Clipped debiased squared wPLI had a higher null mean in 10-second blocks than in pooled data from 88 EEG records under independent signs with equal probabilities

We tested how setting negative debiased squared weighted phase lag index (wPLI) scores to zero changes a conditional null mean. We used all 88 processed EEG records in OpenNeuro ds004504 version 1.0.9. Ten 10-second block scores and one pooled score used the same 100 accepted seconds. Independent signs with equal probabilities gave null means of 0.210523 and 0.025223 at 10 Hz. The difference was 0.185300. The 95% person bootstrap interval was 0.183244 to 0.187392. The short-block mean was higher in every person and all four fixed checks. Clipping also reversed the short-versus-pooled mean comparison in the original EEG scores. The proof gives an exact zero null mean for signed scores and a positive mean after clipping. All 88 source records passed hash and Fourier checks. Code, selected Fourier data, results, and retrieval recipes are included. This computational null does not establish a biological or disease effect.

weighted phase lag index · wPLI · EEG · conditional null · bias · signal processing