Evidence Review
HRV Biofeedback: Meta-Analytic Evidence
The most comprehensive review to date is Lehrer et al. (2020), a systematic review and meta-analysis of 58 RCTs identified from 1,868 papers [1]. The analysis found a significant small-to-moderate effect size favoring HRV biofeedback versus controls across a range of outcomes. The largest effects were observed for anxiety, depression, anger, and performance (athletic and artistic); smaller effects were found for PTSD and sleep outcomes. The consistency of effects across independent RCTs with varied populations and settings strengthens confidence in the overall finding. The authors note HRV biofeedback is a "useful complementary treatment" — not a replacement for other interventions — and that it works best when combined with paced breathing at the individual's resonance frequency.
Stress and Anxiety
Ratanasiripong et al. (2012) conducted an RCT in 60 nursing students randomized to biofeedback intervention or waitlist control during a five-week clinical training period [2]. Students in the biofeedback group showed stable stress and anxiety scores while controls showed significant increases in both, with clear between-group divergence by week 5. The study used standardized measures (STAI, PSS) and had clean parallel-group design. Limitation: short duration and single population; generalizability to clinical anxiety disorders is uncertain.
Blood Pressure
Tsai et al. (2007) enrolled 38 participants with mild hypertension in a randomized controlled trial comparing active blood pressure biofeedback to sham biofeedback [3]. At 12-week follow-up, the active group showed a systolic BP reduction of 12.6 ± 8.8 mmHg vs. 4.1 ± 5.7 mmHg in sham controls (p=0.001). Mean arterial pressure reduction was also significantly greater (8.2 vs. 3.3 mmHg, p=0.017). The active group additionally showed improved skin conductance reactivity and reduced systolic BP stress reactivity, suggesting lasting autonomic regulation changes.
Nolan et al. (2010) tested behavioral neurocardiac training — an HRV-based biofeedback protocol — against autogenic relaxation in 65 hypertensive patients across 6 sessions [4]. Published in the high-impact journal Hypertension, this trial showed HRV biofeedback produced significant reductions in daytime systolic BP (−2.4 mmHg, p=0.009) and 24-hour ambulatory systolic BP (−2.1 mmHg, p=0.03) versus no significant change in the relaxation control. Heart rate modulation improvements were also observed. The effect size is modest but achieved with a remarkably brief 6-session protocol, suggesting continued training could produce larger effects.
Headache
Rausa et al. (2016) conducted a pilot RCT in 27 patients with medication overuse headache (MOH), a particularly difficult condition where the treatment itself becomes a trigger [5]. Patients were randomized to pharmacological treatment alone or pharmacotherapy plus biofeedback. By end of treatment, significantly more biofeedback patients had reverted from chronic to episodic headache status. The biofeedback group also showed greater reductions in headache frequency and analgesic intake, with effects sustained at 4-month follow-up. Limitation: small sample (N=27), pilot design; adequately powered trials are needed.
Biofeedback for non-medication headache and tension-type headache has been studied more extensively in the older literature and is considered an evidence-based approach by several clinical practice guidelines.
Neurofeedback and Cognitive Function
Hasslinger et al. (2022) ran a large multi-arm pragmatic RCT (N=202, ages 9–17) testing two neurofeedback protocols (slow cortical potential, live Z-score), working memory training, and treatment as usual in children with ADHD [6]. Neurofeedback produced some immediate post-treatment improvements in cognitive domains, but these were not fully sustained at follow-up. The study highlights an ongoing debate in the field: near-transfer (trained tasks improve) vs. far-transfer (daily function improves) effects are difficult to demonstrate consistently. The authors call for more personalized protocols. This area requires more rigorous, adequately powered trials before strong conclusions can be drawn.
Overall Assessment
Biofeedback has its strongest evidence base in HRV biofeedback for anxiety and stress (meta-analysis level, 58 RCTs), with meaningful but smaller evidence for hypertension and headache. The neurofeedback literature for ADHD and cognitive performance is promising but not definitive. Biofeedback is unlikely to cause harm and has no pharmacological side effects, making it a reasonable adjunct therapy for motivated patients willing to invest time in repeated sessions.