Directional Brain Connectivity Changes Induced by Heart Rate Variability Biofeedback: A Spectral Dynamic Causal Modeling Study

Abstract

Heart rate variability biofeedback (HRVB) is a non-pharmacological intervention that trains individuals to breathe at their cardiovascular resonance frequency to enhance autonomic regulation. While HRVB improves cardiovascular function across clinical conditions, the directional neural pathways through which it modulates the central autonomic network (CAN) remain poorly understood. Using a publicly available open-access dataset, we conducted a secondary analysis of effective (directional) neural connectivity (EC) data from 32 healthy participants randomized to an 8-week HRVB intervention (n = 15) or a video game Control intervention (n = 17). Pre- and post-intervention resting-state functional magnetic resonance imaging (fMRI) data were analyzed using spectral dynamic causal modeling (DCM) combined with parametric empirical Bayes to examine group differences in EC changes (ΔEC). We also evaluated whether connectivity changes correlated with changes in peripheral autonomic regulation, indexed by the standard deviation of normal-to-normal intervals (ΔSDNN). Compared with the Control group, HRVB showed strong evidence (posterior-probability = 1) of increased ΔEC from the medial prefrontal cortex to the anterior cingulate cortex and bilateral thalamus post-treatment, along with reduced ΔEC from the right hippocampus to the right ventrolateral prefrontal cortex (VLPFC). Furthermore, the regression coefficient linking hippocampus-to-VLPFC ΔEC with ΔSDNN was more (posterior-probability = 1) negative in the HRVB group than in Controls, indicating an intervention-dependent association between autonomic changes and effective connectivity. These findings indicate that HRVB alters directional fronto-limbic communication within the CAN, providing a circuit-level framework for understanding how resonance breathing modulates central autonomic control.