Fixed tests did not confirm more than one category response direction in 14 SEEG patients and five fMRI adults

We tested if object categories produce more than one stable spatial response direction after a constant animal/person-versus-other pattern is removed. Separate public studies supplied SEEG high-gamma amplitude and BOLD fMRI. Two pictures per category set site scales. Different pictures estimated two test halves. Disjoint within-class category contrasts gave one signed area score per person. Development used three patients. A revised plan and code lock preceded tests in 14 other eligible patients and all five MRI adults. Mean areas were -0.00008564 and 0.00015700. Their 95% person Student limits were [-0.00023919,0.00006790] and [-0.00008392,0.00039791]. The shared positive-mean and sign-flip rule failed, p=0.87516. Controls with eight sites and separate MRI runs were also uncertain. Negative and discordant results remain. An exact model exposed an interpretation limit. Independent picture halves with zero mean gave positive area in all four sign outcomes when categories shared errors. Four independent groups removed that model bias. The initial four-group human design failed contact coverage in development. Independent source calculations matched all 19 included scores within 4.21e-19. A full source repeat matched recorded result bytes. The contribution is a negative visual benchmark, checked area tools and an exact error model. Separate cohorts and pictures, source order, missingness, sparse clinical sites, five MRI people and unequal filters limit scope. No positive cross-modal neural effect, common source count or cause was established. Code, source versions and terms, all scalar results, figures and repeat steps are supplied. Original neural arrays remain external.

human neuroscience · SEEG · fMRI · visual categories · response rank · signal moments · independent errors · negative result