The face inversion effect and the anatomical mapping from the visual field to the primary visual cortex
- Martha Gahl, CSE Department 0404, UCSD, La Jolla, California, United States
- Meilu Yuan, Computer Science, University of Illinois at Urbana-Champaign, Champaign-Urbana, Illinois, United States
- Arun Sugumar, CSE Department 0404, UCSD, La Jolla, California, United States
- Garrison Cottrell, CSE Department 0404, UCSD, La Jolla, California, United States
AbstractThe face-inversion effect, or the drastic decrease in accuracy seen when identifying inverted faces compared to upright faces, is not mimicked in object inversion. We introduce a new approach using computational models to explore the anatomical mapping of the visual field onto primary visual cortex. We propose that inverted faces are mapped onto the cortex in a fundamentally different way from upright faces. Our work first shows the validity of using this mapping with convolutional neural networks to complete recognition tasks by exploring scale invariance and realistically constrained rotation invariance. We confirm that a participant’s accuracy decrease when identifying inverted faces is not seen in the network with inverted objects. With the support of these findings, we test the face-inversion effect on our network and show the unique decline in accuracy, suggesting the mapping of the visual field onto the primary visual cortex is a key facet in this effect.