Empirically constrained network models for contrast-dependent modulation of gamma rhythm in V1

• DOI: 10.1016/j.neuroimage.2021.117748
• Publication Date: 2021-01-15T14:10:22Z
• Authors: Margarita Zachariou; Mark J. Roberts; Eric Lowet; Peter De Weerd; Avgis Hadjipapas
• Keywords: Ping (video games); Network model; Interneuron

Gamma oscillations are thought to play a key role in neuronal network function and communication, yet the underlying generating mechanisms have not been fully elucidated date. At least partly, this may be due fact that even simple models of interconnected inhibitory (I) excitatory (E) neurons, many parameters remain unknown set based on practical considerations or by convention. Here, we mitigate problem requiring PING (Pyramidal Interneuron Network Gamma) simultaneously satisfy broad criteria for realistic behaviour empirical data spanning both single unit (spikes) local population (LFP) levels while varied. By doing so, were able constrain parameter ranges select empirically valid models. The derived model constraints implied weak rather than strong as mechanism gamma, connectivity between E I neurons within specific bounds, variations external input but neurons. Constrained showed behaviours, including gamma frequency increases with contrast power saturation decay at high contrasts. Using an empirically-validated studied route instability This involved increased heterogeneity increasing triggering breakdown neuron pacemaker function. Further, illustrate model's capacity resolve disputes literature concerning oscillation properties GABA conductance proxies. We propose our study will useful other modelling studies, approach constraining can expanded when richer datasets become available. As networks building blocks larger aim understand complex cognition through their interactions, there is considerable value improving these blocks.