Australasian Mathematical Psychology Conference 2019

The effect of feature separation on processing architecture and implications for models of visual attention

Sarah Moneer
Melbourne School of Psychological Sciences, University of Melbourne
Daniel Little
Melbourne School of Psychological Sciences, University of Melbourne

Changing the physical separation between features that are perceptually separable affects processing architecture in a categorization task. When relevant features were presented as parts of the same object but separated by a large distance, the features were processed serially. When the relevant features overlapped in space, however, we observed evidence of parallel processing. This raised the question of whether features must necessarily be overlapped in order to be processed in parallel. We therefore manipulated the separation between a pair of visual features (saturation and orientation) to investigate how this affects processing architecture. Comparison of logical rule models and Systems Factorial Technology analyses were used to determine processing strategies. We found evidence of parallel processing at all the tested separation values in the absence of distractors, and some evidence to suggest that bilateral presentation of the features may facilitate parallel processing. Implications of these findings for models of visual attention will be discussed.