Closed-loop systems for Epilepsy

Closed-loop approaches come in a variety of forms; on-demand responsive intervention strategies are one example of a closed-loop approach. Through online seizure detection, on-demand intervention provides temporal specificity. On-demand optogenetic intervention provides additional cell-type, spatial, and direction of modulation specificity. Such a highly selective intervention strategy can be a successful strategy, as illustrated by selective on-demand optogenetic inhibition of dentate gyrus granule cells near the seizure focus. On-demand optogenetic interventions can also be used to gain deeper insight into the epilepsies, and to explore potential intervention targets distant from the seizure focus, as illustrated by on-demand optogenetic work targeting the cerebellum. This work additionally illustrates that specificity of intervention can also be critical to providing robust seizure inhibition. While optogenetics is an excellent research tool, it is not yet clinically available to epilepsy patients. Electrical stimulation provides an opportunity for more immediate translation, but deciding on appropriate stimulation parameters (e.g., frequency, pulse width, amplitude) can be critically important to success. Bayesian optimization with Gaussian process regression is an additional example of a closed-loop approach, and can be coupled with on-demand intervention strategies. Using this approach, we demonstrate robust inhibition of hippocampal seizures in a mouse model of temporal lobe epilepsy via on-demand electrical stimulation of the cerebellum. Closed-loop optimization approaches have the potential to aid in the clinical realization of successful stimulation interventions.