Presentation Details
| Choice Signals Emerge in Mouse Piriform Cortex During Delayed Olfactory Decision Making Srividya Pattisapu, Braden Brinkman, Alfredo Fontanini. Stony Brook University, Stony Brook, NY, USA |
Abstract
Piriform cortex (PC) multiplexes respiration-entrained activity with odor identity and intensity, but its role in encoding choice during odor-guided decisions remains debated. Prior work relied on go/no-go tasks where sensory evidence, choice, and movement are coupled. To dissociate these factors, we recorded respiration and PC spiking from large ensembles of neurons in mice performing a delayed 2-alternative forced-choice task. Stimuli were two pure odors and mixtures across four concentrations (+Limonene/-Limonene: 0/100, 30/70, 70/30, 100/0) and mice chose left for the first two and right otherwise. This task separates graded sensory signals tracking odor concentration from categorical signals tracking choice. To demix sensory (respiratory and odor-related) and choice signals, we fit generalized linear models incorporating respiration, odor features, and choice. PC neurons showed diverse coding, including units whose firing rate increased linearly with odor concentration and units with all-or-nothing categorical responses. Some neurons showed linear encoding early after odor onset and switched to categorical, choice-predictive coding later in the delay. To probe a mechanism for categorical activity, we trained a recurrent neural network with PC-inspired cell-types and connectivity on the same task. The model reproduced linear sensory and categorical choice encoding. Suppressing recurrent excitation impaired performance and reduced categorical activity, suggesting that recurrent circuitry promotes choice coding. We are testing this in vivo by expressing tetanus toxin light chain in PC pyramidal neurons while monitoring behavior and neural dynamics. Together, these results show that choice signals emerge in PC during delayed olfactory decisions and may be facilitated by local recurrent circuitry.
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.