Time-Domain Heart Rate Variability Analysis as a Tool for Sleep Apnea Detection

• Main Authors: Jarwin-Jim Ang Tee, 鄭嘉輝
• Other Authors: Yuh-Show Tsai
• Format: Others
• Language: en_US
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/04701869442905345489

碩士 === 中原大學 === 生物醫學工程研究所 === 101 === Obstructive Sleep Apnea (OSA) is a syndrome in which there is a repeated event of a partial or complete obstruction of the upper airway during sleep, resulting in intermittent hypoxia and transient repetitive arousals from sleep. The characteristic heart rate pattern, known as the cyclic variation of heart rate (CVHR), that is known to accompany OSA episodes had been demonstrated in earlier studies to be an effective tool in the detection of OSA due to the high correlation between the CVHR index (CVHR per hour) and the apnea-hypopnea index. Moreover, Time- domain HRV analysis has been proven as powerful tool in definitive diagnosis and classification of OSAS by using R-wave detection to extract and analyze the RR intervals of ECG readings. In this study, the So and Chan algorithm for QRS detection was implemented along with time-domain HRV analysis in order to develop a system capable of deriving the required HRV characteristics for reliable diagnosis from ECG signals. The system was tested by using ECG recordings from Physionet’s Apnea-ECG database and also from ECG recorded using through the system. The results of the diagnosis from the Physionet data were then compared to the minute by minute classifications found in the Physionet database in order to test the reliability of the algorithm. The findings in the tests conducted have shown high accuracy, as high as 84% for recordings with severe apneas, and high sensitivity and specificity, around 90% and around 80% respectively. Real ECG data that was recorded using the National Instruments USB DAQ-6008 data acquisition device gave us similarly good results as with the analysis of the Physionet database.