Representation Learning for Physiological Signals and Multi-Region Neural Dynamics

Abstract

Biomedical Signal Representation has become a test of how well researchers can connect performance with evidence quality, resource limits, and transfer across settings. The discussion is organized around separating clinically meaningful dynamics from nuisance variation in respiratory and neural measurements. The evidence base combines 2 focal papers with 11 independently retrieved publications reviewed against traceable publication metadata. The analysis is organized around signal decomposition, latent disentanglement, temporal coupling, interpretability, and clinical validation. The comparison does not regard published metrics as automatically comparable, the review compares problem definitions, methodological assumptions, and validation boundaries. Across the literature, the central lesson is that advances in biomedical signal representation become credible when measurement, model, and clinical decision are evaluated together and when uncertainty about calibration is reported explicitly. The synthesis ties method selection to operational consequence while identifying external-validity hazards, and proposes a research agenda centered on traceable baselines, scenario testing, and reusable evidence.