CNS*2026 Halifax

Introduction
The neocortex not only has the ability to represent stimuli, but it also needs to be able to categorize them for fast and efficient processing. Research has shown discrete representation in the primary auditory cortex when presented with ambiguous stimuli [1]. We hypothesize that this mutually exclusive dynamic is possible through competitive interaction between different neuronal assemblies representing the stimuli, mediated by inhibition to opposing assemblies via Martinotti cells (MC).


Methods
This study employs an existing, experimentally-valided, large-scale biophysical model of the non barrel primary somatosensory cortex (nbS1) of juvenile rates [2,3]. This level of detail allows for a manipulation of the connectome to mimic different hypotheses for how learning could affect the connectivity between different neuronal populations. Based on a previous method, the circuit presented with “pure” patterns to identify assemblies then ambiguous patterns generated from interpolation [4]. Different modifications are done to the circuit like the removal of connections between different populations of neurons, allowing for a study into how these modifications change MC’s ability to inhibit different excitatory populations.


Results
When presented with the interpolated patterns, the unmodified and naïve circuit followed it while displaying a transitional representation. However, the modified circuit with changes to connection of MC also exhibited the same behavior. This is unexpected behavior which prompted further experiments to see how competitive dynamics can be achieved in this circuit.


Discussion
Our results suggest further experimentation and a possible revision of our hypothesis. We would like to further analyze the role of top-down projections, VIP+ neurons, and hypothetical changes to synapses due to plasticity in process of learning to categorize.


References

1. Bathellier, B., Ushakova, L., & Rumpel, S. (2012). Discrete neocortical dynamics predict behavioral categorization of sounds. Neuron, 76(2), 435–449. https://doi.org/10.1016/j.neuron.2012.07.00
   
2. Reimann, M. W., … Ramaswamy, S. (2026). Modeling and simulation of neocortical micro- and mesocircuitry (Part I, anatomy). eLife, 13, RP99688. https://doi.org/10.7554/eLife.99688
   
3. Isbister, J. B., … Reimann, M. W. (2026). Modeling and simulation of neocortical micro- and mesocircuitry (Part II, Physiology and experimentation). eLife, 13, RP99693. https://doi.org/10.7554/eLife.99693
   
4. Ecker, A., ..., Reimann, M. W. (2024). Cortical cell assemblies and their underlying connectivity: An in silico study. PLoS Computational Biology, 20(3), e1011891. https://doi.org/10.1371/journal.pcbi.1011891
   

Acknowledgement
This research project is supported by funding from the Fondation Courtois, NSERC, IVADO, the CHU Sainte-Justine Research Center,  FRQS, the Canada CIFAR AI Chairs Program, Mila, and Google. Their compute infrastructure was supported through a grant from the Canada Foundation for Innovation (John Evans Leader Fund), and a grant of computing time awarded from the Digital Research Alliance of Canada.