Prolonged Amygdala Beta-power May Underlie Post-ictal Respiratory Suppression and SUDEP Vulnerability in Temporal Lobe Epilepsy

Introduction: The amygdala plays a critical role in seizure-induced respiratory arrest, which can persist postictally without patient awareness or air hunger. Seizure-related respiratory complications represent a major mechanism underlying sudden unexpected death in epilepsy (SUDEP), with unaware seizures conferring substantially higher SUDEP risk than aware seizures. However, the network dynamics differentiating aware from unaware seizures remain poorly understood. While the amygdala's causal role in ictal apnea is well-supported, the electrophysiological signatures that distinguish these seizure types and identify patients at risk for persistent respiratory suppression remain unknown. We hypothesized that sustained post-ictal amygdala beta power represents pathological network dysfunction that maintains respiratory suppression, thereby increasing SUDEP risk.

Methods: We analyzed SEEG recordings from 58 drug-resistant focal epilepsy patients at Vanderbilt University Medical Center. Seizures were classified as aware or unaware based on clinical semiology and responsiveness testing. We computed dynamic beta power (13-30 Hz) and alpha-band (8-13 Hz) partial directed coherence (PDC) for the seizure onset zone (SOZ), amygdala, and non-involved zones (NIZs). Time-frequency analyses examined network dynamics from 50 seconds pre-ictal through 150 seconds post-ictal, using 10-second smoothing windows, 95% confidence intervals, and Bonferroni-Holm correction for multiple comparisons.

Results: At seizure onset, the SOZ abruptly transitions into a primary network target, evidenced by a spike in net inward connectivity that was significantly more pronounced in aware seizures (p < 0.01). PDC analysis identified the amygdala as a key source node providing directed connectivity to the SOZ, with aware seizures demonstrating significantly higher amygdala-to-SOZ causal influence during ictal and early post-ictal periods relative to unaware seizures (p < 0.01). Moreover, the local consequences were evident in beta power dynamics—while NIZs remained stable and returned to baseline postictally, amygdala beta power demonstrated divergence between seizure types. In aware seizures, amygdala power rapidly normalized, whereas in unaware patients it remained persistently elevated for over 100 seconds (p < 0.01).

Conclusion: Our findings identify persistent post-ictal amygdala beta power as a potential electrophysiological biomarker distinguishing seizures with impaired awareness. This sustained elevation suggests pathological network inertia, where the amygdala remains hyperexcitable but functionally isolated after SOZ activity resolves. The differential amygdala-SOZ connectivity patterns and prolonged post-ictal beta elevation in unaware seizures provide mechanistic insights into awareness preservation during seizures. Given that most SUDEP cases occur postictally and involve respiratory suppression without arousal, persistent amygdala beta elevation represents a promising biomarker for SUDEP vulnerability and potential target for real-time monitoring and targeted neuromodulation interventions in high-risk patients.