A fixed test did not confirm greater thumb contrast energy in two ECoG patients and six fMRI adults

We compared thumb contrast energy with mean energy per contrast among the other four fingers. Separate public human studies supplied high-frequency ECoG power and M1 fMRI patterns. A local plan preceded new endpoint scores. Development used two patients and one MRI adult. The fixed four-trial cell rule kept only two of six reserved patients, but all six reserved MRI adults. An exact count model predicts 2.03 calibration passes under uniform cue order. Two passed. Mean thumb-minus-nonthumb scores were 0.102757 and 0.099706. Their pointwise 95% person Student intervals were [-0.822965,1.028479] and [0.043266,0.156146]. The shared rule also required positive thumb energy in both modalities. It failed, intersection p=0.196317. An added two-trial rule kept all six patients. Their mean difference was 0.476392 [0.014975,0.937809], with six positive people. This was exploratory after the primary result. The four newly included patients had interval [-0.010119,1.336538]. All ten simultaneous added-check intervals included zero. Movement adjustment and feature-mean retention weakened the added scores. Negative energies remain. Independent contrast, distance and raw Fourier calculations checked the arithmetic. Exact models show that one response axis or shared errors can give a positive score. The package preserves source versions and terms, code, all results, figures and repeat steps. Source arrays stay external. This is a failed fixed test and a count-sensitivity benchmark. It does not confirm cross-modal generalization or causality.

human neuroscience · ECoG · fMRI · finger representation · thumb · representational geometry · calibration · negative result