Human iEEG and fMRI did not show a shared time-scale rank across 86 sampled cortical areas

We tested whether fast intracranial EEG and slow BOLD fMRI share a regional time-scale order. Public MNI iEEG estimates from 106 patients and HCP 7 T maps from 180 released rows supplied 86 areas with at least ten iEEG patients. Rank correlation was -0.135, with a 95% whole-person/row resampling range of -0.311 to +0.069. After tissue-contrast and position controls, correlation was +0.170, with range -0.187 to +0.393. Moran spatial-reference Holm p values were 0.9161 and 0.6094. Rotation tests and a second fMRI estimator also gave no support. iEEG split-half map stability was much lower than fMRI stability. We then selected an orbitofrontal-versus-lateral-frontal contrast from these maps and fixed new-cohort tests. In 107 RAM depth-electrode patients, the mean log10 time-scale contrast was +0.11763, with range +0.05841 to +0.17539. In ten people with two processed movie-fMRI runs, it was -0.01694, with range -0.02926 to -0.00403. Holm p values within this two-test follow-up were 0.0004 and 0.037109. Position plus spectral-exponent adjustment removed the SEEG contrast. The movie ACF-sum check included zero. A later NORDIC check retained both negative movie estimates in the same people. Correction across all four primary tests gave movie p=0.111328. Thus the original shared-order prediction was unsupported, and the selected opposite contrasts have limited evidence. Separate cohorts, tasks, processing, sparse electrodes, and BOLD signal quality prevent a causal modality claim. The package preserves code, plans, source hashes, draws, figures, negative results, and repeat instructions. Independent numerical checks and a full offline repeat support the reported calculations.

human neuroscience · intracranial EEG · SEEG · fMRI · time scales · spatial statistics · independent confirmation · negative result