At 10–30 mm, SEEG pairs on the same shaft were more similar in band power and aperiodic exponent

In archived RAM SEEG data, we compared bipolar pairs on the same shaft with pairs on different shafts within patients, matched by MNI midpoint distance. The analysis used a per-pair spectral feature table derived from archived Welch spectra for OpenNeuro ds004789 (RAM FR1) and ds004809 (RAM catFR1), with 60 seconds from the first WORD event for each of 256 patients. The exact dataset releases used for those archived spectra are not recorded. After excluding 3,345 pairs without a clear shaft prefix, 214 patients had at least one 5–30 mm distance bin with at least three comparisons of each type. Across 5–30 mm, same-shaft pairs were more similar in relative theta power (excess similarity 0.095, 95% CI 0.032–0.156; Holm p=0.015), relative alpha (0.145, 0.082–0.202; p=0.0003), relative beta (0.138, 0.075–0.200; p=0.0003), and the 2–40 Hz aperiodic exponent (0.071, 0.013–0.127; p=0.040). The result for log 2–45 Hz power was not significant (0.068, −0.003–0.139; p=0.073). There was no same-shaft excess at 5–10 mm; descriptive estimates at 10–30 mm ranged from 0.07 to 0.21. A non-pre-registered diagnostic found that, for fixed contact-number steps, excess similarity was near zero at the usual MNI distance for a step and increased at larger-than-usual MNI distances. Shared tissue along the track, shaft-level recording factors, and MNI coordinate or contact-spacing error may contribute; these data cannot separate the causes. The findings indicate that spatial null models based only on MNI distance may underestimate dependence between pairs on one shaft. Limits include one 60-second task segment per patient, no tissue labels, unknown contact spacing, and the assumption that contact-name prefixes identify shafts.

SEEG · bipolar pairs · spectral similarity · MNI distance · shaft dependence · spatial null models · RAM · aperiodic exponent