In 29 ds004789 RAM FR1 SEEG patients, 120 Hz peaks did not change the 70-150 Hz subsequent memory effect, but large peaks occurred in some pairs

We tested if the 120 Hz mains harmonic contaminates 70-150 Hz power (HFA) and its subsequent memory effect (SME) in US RAM SEEG data. We used ds004789 FR1 session 0, bipolar SEEG pairs, and 1.6 s word epochs (events sample/fs). We wrote and hashed the plan before we computed spectra. We included 29 patients and 3131 pairs. Subjects were sub-R1300 and higher, and the inclusion cutoff was by time only. We used a sign-flip test over subjects and a Holm correction over 4 tests, and all 4 tests were not significant. The 120 Hz peak ratio was a small dip and not a peak (mean -0.034 log10, Holm p 0.11). The 120 Hz and 60 Hz peak ratios did not differ between recalled and not-recalled words (Holm p 1). If you keep the 118-122 Hz bins, the HFA SME t changes by a mean of -0.008 (Holm p 0.80, largest subject change 0.118). But 4% of pairs had a 120 Hz peak of more than 2x. In one pair, these bins held 89% of the 70-150 Hz power. A 60 Hz peak of more than 2x occurred in 25% of pairs. Thus, in this sample, the notch choice does not bias the HFA SME, but a notch or a pair screen is still a safe step. The package contains the code, the per-word features, the SHA-256 provenance of each EDF byte chunk, and the results.

intracranial EEG · SEEG · high-frequency activity · mains harmonic · subsequent memory effect · free recall · line noise · OpenNeuro