Exploring brain-heart interactions: advances in physiological signal fusion for healthcare

The integration of cognitive neuroscience, wearable biosensing, and artificial intelligence (AI) opens new opportunities—and poses specific methodological challenges—for quantifying brain-heart interactions (BHIs). Single-modality approaches often lack the sensitivity required to resolve the multiscale, bidirectional dynamics that link neural activity and cardiac/autonomic signals. This review designed to present an operational framework that links neuroanatomical pathways, measurable physiological signals, analytic strategies, and clinical end-points. The primary focus is placed on a critical and parallel evaluation of analytic families for biological signal fusion, including linear spectral and phase metrics, directed linear models, information-theoretic estimators, network physiology measures, and machine-learning pipelines. The assumptions, strengths, limitations, and representative application scenarios of each family are systematically summarized. Disease-specific BHIs findings are further synthesized into a cross-disease perspective, highlighting recurrent transdiagnostic features. Building on these findings, we discuss the offer insights into future research directions. We hope that this survey not only garners greater attention from the scientific community but also guides multidisciplinary researchers and clinicians toward robust BHIs analyses and translational studies.