Most SEEG contacts lacked visual responses, and responsive contacts showed no consistent d′ difference from ECoG in two patients

This offline analysis compared single-trial visual broadband responses (70–200 Hz) from SEEG depth and ECoG surface contacts in nine patients performing the OpenNeuro ds004194 spatialpattern task. Trial response was the change in log10 broadband power from −0.25–0 s to 0.05–0.30 s after stimulus onset. In a pre-specified split-half procedure, responsive contacts were selected by a one-sided test (p < 0.001), then d′, response magnitude, and post-versus-pre AUC were measured on held-out trials. Only p05 and p09 met the criterion of at least three responsive contacts of each type. Their median held-out d′ differences (SEEG minus ECoG) were +0.060 and −0.396; the exact sign-flip p-value was 1.0, and the contact-label permutation p-value was 0.059. Most SEEG contacts did not respond, while the responsive-contact results did not show a consistent advantage for either type. The data therefore establish neither equivalence nor a difference between SEEG and ECoG. With only two eligible patients, the sign-flip test has no power to detect p < 0.5; a two-sided p < 0.05 requires at least six patients. Anatomical matching was unavailable, some depth contacts may have been in white matter or sulci, and the two contact types used different common-average-reference groups. Results apply to this offline analysis of one task and cohort.

SEEG · ECoG · visual cortex · broadband power · single-trial response · intracranial EEG · functional matching · brain-computer interface