In-silico EEG biomarkers of reduced inhibition in human cortical microcircuits in depression

Abstract

Reduced cortical inhibition by somatostatin-expressing (SST) interneurons has been strongly associated with treatment-resistant depression. However, due to technical limitations it is impossible to establish experimentally in humans whether the effects of reduced SST interneuron inhibition on microcircuit activity have signatures detectable in clinically-relevant brain signals such as electroencephalography (EEG). To overcome these limitations, we simulated resting-state activity and EEG using detailed models of human cortical microcircuits with normal (healthy) or reduced SST interneuron inhibition (depression), and found that depression microcircuits exhibited increased theta, alpha and low beta power (4–16 Hz). The changes in depression involved a combination of an aperiodic broadband and periodic theta components. We then demonstrated the specificity of the EEG signatures of reduced SST interneuron inhibition by showing they were distinct from those corresponding to reduced parvalbumin-expressing (PV) interneuron inhibition. Our study thus links SST interneuron inhibition level to distinct features in EEG simulated from detailed human microcircuits, which can serve to better identify mechanistic subtypes of depression using EEG, and non-invasively monitor modulation of cortical inhibition. Author summary Reduced somatostatin-expressing interneuron (SST) inhibition has been implicated in depression. However, it is impossible to establish experimentally in humans how these inhibitory changes are reflected in clinically relevant brain signals (electroencephalography, EEG). We performed detailed simulations of human neuronal network activity and EEG signals in health and depression. We found that reduced SST inhibition led to significant changes in EEG, which accounts for changes seen in depression. Our study thus provides biomarkers that link inhibition level of a key interneuron type to measurable signatures in EEG. These biomarkers can be used to identify subtypes of depression and non-invasively monitor cortical inhibition modulation.