Neuron Network Simulation of Visual Working Memory
From Montana Tech High Performance Computing
MSU Cell Biology and Neuroscience
The project consists of developing a neuronal network model to investigate potential neurophysiological mechanisms that will emulate the encoding of short-term memories into coherent oscillations of neuronal populations. The goal is to simulate coherent oscillations between distant populations of neurons. The long-range projections induced a reciprocal inhibition of the local beta oscillations, thereby preventing any coherence between the columns. After estimating the numerous potential causes for this effect, it was decided to shift our goal to a shorter time‐frame plan that exploits the behavior of a single column. An increase in the power of oscillations during the memory delay is induced by spike‐time‐dependent plasticity of synapses (STDP) during the memory delay, and decided to use our single‐column network to test hypothesis. Simulations were run with and without STDP to compare the effects on local oscillations. Using the HPC cluster, it was found that the time necessary for synapses to return to baseline conductance is the most relevant parameter in determining the power of oscillations in the beta frequency range. Increasing this parameter leads to a decrease in the beta power. More precisely, when the return to pre-STDP conductance is slow, the effects of STDP on beta frequency power are minimized, although still negative.