For two fixed non-vocal sound panels, pair relations agreed across 15 iEEG patients and 12 fMRI adults after one training pattern was removed

We tested whether sound pair relations repeat across people after one training response pattern is removed. Separate human studies supplied signed intracranial EEG and BOLD fMRI. Each study used four reference people and four development people. Calibration sounds set site scales and the leading response direction. A fixed test used other non-vocal sounds in 15 patients and 12 adults. Category-pair means and individual-sound additive effects were removed from cosine similarities. Mean agreement was 0.09229 in iEEG and 0.28483 in fMRI. The 95% person t intervals were 0.05135 to 0.13323 and 0.24058 to 0.32907. The fixed person and common-label rule passed, conjunction p = 0.0036926. Resampling people, reference people and sounds gave ranges of -0.00682 to 0.16362 and 0.14289 to 0.32898. Thus new-sound transfer was not confirmed in iEEG. An early electrical window failed the sign rule. Four-site subset averages were small and positive; many individual subset scores were negative. A one-source model with independent noise also gave positive residual agreement. A population calculation shows that site scaling can keep one factor from aligning with the leading principal component. The result supports measured pair relations for two fixed panels. It does not identify several neural sources, a common cross-modal map or a cause. The package gives code, permitted electrical inputs, fixed source versions and terms, plans, all outcomes, figures and repeat instructions.

human neuroscience · intracranial EEG · fMRI · auditory cortex · sound geometry · measurement noise · principal component analysis · reproducibility