Back to U19 Main Page

brainCOGS - circuits of COGnitive Systems. Carlos Brody. Mechanisms of neural circuit dynamics in working memory and decision-making

Abstract 1:

Title: Multi-region calcium imaging during two decision making tasks

Abstract: I am using a two-photon mesoscope to record multi-region single-cell resolution calcium signals from 3 cortical regions while mice perform a decision making task in virtual reality. In this ‘Accumulating Towers task’ head-fixed mice are required to gradually accumulate visual evidence as they navigate in a virtual T-maze. The side on which the majority of the evidence appears informs them which maze arm the reward is located in. In an alternate version of the task (‘Visually Guided task’), mice navigate the same virtual T-maze and receive the same visual evidence cues but they do not have to accumulate evidence. Rather, they simply have to turn in the direction of a large visual guide. While mice perform these tasks, we record calcium signals simultaneously from the secondary motor area, retrosplenial cortex, and anterolateral visual cortex. This work provides a rare opportunity to explore how the neural underpinnings of decision making emerge on every level from single-cell responses to between region interactions.

Previous work from our lab comparing these tasks has shown that optogenetic inhibition of nearly any dorsocortical region impairs performance in the Accumulating Towers task whereas only inhibition of visual regions impairs performance in the Visually Guided task. Additionally, widefield calcium signals across cortical regions are less correlated in the Accumulating Towers task. Lower correlations were also observed during more difficult trials and more difficult task epochs. This seems to indicate that higher cognitive loads are supported by decreased neural correlations. A current objective is to replicate this last result with single-cell resolution data.

Questions that would benefit from theoretical modeling: If true, why are neuron-neuron correlations lower in the accumulating towers task? Is task-related information shared between brain regions (e.g. via a communication subspace)? How does neural activity coordinate between regions to produce behavior? In these tasks, an individual’s performance can vary from session to session and even between blocks of the same session. Can we identify performance-related neural correlates?

Back to U19 Main Page