Abstracts

Rationale: Responsive neurostimulation (RNS) is an established treatment for drug-resistant epilepsy with variable response. It remains unclear which patients will respond best to the therapy. Growing evidence suggests that a patient's response to RNS depends on several factors, including the brain's structural connectome and state-dependent brain activity. Neurotransmitters are crucial as they act as a key regulator of brain function, influencing how different regions communicate and how the brain reacts to perturbations like neurostimulation (Stoof et al., 2024). Here, we test the hypothesis that regional neurochemical characteristics are associated with RNS responsiveness.

Methods: We retrospectively analyzed 47 patients with drug-resistant focal epilepsy implanted at the University of Kansas Medical Center (KUMC) with cortical RNS leads, all with at least one year of follow-up. Electrode locations were mapped to 19 regional neurotransmitter density profiles derived from publicly available normative PET-based atlases (Hansen et al., 2022). Patients were classified as responders (≥50% seizure reduction, ILAE Class ≤4) or non-responders (ILAE Class 5–6). For each brain parcellation (100- and 400-region atlases; Schaefer et al., 2018), we compared the mean neurotransmitter values of the regions underlying each electrode contact between responder and non-responder groups. Statistical significance was assessed using permutation testing (n =10,000) with randomized responder/non-responder electrode assignments and corrected for multiple comparisons using False Discovery Rate (FDR).

Results: Distinct neurotransmitter signatures differentiated responder from non-responder sites. Norepinephrine (NE) profiles showed the most robust association with responsiveness, significant in both parcellations (100-ROI: p=0.004, FDR q=0.09; 400-ROI: p=0.001, FDR q=0.045). Dopamine D2 receptor density also overlapped with responder sites in the 100-ROI analysis (p=0.001) but did not pass FDR correction.

Conclusions: These findings suggest that regional neurotransmitter environments, particularly NE and Dopamine D2 distributions, may underlie differential RNS responsiveness. Neurotransmitter-informed models could therefore provide a novel biomarker framework for patient stratification and provide a mechanistic basis for personalized neuromodulation strategies.

Funding: AES BRIDGE program