Tests in seven ECoG patients and eight fMRI adults did not establish shared coding of individual stimuli

We tested whether responses to individual stimuli repeat after category means and the spatial mean are removed. Public data came from separate visual ECoG and auditory fMRI studies. A local plan fixed one shared test before neural outcomes. The mean repeat score Q was 0.024587 in seven ECoG patients (95% person interval 0.007677 to 0.043080) and 0.167217 in eight fMRI adults (0.141784 to 0.196546). The fixed numerical gate failed; its shared person-sign p was 0.226562. However, all 7 ECoG records repeated the same 100-image order in three runs. Image and position indicator matrices each had rank 100. Their combined rank was also 100. A position-only model gave Q=1 after the same controls. Thus the ECoG score cannot separate image identity from sequence position or preceding-image context. The fMRI test used scan 3 after scans 1 and 2 supplied calibration. It does not remove the electrical design limit. These data do not establish a shared exemplar code or a common neural cause. The contribution is a fixed cross-study response benchmark with an exact source-design audit. The package keeps code, source versions and terms, all person and control results, uncertainty draws, figures, and repeat steps. Source data stay external.

ECoG · fMRI · stimulus identity · fixed order · repeatability · source audit