Cardiac autonomic state and neural manifold geometry co-vary non-monotonically across working-memory task phases
Description
Using multimodal EEG–ECG–PPG–pupillometry recordings from 62–65 healthy adults during a digit-span working-memory task, we show that cardiac autonomic state and neural manifold geometry are both systematically task-modulated, with a non-monotone, stage-shifted profile: the RMSSD-driven vagal proxy peaks during passive listening while neural trajectory efficiency peaks at the easiest active memory condition — a 24-fold rest-to-task difference. Direct autonomic–manifold coupling is modest and between-subject correlations are null, indicating co-modulation by task phase rather than causal autonomic gating.
Files
Cardiac autonomic state and neural manifold geometry co-vary non-monotonically across working-memory task phases.pdf
Files
(2.5 MB)
| Name | Size | |
|---|---|---|
|
md5:707c7acb5b1754ed314e81241b11519c
|
824.6 kB | Preview Download |
|
md5:25c46d826a6c28821d567768fea92617
|
225.9 kB | Preview Download |
|
md5:0d10e87cfb47e68ce8aff7495d4d6be8
|
605.8 kB | Preview Download |
|
md5:f73ac2997e5fe357fc6091690dedc895
|
361.3 kB | Preview Download |
|
md5:d501eface6a7ec4efb602e7d71e007bf
|
212.0 kB | Preview Download |
|
md5:4509a03a6b3743f982f340bd0ef474d7
|
277.1 kB | Preview Download |
Additional details
Related works
- Is supplemented by
- Working paper: 10.5281/zenodo.20112950 (DOI)
Software
- Repository URL
- https://github.com/NoeticDiffusion
- Programming language
- Python
- Development Status
- Active