CookiesWe use cookies to enhance your experience and the functionality of our website. By continuing to browse, you are agreeing to our use of cookies. Learn More

CookiesWe use cookies on our website. By continuing to browse, you are agreeing to our use of cookies. Learn More

New Inner Balance™ Coherence Plus Sensor and App: Includes Free Lifetime Access to the New HeartMath® App Learn More

Research Library

Novel Methodological Tools for Behavioral Interventions: The Case of HRV‑Biofeedback. Sham Control and Quantitative Physiology‑Based Assessment of Training Quality and Fidelity

    • Published: 2021
    • Ewa Ratajczak1,2,3,5, Marcin Hajnowski4,5, Mateusz Stawicki3,5, and Włodzisław Duch2,3,5
    • Sensors 2021, 21, 3670. DOI: Institute of Psychology, Faculty of Philosophy and Social Sciences.2. Department of Informatics, Faculty of Physics, Astronomy and Informatics.3. Centre for Modern Interdisciplinary Technologies.4. Institute of Information and Communication Research, Faculty of Philosophy and Social Sciences.5. Nicolaus Copernicus University, Toruń, Poland.
    • Download the complete paper, click here.


Scientific research on heart rate variability (HRV) biofeedback is burdened by certain methodological issues, such as lack of consistent training quality and fidelity assessment or control conditions that would mimic the intervention. In the present study, a novel sham HRV-biofeedback training was proposed as a credible control condition, indistinguishable from the real training. The Yield Efficiency of Training Index (YETI), a quantitative measure based on the spectral distribution of heart rate during training, was suggested for training quality assessment. A training fidelity criterion derived from a two-step classification process based on the average YETI index and its standard deviation (YETISD) was suggested. We divided 57 young, healthy volunteers into two groups, each subjected to 20 sessions of either real or sham HRV-biofeedback. Five standard HRV measures (standard deviation of the NN (SDNN), root mean square of the standard deviation of the NN (RMSSD), total power, low-frequency (LF), and high-frequency (HF) power) collected at baseline, after 10 and 20 sessions were subjected to analysis of variance. Application of a training fidelity criterion improved sample homogeneity, resulting in a substantial gain in effect sizes of the group and training interactions for all considered HRV indices. Application of methodological amendments, including proper control conditions (such as sham training) and quantitative assessment of training quality and fidelity, substantially improves the analysis of training effects. Although presented on the example of HRV-biofeedback, this approach should similarly benefit other behavioral training procedures that interact with any of the many psychophysiological mechanisms in the human body.