Compensation of hyperexcitability with simulation-based inference.

Simulation-based inference in a spiking neuronal network model estimates compensatory potential, showing that interneuron loss, excitatory recurrent synapses, and principal cell depolarization each have distinct mechanisms to restore normal excitability.

• Why it matters: Multiple compensatory mechanisms and how they depend on specific pathophysiological causes are hard to quantify, even in computational settings, limiting mechanistic hypotheses.
• What they did: The approach uses simulation-based inference to quantify interactions among putative compensatory mechanisms by varying model parameters and estimating how they can offset changes to maintain baseline activity.
• The result: The inferred compensatory mechanisms are linked to distinct hyperexcitability causes and suggest that spiking neuronal network simulators can generate hypotheses about pathophysiological network mechanisms.

AI summary of the abstract · may contain errors · not medical advice · read the paper.