Fixed tests did not confirm shared positive session gaps or object label associations in 6 EEG and 9 fMRI adults

We tested repeated responses to up to 102 fixed videos in separate human EEG and fMRI studies. A local code lock preceded tests in 6 EEG adults and 9 other fMRI adults. One MRI adult supplied development. A matched score compared pairs of single trials within sessions with four pairs across sessions. Mean within-minus-across cosines were +0.00030 and +0.00885. Their 95% person Student intervals were [-0.01609 to +0.01670] and [+0.00335 to +0.01435]. The shared gap rule failed; its two-claim Holm p was 0.9637496. A separate test related object-label agreement to consistency of session-group means. It controlled fixed label, motion and repeat-count terms. Adjusted group correlations were +0.07952 in EEG and -0.08115 in MRI. Its shared positive rule failed, Holm p=0.9922079. All negative person scores and all checks remain. Video-and-person resampling gives gap ranges [-0.02154 to +0.02104] and [+0.00116 to +0.01621]. Both label-association ranges include zero under this resampling. A source-event audit checks session boundaries. Selected development values are checked against estimates before source organization. A fixed latent-pattern model gave a gap from gains of final pattern features alone. Thus a measured gap does not identify neural drift. Source fitting, calibration, trial order, attention, sound, spatial support and cohort differences limit scope. The package gives compact Gram data, source pins and terms, code, all results, figures and repeat steps. The score is an established two-sample construction; the contribution is this cross-study test and benchmark.

EEG · fMRI · video responses · session consistency · repeatability · object labels · maximum mean discrepancy · open data