Breathe better, live better: the science of slow breathing and heart rate variability.
Level 5 - mechanism / opinion, no new human data
Scoping review of six studies without formal systematic review grading or meta-analysis
PubMed 40252198 · doi:10.1007/s13760-025-02789-w
What was done
A scoping review synthesized findings from six studies identified through databases including MEDLINE (PubMed), Scopus, and Web of Science. The studies evaluated healthy adults aged 18 to 60 years free from major cardiovascular, respiratory, or neurological disorders. Interventions included slow breathing and heart rate variability (HRV) biofeedback delivered either acutely (single or short-term sessions) or chronically (daily sessions over 4 to 8 weeks). Evaluated outcomes included HRV parameters, respiratory sinus arrhythmia (RSA), baroreflex sensitivity, and neural synchronization.
What was found
All six studies reported statistically significant improvements in HRV metrics, particularly in the high-frequency (HF) band reflecting increased parasympathetic activity. RSA increased in protocols emphasizing tailored breathing patterns with longer exhalations. Baroreflex sensitivity improved in interventions combining slow breathing with HRV biofeedback. One study reported neural synchronization between cortical potentials and HRV during slowed respiration. The abstract reports directional improvements across all six studies but provides no quantitative effect sizes, variances, or p-values.
Why it matters
The findings synthesize evidence that structured slow breathing and HRV biofeedback acutely and chronically improve autonomic regulation and parasympathetic tone in healthy individuals. This supports the utility of respiratory pacing as a non-invasive tool for autonomic flexibility.
Limits
The review included only six studies, and the total participant count across studies was not reported in the abstract. Results are restricted to healthy adults aged 18–60 and cannot be directly generalized to clinical populations with chronic conditions. Interventions varied substantially in duration and delivery method, and no numerical effect sizes or risk-of-bias assessments were detailed.