Embedded realization of a real time Heart Rate Variability logger for at-home sleep studies

Sleep has no substitute and quality of sleep is a major concern for healthy living of human beings. Sleep breathing disorders are events characterized by pauses of breathing during sleep. Sleep breathing disorders like Obstructive Sleep Apnea (OSA) may result in cardiac disorders and fatalities. Though Polysomnography is considered as the gold standard for conducting sleep studies, current research directs that the trend of Heart Rate Variability (HRV) during sleep is indicative of sleep breathing disorder. Hence, reliable HRV recorders with ease of use may contribute to early screening of these disorders. This paper reports the prototype development of an embedded system for logging HRV during sleep for screening during sleep. The system is built with open source Arduino platform consisting of an ATMEGA328 microcontroller along with a provision for storage on a Secure-Digital card. 'R' peak detection is carried out using a combination of dynamic threshold and amplitude threshold. The logger is able to work on two modes: (1) plain, long duration logger and (2) HRV Logger. The estimated duration of logging is 72 hours with a +9 V battery supply. The system performance is compared with a commercially available Electrocardiogram (ECG) recorder system and a MATLAB based R peak detection system with real time recordings of 30 healthy adults. The system code is optimized to achieve a logging time of 6.25 milliseconds per sample and 0.98 seconds for each 'R' peak detection and storage. The proposed system was also tested with Sleep ECG samples from Physionet database and it achieved a maximum sensitivity of 97.7% and specificity of 95.56%. The maximum recorded percentage error of detection was 2%. The results indicate that the proposed system and software design can be developed as a compact, economical and portable device for early detection of sleep breathing disorders.