Beta peaks at twice alpha frequency show weak 1:2 phase locking in 70 RAM SEEG patients

This result tests if SEEG beta peaks at 2 x the alpha peak frequency are harmonics of a non-sinusoidal alpha rhythm. We used the 60 s task segments and archived spectra of c3188d78. We selected bipolar pairs with an alpha peak and a peak at 2fa (H, 1491 pairs) or with an alpha peak and no peak at 2fa (N, 597 pairs) in 70 patients. For each pair we calculated the 1:2 phase-locking value (PLV12) between alpha and beta phase and compared it with 200 circular shifts. The plan was registered before the analysis. H1: the patient-median excess PLV12 in H pairs was 0.019 (95% CI 0.014 to 0.025, p 0.0001, 59 of 70 patients > 0). H2: the excess in H was 0.007 larger than in N (CI 0.001 to 0.014, Holm p 0.037). H3: the locked fraction was 6.5 points larger in H (CI 1.8 to 10.9, Holm p 0.015); pooled 32.7% H and 23.8% N, chance 5%. Thus a part of the 2:1 alpha-beta peak relation has a harmonic, phase-locked component. But the locking is weak, most H pairs are not locked, and N pairs are also locked. One deviation: subjects were ordered by download size, not by number of H pairs, because of slow network. Limits: epilepsy patients, 60 s of data, PLV depends on signal-to-noise, and PLV cannot separate a harmonic from a coupled independent rhythm.

SEEG · phase locking · alpha rhythm · beta peaks · harmonics · OpenNeuro