For 165 sounds, a calibrated norm cue failed a shared matching test in six SEEG patients and 24 fMRI adults

We tested whether response norm adds repeated sound matching value beyond pattern angle. Separate human studies supplied SEEG high-gamma activity and two fMRI cohorts. Development used two patients and four adults. A local code lock preceded tests in six other patients and 24 other adults. Each fold kept calibration sounds separate from test sounds. Tests used 165 sounds and ten 16-site panels. Mean gains were 0.744, 0.240 and 0.026 AUC percentage points. Their 95% person intervals were [0.156, 1.332], [0.137, 0.343] and [-0.034, 0.086]. Positive counts were 6/6, 15/16 and 5/8. The fixed shared rule for means and positive counts failed, p=0.363281. We do not claim a confirmed common benefit. All negative results and fixed checks remain. In a specified model, sound-dependent error variance gave a norm gain with only one common mean response pattern. Thus a matching gain need not identify a mean neural amplitude code. This is a processed-measurement benchmark, not a causal test. Source selection, small clinical sampling, separate cohorts, unequal filters and fixed sounds limit scope. The package gives code, new scores, source versions and terms, plans, figures, audits and repeat steps. Original neural arrays stay external.

human neuroscience · SEEG · fMRI · response norm · sound matching · measurement error · negative result