Time since the previous response click did not change idle-window SEEG movement-detector scores in 15 patients

We tested whether time since the previous response click changes idle-window movement-detector scores in 15 patients with stereo-electroencephalography (SEEG) during a mouse-movement task. We used archived 70–150 Hz high-gamma activity window features, with REST defined as the 0.5 s before stimulus onset and MOVE as 0–0.5 s after movement onset. A shrinkage linear discriminant analysis (LDA) detector used five contiguous time-ordered folds. The primary outcome was the partial Spearman correlation between REST-window score and log gap, controlling for trial index. Its mean was -0.002 (95% CI -0.027 to +0.023; 6 of 15 patients negative; exact one-sided sign-flip p = 0.44), and the interval was inside the pre-registered equivalence bound of +/-0.1. The false-positive rate difference between the shortest-gap and longest-gap thirds was -0.015 (95% CI -0.044 to +0.017; p = 0.82). MOVE-window scores also did not change (mean rho -0.009, 95% CI -0.047 to +0.029; p = 0.67). Exploratory analysis found higher raw mean log high-gamma activity at rest after longer gaps (mean rho +0.091, 95% CI +0.052 to +0.133; 14 of 15 positive; p = 0.0004), while the LDA projection did not follow that change. A pooled-LDA robustness analysis gave the same conclusion. These offline results support earlier null results in a second dataset with a different task. Limits include one session per patient, gaps of about 2–15 s (the shortest about 2–3 s), gaps not controlled by the experiment, and a five-fold scheme that trains on later trials for early folds. The archived report, pre-registration, code, results, per-patient values, provenance, and source hashes are preserved in the supporting files.

stereo-electroencephalography · high-gamma activity · movement detection · brain-computer interface · false positives · linear discriminant analysis · idle-state control