A fixed test did not confirm shared residual boundary response agreement in 18 patients with spikes and ten fMRI adults

We tested whether residual hippocampal responses to individual film boundaries agree across viewers in two separate human studies. Seven patients with intracranial spikes and five fMRI adults formed reference groups. A local plan fixed controls for durations, content changes, cuts and film time. Reserved tests used 18 other patients and ten other adults. Each film had its own events and response measure. Mean Fisher z agreement was +0.0344 in spikes and +0.1479 in BOLD. Scores were positive in 11/18 and 10/10 people. With reference people and events fixed, 95% person intervals were -0.0647 to +0.1334 and +0.0843 to +0.2114. Intervals that also varied reference viewers and common movie blocks were -0.1314 to +0.1671 and -0.0094 to +0.2403. The fixed shared rule failed; joint one-sided person sign p was 0.240341. All ten positive BOLD scores did not remove reference and event uncertainty. Fixed checks preserved the uncertain spike result. These are new-endpoint person holdouts in reused cohorts, not new-study confirmation. Sparse clinical sampling, different films and event labels, source processing, BOLD physiology and approximate block resampling limit scope. We do not claim absent event coding, a common neural mechanism, a causal modality difference or cross-modal generalization. The package preserves source tables, all scores and draws, figures, plans, code, source terms and repeat instructions. An independent method checked all 28 scores; fresh source extraction matched feature tables for two spike patients.

hippocampus · event boundaries · single neurons · fMRI · human neuroscience · reference uncertainty · cross-modal test · reproducibility