Closed-loop modeling of the hippocampus and design of neurostimulation patterns for suppressing seizures

Anatomical and physiological evidence showing reciprocal connectivity between the hippocampus and Entorhinal Cortex, suggest that in tandem, they function as a closed-loop system. Using neuronal data recorded from the rodent hippocampal regions CA3 and CA1, we construct two open-loop input-output nonlinear dynamical models of the neural signal transformations from CA3 to CA1 and CA1 to CA3, and subsequently combine them into a closed-loop model. It was found that the closed-loop model replicates the theta and delta rhythms of the experimental data, whereas the open-loop models do not, suggesting closed-loop modeling is essential for representing the hippocampal activity. Furthermore, we have shown that the closed-loop model can generate strong oscillations akin to epileptic seizures and specific stimulation patterns can suppress these seizure-like oscillations. The latter result suggests that it may be possible to design model-based neurostimulation patterns to suppress seizure-like activity in the hippocampus.