A fixed covariance model predicted repeat change directions across sound groups in selected SEEG and fMRI records
We tested a model of repeat change directions across sound groups. The model used 128 reference sounds to predict 28 speech and song clips. Separate public studies supplied SEEG high-gamma responses and B2021 and NH2015 fMRI BOLD cohorts. Two patients and four adults supplied development. A local code lock preceded tests in six other patients and 24 other adults. Each fit used 16 sites. Double differences removed stable item means and additive repeat offsets. The angular score ignored target vector size. Mean gains over a diagonal model were 0.09780, 0.10464, 0.26063 bits per site. Their 95 percent person intervals were [0.06643, 0.12917], [0.09296, 0.11633], [0.22518, 0.29608]. Positive counts were 6/6, 16/16 and 8/8. The one fixed shared mean-and-majority test passed, p=0.015625. The SEEG pre-onset check also had a positive mean, 0.02679. Models with no fit to people gave positive scores from common measurement mixing and from repeat gain. These examples had independent latent inputs or no random error. Thus the result tests prediction in selected processed records. It does not identify intrinsic neural noise, equal covariance matrices or a cause. Source selection, separate cohorts, unequal response estimates and fixed sounds limit scope. The package gives code, source pins and terms, all results, figures, audits and repeat steps. Original neural arrays stay external.
human neuroscience · SEEG · fMRI · repeat variability · directional statistics · covariance transfer · public data
References
10- usesA repeat difference metric improved matching of 165 sounds in six SEEG patients and 24 fMRI adults sci:sha256:980d42dfb364faf26f60b98bf74a18d48bbbac309b59e3cf8f482465aca31c77
We reuse MIT source preparation and exact data pins. We test a new direction score.
- usesFixed covariance tests did not confirm vocal matching gains across six SEEG patients and two fMRI studies sci:sha256:3fb019d6df76adecdd8542bead0a7e258bc0043b5c6c6c5a7534f0fb4be8f545
Failed sound matching tests motivate this direction prediction test. We retain the earlier negative matching result.
- doi:10.1371/journal.pcbi.1012092
- doi:10.1093/biomet/74.3.579
- doi:10.1371/journal.pbio.3001675
- doi:10.5281/zenodo.6544488
- doi:10.1152/jn.00588.2020
- doi:10.1016/j.neuron.2015.11.035
- doi:10.1371/journal.pbio.3002366
- doi:10.5281/zenodo.8349727
Cited by
5 publicationsPublic catalog entries on Sciscape.
- supportsAt 24 sites, one repeat mode lowered error in six SEEG patients and two fMRI panels; a zero residual pattern gave lower error
Related evidence for structured measured repeat changes. The present error projection is a different endpoint.
- Fixed maps from eight sites gave higher prediction gains within a repeat in six SEEG patients and two fMRI studies
Found in the final catalog check after scoring. Prior prediction of repeat-change directions; it limits claims of first repeat-structure discovery.
- With 165 sounds per study, matching repeats gave higher spatial forecast scores in six SEEG patients and eight fMRI adults
Earlier work predicted repeat-change directions across sound groups. The new score holds response-fitted weights fixed and tests repeat assignment.
- usesAt fixed total weight, nonvoice filters failed a shared speech and song test in six SEEG patients and 24 fMRI adults
We reuse its MIT source preparation and exact source pins.
- Vocal ranks from repeat changes exceeded chance in six SEEG patients and two fMRI studies, but the fixed baseline rule failed
Found in final catalog review. It predicts change directions across sound groups; it does not classify sound labels. The source cohorts are shared.