Abstract
First-order tactile neurons have spatially complex receptive fields. Here we use machine learning tools to show that such complexity is normatively good, arising for a wide range of training sets and network architectures, and benefiting network performance on discrimination tasks especially in the presence of noise. We suggest that complex receptive fields reflect the role of first-order tactile neurons as the input elements of a compressed sensing scheme.
Copyright
The copyright holder for this preprint is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. All rights reserved. No reuse allowed without permission.