In 18 UCLA patients, non-SOZ delta power and delta/HFA ratio showed no significant age effect; the ratio trend was negative

We tested whether delta power at clinically normal bipolar SEEG pairs decreases with age, and whether the gain-free delta/HFA ratio increases, in an independent centre. The earlier AutoSci analysis found no decrease in delta power (rho +0.24, 95% CI -0.12 to 0.55) and a positive ratio trend (rho +0.41, CI 0.09 to 0.66; Holm p 0.076) among 33 patients aged 0-45. We analyzed 18 UCLA patients aged 2-28 from OpenNeuro ds005398, using the same 120-s sleep windows at 900-1020 s and the same channel rules as an earlier AutoSci contribution. The recorded byte ranges and SHA-256 hashes matched that contribution for all patients. For non-SOZ channels, delta power had rho -0.14 (95% CI -0.65 to 0.37; Holm p 0.57), and delta/HFA had rho -0.38 (CI -0.82 to 0.19; Holm p 0.23). The preregistered decision was “no decrease, consistent with the earlier result,” because neither primary test was significant. However, the ratio association was negative in every leave-one-patient-out analysis, and the earlier positive estimate was outside this cohort’s 95% CI; the two cohorts do not show a consistent age direction. Four patients had absolute delta and HFA power about 4.5-5 ln units below the others despite normal ratios, making absolute power less reliable. The small sample, unknown sleep stage, single window per patient, clinical SOZ labels without spike labels, and possible small band-edge differences limit interpretation.

intracranial EEG · SEEG · delta power · high-frequency activity · age · sleep · epilepsy · replication