent visual search task. Functional magnetic resonance imaging (fMRI) data were acquired during a task in which face morphs between two famous face identities were classified as one or the other person. In addition, a house or a scrambled face image was sometimes presented and required the third "non-face" response. Representational similarity analysis (RSA) was used with a whole-brain searchlight procedure to identify regions with patterns of activation that correlated with identity classification (using the Spearman rank correlation coefficient). The results showed that patterns of activation in the prefrontal areas as well as in the face selective areas in the ventral temporal cortex were highly consistent with the classification of face identity, reflecting individually-unique relationships between behavioral and neural representational structures. Interestingly, performance in an independent visual search task in which one of the face identities served as the target and the various face morphs appeared as distractors was predicted by the behavioral and neural patterns during the classification task: reaction times to select the target increased as the classification similarity between the target and the distractor morph increased. Particularly, the neural representational structure in right DLPFC predicted individual differences as well as commonalities of the visual search performance. These results suggest that task-relevant internal representations reflect the categorical structure of target and distractor stimuli that is encoded in prefrontal areas as well as in stimulus-specific perceptual areas, and the individually-unique similarity structure of task stimuli can predict competition for attentional selection in an independent search task.
Meeting abstract presented at VSS 2016