Human iEEG knee and denoised BOLD time scale maps showed little regional agreement across 142 cortical parcels

We tested whether fast iEEG knee time scales and slow BOLD time scales share a cortical order beyond anatomy. We reused published estimates from 106 clinical iEEG patients and 180 HCP fMRI subjects. The cohorts are separate. The main test used 142 parcels with at least five covered iEEG patients; 105 patients contributed in that set. With a time-domain BOLD fit, the regional rank correlation was -0.080, with a 95% participant resampling range of -0.266 to +0.109. After T1w/T2w adjustment it was -0.090, with a range of -0.279 to +0.114. Two spatial tests gave large probabilities; both primary Holm probabilities were 1. A second BOLD fit gave -0.071 overall and -0.089 after adjustment. The iEEG anatomical relation was present within patients, but the corresponding BOLD relation was near zero. Sampling, hemisphere, region deletion, and electrode-level checks did not establish positive transfer. Patient-effect and position controls gave different directions, which we retain. A later diagnostic found high BOLD map consistency and moderate iEEG map consistency across subject splits. These splits are exploratory. HCP family links were unavailable, so bootstrap limits are conditional resampling ranges. The result does not show equivalence or reject a hierarchy under other signal definitions. There is no independent external confirmation. The package contains code, fixed source hashes, the plan, all scalar results, figures, and repeat instructions. Source arrays stay external under their original terms.

human neuroscience · intracranial EEG · fMRI · time scales · spatial statistics · cross-modal generalization · negative result