Two fixed Hurst asymmetry tests did not confirm a shared pattern in separate iEEG and fMRI patient groups

We tested two predictions about left and right cortical Hurst maps in human epilepsy. Separate iEEG and fMRI patient groups came from one public release. The first local plan predicted greater absolute regional deviation on the treated side outside all cortical ROI names with a treated contact. Nine of 24 electrical candidates had valid spared homologous regions. All 12 other fMRI patients had whole-region maps. Mean contrasts were +0.1300 and +0.0028 reference SD units. Their 95% person bootstrap intervals were [-0.5801, +0.8009] and [-0.1200, +0.1438]. The shared test failed. An exploratory right-side excess in 11 of those 12 MRI patients prompted a fixed test in other people. Before new scores, we assigned eight source Bad patients to SEEG and 15 different Bad patients to fMRI. Whole available cortex gave mean right-minus-left contrasts +0.1765 and +0.0423, with intervals [-0.8545, +1.7618] and [-0.0160, +0.1018]. Only 2/8 and 10/15 were positive. Both shared claims had Holm p=0.9751. Whole-reference-person weight checks and processing checks retained uncertainty. Every negative value remains. A separate scalar calculation matched all 44 person scores and exact sign enumeration matched all four probabilities. A full replay matched 141 result files. The contribution is an anatomical comparison tool and benchmark, with no confirmed common neural pattern. The follow-up uses new people in reused data, with different outcome status and masks. It is not independent-study replication. Sparse clinical coverage, unequal spatial filters, limited normal reference support and source outcome-based harmonization limit scope. Code, public table exports, source pins and terms, local plans, figures, all results and repeat instructions are preserved.

epilepsy · intracranial EEG · SEEG · fMRI · Hurst exponent · hemisphere asymmetry · reference uncertainty · negative results · reproducible analysis