SEEG speech-related high-gamma activity was bilateral during spoken free recall

We tested if speech-related high-gamma activity in SEEG is left-lateralized or bilateral. Data: archived per-pair 70-150 Hz traces (dB re -2.0 to -1.5 s) around annotated vocal onset of isolated spoken recalls from publication 83bd088b (OpenNeuro ds004789 v3.1.0, RAM FR1, session 0, pure-SEEG subjects). We added the hemisphere and MNI coordinates of each contact from the dataset electrodes.tsv files. We wrote the plan (PLAN.md, SHA-256) before we joined hemisphere to the HFA values. Unit: subject x hemisphere x region. Sensorimotor cortex (SMC, precentral/postcentral) uses -0.5 to 0 s; auditory cortex (AUD, superior/transverse temporal) uses 0 to 0.5 s. The left minus right test permutes the hemisphere labels within each subject. We used Holm correction over 6 tests. Result (54 subjects, 1249 pairs): SMC power increased in the left (+0.28 dB, 31/36 units, Holm p = 0.0006) and in the right hemisphere (+0.16 dB, 21/29, Holm p = 0.010). AUD power increased in the left (+0.67 dB) and right (+0.70 dB) hemisphere (both Holm p = 0.0006). Left minus right was not significant: SMC +0.12 dB (95% CI -0.004 to 0.24, Holm p = 0.14); AUD -0.03 dB (CI -0.37 to 0.30, p = 0.86). Thus this activity is bilateral. A ventral SMC subset (secondary, uncorrected) showed left > right (+0.24 dB, p = 0.021). Thus a moderate left amplitude bias is possible. Limits: most subjects have one hemisphere only (11 SMC and 19 AUD subjects have both); language dominance and handedness are not known; atlas labels; annotated, not acoustic, vocal onset. The package has the code, inputs with hashes, results and reproduction steps (numpy only, about 5 s).

spoken free recall · SEEG · high-gamma activity · hemispheric lateralization · sensorimotor cortex · auditory cortex · speech production