As described in my data gathering process, I sometimes track my heart rate during night, motivated by studies which clearly pointed out the correlation between sleep disorders and reduced heart rate variability [1, 2, 3].
In this very first experiment I tracked my heart rate during several nights with SweetBeat and processed the data collected with the KubiosHRV software. As a baseline for comparison, I relied on the study conducted in [2], which points out several significant time and frequency-domain parameters of HRV in patients suffering of Chronic Fatigue Syndrome (CFS). Since these parameters have been identified with a high statistical significance in comparing subjects with CFS and control ones, and computed with KubiosHRV as well, I believed it was interesting to compare with those obtained from my side.
The results I got are given on the Fig. 1. The first three columns come from the study conducted in [2], the last one presents the average of my HR parameters tracked over 19 nights (about 158 hours in total). And I can draw basically the same conclusions as in the study, my RMSSD and HF parameters are significantly reduced compared with healthy controls (which like in subjects with CFS would suggest a decrease in vagal modulation of heart rate as detailed in the article). I am a bit concerned by the results obtained, notably for my HF parameter which is significantly below the one from patients with CFS. Indeed, I am wondering if it could not come from the poor quality of my HR data collected during the night with my chest strap. Indeed, the signal used as input of KubiosHRV might be quite noisy, and in the software, I simply apply a "strong" artifact correction (from a feature embedded in KubiosHRV) and isolate the chunks of the input signal which are the less noisy, but I am wondering if more pre-processing would not be required (since in [4] is mentioned that "HRV is highly sensitive to artifact and errors in as low as even 2% of the data will result in unwanted biases in HRV calculations"). An example of results processed and generated by KubiosHRV for a night is given Fig. 2.
Fig. 1: Comparison between my HRV parameters and these presented in [2].
Fig. 2: Example of a result sheet generated by KubiosHRV.
References & Links
[1] Heart Rate Variability (HRV): Sleep Disordered Breathing. Homer Nazeran, Yvonne Pamula, Khosrow Behbehani. 2006.
[2] Reduced heart rate variability predicts poor sleep quality in a case-control study of chronic fatigue syndrome. Burton AR, Rahman K, Kadota Y, Lloyd A, Vollmer-Conna U. 2010.
[3] SleepImage Monitoring System. http://www.sleepimage.com.
[4] VivoSenseTM - User Manual – Heart Rate Variability (HRV) Professional Edition.