Weak Perturbations in Large-Scale Cortical Network Models

Abstract

Event Abstract Back to Event Weak Perturbations in Large-Scale Cortical Network Models Mohsin Ali1 and Flavio Frohlich2* 1 University of North Carolina, United States 2 University of North Carolina - Chapel Hill, Psychiatry, Cell and Molecular Physiology, Biomedical Engineering, Neuroscience Center, United States The importance of the occurrence and timing of a single action potential in cortex has remained unresolved. Given the number of neurons and the weak and unreliable nature of cortical synapses, it has been argued that individual action potentials carry little significance. In contrast, recent experimental and modeling studies suggest that individual action potentials have a measurable effect on the subsequent evolution of the overall network dynamics [1-3]. It is unknown under what circumstances cortical networks exhibit such sensitivity to weak perturbations. In particular, it has remained unclear how the sensitivity of cortical networks to very small perturbations depends on network connectivity and overall network state (e.g. awake versus anesthetized). We here used large-scale simulations of Izhikevich neurons [4] to determine the effect of weak perturbations on the overall network activity structure. Specifically, the model consisted of a two-dimensional layer of excitatory pyramidal cells (PYs) and a two-dimensional layer of fast-spiking inhibitory interneurons (INs). Fast, excitatory synapses connected PYs to PYs, and PYs to INs whereas INs provided fast inhibitory input to PYs. The connection topology for the PY-PY connections was chosen to be either local, small-world, or global random. Perturbations to the membrane voltage were applied by injection of weak currents. We assessed how weak perturbations propagated through the network as a function of the network configuration and what perturbation parameters enabled transitions between different macroscopic activity states. Understanding the effect of weak perturbations will help to elucidate sensory processing where sensory input has to overcome ongoing spontaneous network activity that may dominate overall activity structure in the awake animal. Furthermore, more refined brain (micro-)stimulation paradigms may be designed to specifically exploit regimes of high sensitivity to efficiently modulate pathological network dynamics as a treatment of neurological and psychiatric disorders with underlying functional pathologies in cerebral cortex. Acknowledgements Funding for this work from UNC Chapel Hill, Department of Psychiatry, and the Foundation of Hope is gratefully acknowledged.