Tests did not confirm repeatable finger distances beyond a fixed hand-use model in five ECoG patients and six fMRI adults

We tested whether finger response distances have repeatable structure beyond a fixed hand-use model. Separate public studies supplied high-frequency ECoG power and 7 T fMRI. Two patients and one MRI adult supplied development. A local plan and code lock preceded reserved endpoint scores. Five other patients and six other MRI adults passed the fixed rules. One patient failed the trial-count rule. The model allowed arbitrary additive finger terms, any effect of separation in finger order, and a fixed target from 22 adults in a third study. Separate early and late recording blocks supplied residual distance products. Independent observations supplied each scale. Full-model mean products were +0.42910 and +0.00304. Their 95% person Student limits were [-0.69362,+1.55182] and [-0.00321,+0.00930]. Both shared primary rules failed, Holm p=0.348446. An exact design check left two contrasts after finger and separation control, and one after the hand-use term. All are blind to reflection-symmetric distance changes. Simulations showed positive bias when distance products shared a block. An added human bias check had opposite mean signs and remained uncertain. Independent calculations checked all 48 person-processing records, 3,108 power features, the external movement model and all MRI input rows. This is a checked method and a negative benchmark. It does not confirm positive neural generalization across modalities, model completeness, or causality. Small cohorts, clinical coverage, different tasks, source transforms and dependent errors limit scope. Code, source pins and terms, all outcomes, figures and repeat instructions are supplied. Neural arrays remain external.

human neuroscience · ECoG · fMRI · finger representation · hand use · model test · independent blocks · negative result