A cortical microcircuit model reveals distinct inhibitory mechanisms of network oscillations and stability

We identify a computational mechanism for network oscillations distinct from classic excitatory-inhibitory networks – CAMINOS (Canonical Microcircuit Network Oscillations) – in which different inhibitory-interneuron classes make distinct causal contributions to network oscillations and stability. A computational network model of the canonical microcircuit consisting of SOM, PV and excitatory neurons reproduced key experimental findings, including: Stochastic gamma oscillations with drive-dependent frequency; precise phase-locking of PV interneurons and delayed firing of SOM interneurons; and the distinct effects of optogenetic perturbations of SOM and PV cells. In CAMINOS, the generation of network oscillations depends on both the precise spike timing of SOM and PV interneurons, with PV cells regulating oscillation frequency and network stability, and delayed SOM firing controlling the oscillation amplitude. The asymmetric PV-SOM connectivity is found to be the key source ingredient to generate these oscillations, that naturally establishes distinct PV and SOM-cell spike timing. The CAMINOS model predicts that increased SOM/PV densities along the cortical hierarchy leads to decreased oscillation frequencies (from gamma to alpha/beta) and increased seizure susceptibility, suggesting a unified circuit model for oscillations across different frequency bands.