2019 Conference on Cognitive Computational Neuroscience

13-16 September 2019    Berlin, Germany

Technical Program

Paper Detail

Paper: PS-2A.40
Session: Poster Session 2A
Location: H Lichthof
Session Time: Sunday, September 15, 17:15 - 20:15
Presentation Time:Sunday, September 15, 17:15 - 20:15
Presentation: Poster
Publication: 2019 Conference on Cognitive Computational Neuroscience, 13-16 September 2019, Berlin, Germany
Paper Title: Neural signatures of coping with multiple tasks in mouse visual cortex
Manuscript:  Click here to view manuscript
License: Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
DOI: https://doi.org/10.32470/CCN.2019.1293-0
Authors: Márton Hajnal, Hungarian Academy of Sciences, Hungary; Duy Tran, Michael Einstein, University of California Los Angeles, United States; Gergő Orbán, Hungarian Academy of Sciences, Hungary; Peyman Golshani, University of California Los Angeles, United States; Pierre-Olivier Polack, Rutgers State University of New Jersey, United States
Abstract: Flexibly adapting to different task requirements is a key challenge of the visual system. In particular, depending on a potentially unobserved context the same stimuli might require different behavior. While task-related activity has been identified as early as the primary visual cortical (V1) activity in mice, it remained unknown if and how the visual cortex primarily dealing with feed-forward input contributes to efficient arbitration between tasks. Mice were trained to perform a multimodal task-switching paradigm where animals were required to make decisions either based on the identity of visual or that of auditory stimuli. Neural activity was recorded from all layers of V1 on 128 channels with extracellular electrodes. Our analysis has identified task-related variables in population responses. Importantly, while task-related variables could be identified mostly during stimulus presentation, the variable that could identify the specific task being performed could be reliably decoded from intertrial intervals, indicating a representation which is aware of the across trial contingency of task context. These results provide insights into how continual learning, the major challenge concerning the acquisition of multiple tasks relying on the same neural circuitry, can be achieved in biological agents.