BIOLOGICALLY CONSTRAINED CORTICAL CIRCUIT MODEL EXPLAINS PARADOXICAL RESPONSES TO MICROSTIMULATION IN HUMAN V1

Intracortical microstimulation (ICMS) has been shown to be a pertinent technology on which to build visual or somatosensory prosthetic systems. However, our understanding of how ICMS activates cortical circuits remains incomplete. While a wide body of literature has demonstrated a wide-spread long-lasting inhibitory effect of electrical stimulation, other studies have shown that stimulation can induce lasting activity in a subset of excitatory cells. Despite the seeming paradoxical nature of these findings, these phenomena were not yet systematically investigated in a single body of data, nor was a clear mechanistic explanation of this phenomena proposed.

To explain this paradox, we built a biologically detailed spiking neural network model of a cortical column incorporating differential excitatory and inhibitory time constants, spike-frequency adaptation and potassium-mediated depolarization block.

The model revealed that while low-amplitude stimulation leads to GABA-mediated lasting inhibition, high-amplitude pulses induce a localized depolarization block in inhibitory interneurons. This local failure of inhibition disinhibits the excitatory population, generating a lasting excitation at the stimulation point, while the surround remains dominated by synaptic inhibition.

To validate this prediction, we recorded responses to ICMS from the V1 of human volunteers. These recordings confirmed the spatiotemporal features seen in the model, with an amplitude-dependent transition from lasting inhibition to lasting excitation at the stimulation site.

We conclude that the interplay between synaptic time constants, differential cellular susceptibility to depolarization blocks and connectivity explains the spatiotemporal response to ICMS, offering insights for optimizing stimulation protocols in cortical prostheses.