With stable means and independent repeat errors, mean finger response area was positive in five ECoG patients and six fMRI adults after feature mean removal

We tested whether five finger cues give repeatable spatial response area in separate public human ECoG and fMRI studies. Two patients and one MRI adult supplied development. A local code lock preceded reserved scores. The fixed count rule retained five patients and six MRI adults. Four separate repeat groups supplied an area crossproduct; separate calibration data set its positive scale. Under stable within-half expected responses and independent repeat errors, one measured response axis gives zero expected area. After feature mean removal, five-finger means were 0.021117 and 0.001977. Their pointwise 95% person Student intervals were [0.002085,0.040150] and [0.001179,0.002776]. All 11 scores were positive. The fixed shared mean test passed two-claim Holm correction, p=0.036908. The thumb-omission test failed, p=0.079099. ECoG intervals also crossed zero when feature means were kept or repeat groups were contiguous. A sign-only two-claim decision would fail. Thus the positive finding depends on the stated mean model and processing. An explicit-minor audit matched all 22 primary scores within 8.54e-18. Exact examples gave positive area with shared errors, changing rank-one means, or nonlinear measurement of one latent scalar. No latent rank, matching neural axes, common mechanism or causality was established. The contribution is a conditional finger benchmark and a checked compact assay. Code, all outcomes, source pins and terms, local plans, figures, audits and repeat steps are supplied. Neural arrays remain external.

finger responses · ECoG · fMRI · representational geometry · independent repeats · rank · measurement error