Differential Stability of Task Variable Representations in Retrosplenial Cortex
Neuroscience ePoster by Luis Franco and Michael Goard
Part of COSYNE 2023 posters, California Institute of Technology; Biology and Biological Engineering; University of California, Santa Barbara
Sunday
Poster
Abstract
Cortical neurons store information across different timescales, from seconds to years. Although the stability of cortical representations is variable across regions, it can vary within a region as well. Association areas are known to multiplex behaviorally relevant variables, but it is not known whether all task variables are stored with similar representational stability. Here, we longitudinally measured the activity of neuronal populations in the retrosplenial cortex (RSC) using two-photon calcium imaging during the performance of a context-choice association task. We find that the activity of neurons tuned to three relevant task variables exhibits distinct levels of stability across days. Moreover, using linear classifiers trained on a reference day, we quantified the coding stability of these task variables. We find that RSC representations of environmental context and trial outcome display higher stability than motor choice, both at the single cell and population levels. Together, our findings show an important characteristic of association areas, where diverse information is stored with varying levels of stability, maintaining an adequate balance between stability and flexibility to subserve behavioral demands.
Topics
Details
- Session
- Poster Session III