A fixed test of 31 object concepts did not confirm a shared heaviness relation in human EEG and fMRI

We tested whether differences in rated object heaviness have a positive relation to neural pattern differences after size, handling ease, other ratings, object offsets, source groups and a visual control. Separate public studies supplied scalp EEG from 10 adults and fMRI from 18 adults for 31 mapped object concepts. A local plan and code lock preceded the first heaviness fit. Mean partial correlations were +0.00407 and -0.01645. Their 95% person t intervals were [-0.04681, +0.05492] and [-0.04127, +0.00839]. Whole-object and crossed person-and-object bootstrap intervals included zero in each modality. The shared exact sign-flip p was 0.91114. The fixed shared rule failed. Raw mean correlations were more positive, +0.05088 and +0.07805; all results remain. A separate dense fit matched all 28 primary effects within 1.5 x 10^-14. This is a new endpoint in reused cohorts, not independent confirmation. Different images, tasks, cohorts, spatial sampling, source processing and concept-level norms limit the result. It does not show absent weight information, equal neural computations or a causal modality difference. The package gives licensed compact EEG inputs and norms, public fMRI source records and fetch code, all scores and draws, figures and repeat steps.

object perception · heaviness · EEG · fMRI · representational similarity · person and object uncertainty · negative result · public data