Sparse SEEG coordinates prevent a group test of NREM slow-wave direction in OpenNeuro ds007118/9/20 v1.0.0

We tested whether 0.5–2 Hz slow waves in stereo-electroencephalography (SEEG) recordings during non-REM sleep showed an anterior-to-posterior lag pattern. Across 78 first-run recordings with SEEG channels from OpenNeuro datasets ds007118, ds007119, and ds007120 (version 1.0.0), 337 of 1,562 SEEG contacts (21.6%) had coordinates, compared with 6,100 of 6,140 ECoG contacts (99.3%). Only two recordings had at least eight normal bipolar SEEG pairs with coordinates. The pre-registered group test therefore had n = 2 and could not produce a two-sided p < 0.05 (minimum p = 0.5). In these two patients, anterior-posterior coordinate and lag projection correlations were positive (Spearman rho = +0.28 and +0.22), opposite to the hypothesized anterior lead; the exact sign-flip p was 0.5 and Holm-adjusted p was 1.0. The correlation sign was not stable when only stronger edges were used. With this release and method, group-level testing of NREM slow-wave direction in SEEG was not possible, and the two testable patients did not show an anterior lead. The analysis used 90 s per patient, assumes the recordings were NREM as described by the dataset, and did not detect individual slow waves.

stereo-electroencephalography · SEEG · non-REM sleep · slow waves · travelling waves · lag projection · OpenNeuro · contact coordinates · metadata coverage · cross-correlation