Visual responses were stronger at the best ECoG contact in 8 of 9 patients

We compared visual 70–170 Hz broadband responses recorded at the same time from ECoG surface contacts and SEEG depth contacts in 9 patients in OpenNeuro ds004194 v3.0.0. Each patient completed two runs of a visual task (72 trials total). We selected the best contact on one run and measured its AUC on the other run, then reversed the runs. In 8 of 9 patients, the best ECoG contact had a higher cross-validated stimulus-versus-baseline AUC than the best SEEG contact (mean SEEG minus ECoG difference −0.266; exact paired sign-flip p = 0.0117). Bipolar SEEG gave a similar result (−0.268, p = 0.0078). Patient p05 was the exception: 17 of 25 SEEG contacts responded, and the best depth contact had AUC 0.98, compared with 0.78 for the best ECoG contact. The group difference in responsive-contact fraction was not significant (−0.033, p = 0.8203). With ECoG subsets matched to SEEG contact counts, the AUC difference was smaller and not significant (−0.154, p = 0.1250, 8 patients). Responsiveness declined with distance from responsive ECoG cortex; at 10–20 mm, 4/15 SEEG and 26/146 ECoG contacts responded. These results are limited to this small clinical cohort and its electrode locations. They indicate that typical depth coverage sampled visual responses less often, while a depth contact near responsive cortex could give a strong signal. Atlas labels were not used, depth coordinates were interpolated, and AUC measures stimulus detection against baseline rather than discrimination between stimulus classes.

electrocorticography · stereo-electroencephalography · visual cortex · broadband power · high gamma · brain-computer interfaces · OpenNeuro