Shared typicality tests failed in eight EEG and three fMRI people; two concepts had large effects on mean scores

We tested whether rated category typicality predicts category affinity in separate human EEG and fMRI studies. A local plan reserved eight EEG people, all three fMRI people, and 144 target concepts in nine categories. Two other EEG people and separate targets supplied development. H1 controlled category, generic neural affinity and an appearance measure. H2 also controlled object effects. A negative H2 direction was added from development before confirmation. All three shared tests failed. Positive H1 and H2 shared Holm p values were 0.11752. H2 mean partial correlations were 0.327 and 0.603. Their 95% person intervals were [0.165,0.473] and [-0.194,0.921]. Joint reference, target and person resampling intervals included zero in both modalities. The membership graph left only 37 contrasts after object and category effects. An exploratory removal of goat and crayfish lowered mean H2 correlations to 0.008 and 0.252. Scores fell in all 11 people. The selected removal was among several large influence pairs; all 780 pair removals are supplied. Its descriptive fMRI person interval for the fall included zero. This check has no independent confirmation or adjusted population claim. The contribution is a tested assay, graph diagnostic and compact benchmark. Small cohorts, sparse overlap, source processing, rapid EEG image timing, image differences and uncertain resampling coverage limit scope. The result does not establish a shared neural typicality code, its absence, or causality. Code, exact source trace, licensed inputs, all outcomes, figures, local plans and repeat steps are preserved.

human neuroscience · EEG · fMRI · object typicality · category overlap · fixed effects · reference sampling · reproducibility · negative result