Virtual Brain Reading: A Connectionist Approach to Understanding fMRI

Abstract

We present a neurocomputational model of visual object processing, which takes photographic inputs and creates topographic stimulus representations on the hidden layer. We perform multi-voxel pattern analysis on the activations of hidden units and simulate contradictory findings from Haxby et al. (2001) and Spiridon and Kanwisher (2002) within a single model. With no special processing mechanism or architecture for faces in our model, we obtain the same results as Spiridon & Kanwisher, who interpreted their results as evidence for a “face module” – something our model does not possess.


Back to Thursday Papers