Optimization of closed-loop electrical stimulation enables robust cerebellar-directed seizure control

Area of Science:

• Neuroscience
• Epilepsy Research
• Biomedical Engineering

Background:

• Temporal lobe epilepsy (TLE) requires new treatments.
• Cerebellar interventions show promise, but electrical stimulation yields inconsistent results.
• Understanding stimulation parameters is key for effective cerebellar therapy.

Purpose of the Study:

• To investigate the impact of varying electrical stimulation parameters on cerebellar control of seizures.
• To efficiently explore a large parameter space for seizure outcome optimization.
• To determine if personalized optimization is necessary for effective cerebellar stimulation.

Main Methods:

• Used a mouse model of temporal lobe epilepsy.
• Conducted on-demand electrical stimulation of the cerebellar cortex, varying charge, frequency, and pulse width.
• Employed Bayesian optimization with Gaussian process regression to efficiently identify optimal settings.

Main Results:

• Bayesian optimization successfully identified effective cerebellar stimulation settings for seizure control.
• Optimal settings were generally consistent across animals, suggesting personalization may not always be needed.
• Settings predicted to be ineffective failed to control seizures, validating the model's predictions.

Conclusions:

• Electrical stimulation of the cerebellum can robustly inhibit seizures in TLE.
• Correctly identifying stimulation parameters through methods like Bayesian optimization is essential for efficacy.
• This study provides a framework for optimizing cerebellar stimulation for seizure control.