English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In large part, rodents see the world through their whiskers, a powerfultactile sense enabled by a series of brain areas that form thewhisker-trigeminal system. Raw sensory data arrives in the form of mechanicalinput to the exquisitely sensitive, actively-controllable whisker array, and isprocessed through a sequence of neural circuits, eventually arriving incortical regions that communicate with decision-making and memory areas.Although a long history of experimental studies has characterized many aspectsof these processing stages, the computational operations of thewhisker-trigeminal system remain largely unknown. In the present work, we takea goal-driven deep neural network (DNN) approach to modeling thesecomputations. First, we construct a biophysically-realistic model of the ratwhisker array. We then generate a large dataset of whisker sweeps across a widevariety of 3D objects in highly-varying poses, angles, and speeds. Next, wetrain DNNs from several distinct architectural families to solve a shaperecognition task in this dataset. Each architectural family represents astructurally-distinct hypothesis for processing in the whisker-trigeminalsystem, corresponding to different ways in which spatial and temporalinformation can be integrated. We find that most networks perform poorly on thechallenging shape recognition task, but that specific architectures fromseveral families can achieve reasonable performance levels. Finally, we showthat Representational Dissimilarity Matrices (RDMs), a tool for comparingpopulation codes between neural systems, can separate these higher-performingnetworks with data of a type that could plausibly be collected in aneurophysiological or imaging experiment. Our results are a proof-of-conceptthat goal-driven DNN networks of the whisker-trigeminal system are potentiallywithin reach.

Toward Goal-Driven Neural Network Models for the Rodent Whisker-Trigeminal System