The abundance of long-range excitatory projections to interneurons promotes synchrony in a cortical attractor network model
CBN (Computational Biology and Neurocomputing) seminars
Friday 17 October 2014
to 11:00 at
Pradeep Krishnamurthy (CB/CSC/KTH)
Memories and mental representations in the cortex are stored as distributed cell assemblies and it still remains unclear how diverse groups of neurons spanning several cortical regions could coordinate their activity. Synchrony is one of the suggested candidate to coordinate activity between different components of an assembly. Irrespective of the functional interpretation of neural synchrony, this dynamical phenomenon has been commonly reported in experimental studies. I will be presenting our findings on long-range synchrony in the framework of a mesoscopic cortical attractor network model. A recent anatomical study showing that the majority of connections found on pyramidal cells in primary visual cortex originate from neurons beyond the column serves as the basis for us to speculate a similar connectivity pattern for interneurons. We find that zero-lag synchrony is possible only when the fraction of long-range connections to interneurons outstrips its local connections. We also demonstrate that the emergent synchronous effects are robust to various parameter distortion, especially to axonal delay distribution and strength of synaptic connections. In addition, we examine whether the level of synchrony, dependent upon the fraction of local versus distant projections targeting interneurons, has any functional consequence in the context of stimulus induced pattern recall in our attractor network. To this end, we quantify the sensitivity of our network model to weak stimuli in terms of detection rate and latency of pattern recall. As a result, we report a facilitatory role of synchrony in pattern completion and competition task.