Presentation Details
| Functional relevance of linear and categorical coding units for taste mixture-based decision-making Liam Lang1, 2, 4, Camelia Yuejiao Zheng1, 2, 3, 4, Jennifer M Blackwell1, 4, Giancarlo La Camera1, 2, 4, Alfredo Fontanini1, 2, 3, 4. 1Department of Neurobiology and Behavior, Stony Brook University, Stony Brook, NY, USA.2Graduate Program in Neuroscience, Stony Brook University, Stony Brook, NY, USA.3Medical Scientist Training Program, Stony Brook University, Stony Brook, NY, USA.4Center for Neural Circuit Dynamics, Stony Brook University, Stony Brook, NY, USA |
Abstract
Gustatory cortex (GC) produces time-varying activity at the population and single neuron levels that mediates food-related behaviors, including taste-based decision-making and taste discrimination learning. However, single unit functional contributions to population activity and behavior, and how such contributions evolve with learning, are underexplored. Here we address this question in mouse GC with a taste mixture-based decision-making task, in vivo high-density electrophysiology, and computational modeling. Mice were trained on a sucrose/NaCl mixture two-alternative choice task where the predominant mixture component cued, after a delay, the correct licking of a lateral spout. Population analyses showed GC collective activity reflected stimuli linearly during taste sampling and choices categorically before lateral lick decisions. Consistent with this, single unit tuning curve analyses revealed some neurons encoded sensory information linearly, and some encoded perceptual categories (sweet vs salty) or choices (left vs right) categorically. To probe the significance of these coding types, we built a recurrent neural network model that performed the task while reproducing the observed neural dynamics. Ablating coding units in the model showed each type, though a small fraction of the network, was required for normal dynamics and behavior, while the remaining ~60% of units were not. To assess how learning alters these units, we trained models further to match steepened psychometric functions, thus simulating neural changes underlying improving task performance. Results from this ongoing work will be presented. Our findings highlight the importance of neurons with specific response patterns in perceptual decision-making and will demonstrate how discrimination learning reshapes their contributions.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
Content Locked. Log into a registered attendee account to access this presentation.