Mirror-matched interictal HUP SEEG pairs show stronger 1–40 Hz coupling in 28 of 29 patients; the 65–115 Hz effect is small and not robust

We compared spontaneous coupling between mirror-matched (homotopic) bipolar stereo-EEG pairs and other contralateral pairs matched for distance in interictal recordings from the HUP iEEG dataset (OpenNeuro ds004100, snapshot 1.1.3). The analysis included 29 bilaterally implanted patients with matched edges, using fsaverage x = 0 as the midline. For each homotopic edge (mirror distance ≤15 mm), coupling was compared with the mean for non-homotopic edges in the same patient within 5 mm of its Euclidean length. The primary 1–40 Hz absolute Pearson correlation difference was +0.0248 (95% bootstrap CI 0.0195 to 0.0306), positive in 28/29 patients (two-sided sign-flip p = 5×10⁻⁵; Holm p = 1.0×10⁻⁴). The raw mean |r| was 0.077 for homotopic and 0.048 for non-homotopic edges. The difference remained after removing seizure-onset-zone pairs (+0.0263), controlling for signal power (+0.0219), and using 10 mm (+0.0366) and 20 mm (+0.0196) mirror thresholds. A post hoc analysis found the low-frequency effect tracked anterior-posterior and superior-inferior matching, not equal depth from the midline. For the 65–115 Hz Hilbert-envelope correlation, the primary difference was smaller (+0.0061, 95% CI 0.0010 to 0.0119; 22/29 patients positive; Holm p = 0.029) and was not robust to power control (p = 0.11) or the 10 mm threshold (p = 0.19); it should not be treated as a positive finding. Coupling values were small and zero-lag correlations may include shared noise or artifact. Segments were short (20–90 s), one per patient; no tissue labels were available, and all patients had drug-resistant epilepsy with clinically determined implant plans. These limits constrain generalization.

stereo-EEG · interictal · interhemispheric coupling · homotopic connectivity · bipolar recordings · epilepsy · OpenNeuro · HUP iEEG