Seizure frequency reduction through Bayesian-optimized cerebellar stimulation and perturbation-based risk modeling

/Abstract Epilepsy’s unpredictable seizures impair quality of life for >50 million people, and one-third of patients remain drug-resistant. Recent work from the Krook-Magnuson and Netoff labs demonstrates that cerebellar stimulation can truncate seizure duration. Additionally, studies from epilepsy-focused dynamical systems labs show that network instability linked to seizures can be revealed through perturbation-based measures, such as line length of evoked responses. Building on these advances, this predoctoral project pursues two independent yet complementary aims. Aim 1 will reduce behavioral seizure frequency by applying Gaussian-process/Bayesian optimization to identify cerebellar stimulation parameters that most effectively suppress behavioral seizure frequency in the ventral intrahippocampal kainate mouse model of temporal-lobe epilepsy. Aim 2 will improve seizure-risk assessment by delivering brief, sub-threshold hippocampal perturbations and quantifying evoked responses to assess “critical-slowing” and model proximity to seizure bifurcation, benchmarking performance against standard passive depth EEG based features. The integrated experimental–computational strategy is innovative in (i) shifting cerebellar neuromodulation from reactive seizure shortening to preventive seizure suppression, requiring optimization to succeed with fewer sampling data points, and (ii) coupling dynamical-systems theory with active probing to expose latent network fragility in real time. Expected outcomes include a stimulation procedure that lowers behavioral seizure incidence and a high-fidelity risk metric to measure seizure risk dynamically. Together they will advance the field toward anticipatory intervention. Training will take place through the Graduate Program in Neuroscience at the University of Minnesota, leveraging resources such as the Center for Neuroengineering and the broader university research infrastructure. Under the mentorship of Dr. Esther Krook-Magnuson (sponsor; epilepsy and circuit physiology) and Dr. Theoden Netoff (co-sponsor; optimization algorithms for neurostimulation), I will receive integrated training across neuroscience, engineering, and mathematics. This includes advanced coursework, hands-on experience with in vivo electrophysiology, and guided development in machine learning for neural data analysis. The fellowship will equip me to build the interdisciplinary expertise necessary to launch an independent research career focused on neuromodulation, predictive modeling, and translational epilepsy research. Project Number: 1F31NS149569-01 | Fiscal Year: 2026 | NIH Institute/Center: National Institute of Neurological Disorders and Stroke (NINDS) | Principal Investigator: Brandon Hoang | Institution: UNIVERSITY OF MINNESOTA, MINNEAPOLIS, MN | Award Amount: $37,548 | Activity Code: F31 | Study Section: Special Emphasis Panel[ZRG1 F01A-N (20)] View on NIH RePORTER: https://reporter.nih.gov/project-details/11386826