Anatomy is Symmetry's Best Friend: Re ections on Modeling Baylis and Driver

An aptitude for the detection of bilateral symmetry is a fairly prominent aspect of the human visual system. Knowledge of the reasons behind this facility is not so well established, however. Some of the behavioral data indicates that processing of symmetric and non-symmetric stimuli is undertaken in two wholly di erent manners (i.e. serial versus parallel). However, the interpretation of this as being due to high level cognitive preferences does not exhaust the list of possible explanations. Using a split-neural network model, we show that instead of cognitive preferences, gross morphological factors may play a large role in underwriting the ability to detect symmetry as a special case of shape perception. The earlier model is consistent with behavioral data, but Occam's razor suggests that we might prefer the newer morphological explanation.