| Paper: | PS-2B.36 |
| Session: | Poster Session 2B |
| Location: | Symphony/Overture |
| Session Time: | Friday, September 7, 19:30 - 21:30 |
| Presentation Time: | Friday, September 7, 19:30 - 21:30 |
| Presentation: |
Poster
|
| Publication: |
2018 Conference on Cognitive Computational Neuroscience, 5-8 September 2018, Philadelphia, Pennsylvania |
| Paper Title: |
Thalamic Modulation of Memory in Recurrent Networks |
| Manuscript: |
Click here to view manuscript |
| DOI: |
https://doi.org/10.32470/CCN.2018.1212-0 |
| Authors: |
Peter Stratton, The University of Queensland, Australia; Michael Halassa, Massachusetts Institute of Technology, United States |
| Abstract: |
During delay tasks, some neurons in the murine thalamocortical system (Schmitt et al., 2017) and hippocampus (‘time cells’) (MacDonald et al., 2011) display transient spike responses with timing that is repeated reliably across trials. In higher mammals and primates, activity in some cells is consistently elevated. These transient responses during delays confer a short term memory of the stimulus. We wondered what neural network structures could facilitate the generation of such dynamic memory patterns. We show that in a simplified formalism of a dynamic recurrently-connected network (DRN), the number of unique dynamic patterns grows exponentially with network size. The DRN formalism emphasises the role in neural function of transient yet repeatable dynamics. Unlike reservoir networks, the connectivity matrix does not need to be finely tuned (random connectivity suffices), and the dynamics implement indefinite (not fading) memory. Gating of input patterns is assumed to be controlled by modulatory signals from the thalamus. In particular, recent experimental evidence suggests that inputs from the MD thalamus convey contextual information and can modulate cortical synaptic strengths. We show in a spiking neural network model that MD modulation of synaptic strength can indeed stabilize dynamic patterns of activity and hence short term memories. |
