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

Published November 3, 2026. In a secondary analysis of 32 healthy adults, eight weeks of HRV biofeedback was associated with changes in selected directional connections within the central autonomic network compared with an active video-game control. The study helps explore how biofeedback may affect brain-heart signaling, but it did not test symptom improvement or treatment effectiveness.

Ma L, Steinberg JL, Eddie D, Hardy A, Zuniga EA, Vassileva J, Bjork JM, Snyder AD, Keen II LD, Gonzalez R, Abbate A, Moeller FG.
Applied Psychophysiology and Biofeedback · Publication date: 2026-11-03
Added to site: 2026-11-05
DOI: 10.1007/s10484-026-09818-1

Peer-reviewed version of record published November 3, 2026. An earlier preprint was posted September 20, 2026 and is the version currently indexed in PubMed. Added to Neurofeedback.com November 5, 2026.

Modalities

HRV biofeedback, fMRI, Spectral dynamic causal modeling

Clinical focus

Brain-heart interactions, Central autonomic network, Biofeedback mechanisms

Population

Healthy adults in a previously collected randomized biofeedback dataset

Sample

32 participants: HRV biofeedback n=15 and video-game active control n=17

Intervention / method

Eight weeks of training, five sessions per week. The HRV-biofeedback group used individually estimated resonance-frequency breathing and heart-rate feedback; the control group followed the same schedule with mobile video games.

Comparator

Active video-game control, not sham biofeedback or a respiration-matched control

Clinician summary

Spectral dynamic causal modeling found group differences in selected medial-prefrontal, cingulate, thalamic, hippocampal and ventrolateral-prefrontal connections. The hippocampus-to-right-VLPFC connectivity change showed an intervention-dependent association with SDNN change. Top-down pathway changes were not associated with HRV change, RMSSD improvement was modest and nonsignificant, and the analysis was restricted to selected network nodes.

How to interpret the findings

This was a small secondary imaging analysis in healthy adults using an existing dataset. The active control addressed engagement but did not match breathing, autonomic practice or biofeedback expectancy. Only selected central-autonomic-network regions were modeled, RMSSD change was not significant, and no clinical symptoms or patient outcomes were tested. Directional-connectivity modeling is mechanistic evidence, not proof that a pathway changed in a way that benefits health or that results generalize to clinical populations.

Equipment and disclosures

The final article identifies smartphone heart-rate feedback, PPG recording during fMRI and analysis software. The record does not attribute the biofeedback system to BrainMaster or establish superiority for any commercial product.

Supported by the NIH National Center for Advancing Translational Sciences under award UM1TR004360. Moeller disclosed consulting and research funding from Boehringer Ingelheim; Abbate disclosed consulting for Kiniksa, Monte Rosa and Novo Nordisk; Eddie disclosed scientific-advisory roles at ViviHealth and Innerworld and a partnership in Peer Recovery Consultants. Other authors reported no conflicts. No correction, retraction or expression-of-concern relation appeared in the bounded Crossref, PubMed, PMC and DOI/title checks on November 5, 2026. AI-assisted source review for Neurofeedback.com, not independent human clinical review.

Related evidence

Compare findings from the same clinical area, including mixed results.

• Effects of daily heart rate variability biofeedback practice on the amygdala's structural covariance network. · Positive with limitations