Authors :
Presenting Author: Ajay Deep Kachhvah, PhD – University of Minnesota
Ganne Chaitanya, MD, PhD – The University of Texas Health Science Center at Houston
Sandipan Pati, MD – University of Minnesota
Jing Wang, PhD – University of Minnesota
Rosana Esteller, PhD – Boston Scientific
Rationale:
Seizures in chronic epilepsy exhibit remarkable autonomy—self-initiating, propagating, and terminating without external triggers. While electrographic features of ictal dynamics are well characterized, it remains unknown whether these transition phases are governed by patient-specific endogenous rhythmic patterns or are modifiable by chronic neuromodulation. This question is critical in light of epilepsy's well-documented chronobiology (from circadian to multidien cycles) and the kindling phenomenon, in which seizures appear to “learn” to self-perpetuate. If neurostimulation is to function as an anti-kindling therapy, it must modulate these fundamental transition rhythms. We hypothesized that seizure onset and termination exhibit distinct rhythmic signatures with differential sensitivity to chronic stimulation, and that their modulation correlates with clinical outcomes.
Methods:
We analyzed 932 electrographic seizures from 18 patients with drug-resistant temporal lobe epilepsy implanted with bilateral hippocampal RNS systems. Intrinsic mode functions (IMFs) were extracted from local field potentials using empirical mode decomposition (EMD). Instantaneous phase relationships were derived via Hilbert transform and clustered using eigenvalue–eigenvector decomposition. Phase-locking values (PLV) were computed during seizure onset, midpoint, and termination, with statistical validation using Fourier-shuffled surrogates. Longitudinal changes in synchronization frequencies were stratified by clinical response: super responders ( >90% seizure reduction), intermediate (50–90%), and poor responders (< 50%).