Development of non-invasive thermometry for continuous body temperature monitoring

Abstract

Body temperature is a vital sign for the assessment of physiological functions. Although other vital signs such as pulse, respiratory rate, and blood pressure can be noninvasively monitored, body temperature is currently being evaluated in an invasive way. Generally, body temperature monitoring is utilized to detect the fever onset, to diagnose the infection, and to prevent hypothermia during operation in clinics. However, as recent works have found that circadian body temperature rhythm is associated with various health conditions such as sleep diseases, mental performance, and female menstrual cycle, the necessity of body temperature monitoring in daily life is increased. Body temperature monitoring during daily life should be noninvasive, comfortable, and economical. And a number of novel techniques have been proposed for noninvasive body temperature monitoring. In this thesis, developments related to a deep body thermometer and core temperature estimation model were proposed. One noninvasive thermometry for body temperature monitoring is the double sensor method-based thermometer. The thermometer is a promising technology that may be applicable to body temperature monitoring in daily life. Despite its considerable potential for body temperature monitoring, its key design features have not been investigated. Therefore, the structural and thermophysical effects on the performance of a thermometer were analyzed using finite element analysis in terms of accuracy, initial waiting time, and the ability to track changes in body temperature. As a result, when the thermometer was composed of an aluminum cover and foam insulator, the accuracy and initial waiting time was improved. In addition, the measurement accuracy was improved when the probes area was larger and height was lower. Our findings may provide thermometer manufacturers with new insights into probe design and help them fabricate thermometers optimized for specific applications. Another noninvasive thermometry for body temperature monitoring involves developing a core temperature estimation model based on environmental or physiological variables. As heart rate is a physiological signal relevant to thermoregulation, the usefulness of heart rate monitoring for estimation of circadian body temperature rhythm was investigated. The performance of regression analysis and the extended Kalman filter were evaluated. As a result, mean R-R interval (RRI) and mean heart rate (MHR) showed better performance (mean RMSE: 0.35 ± 0.10 ˚C) than other heart rate variability (HRV) parameters. Also, the extended Kalman filter, combined with RRI or MHR, provided better accuracy than regression analysis in terms of amplitude and acrophase estimation (amplitude estimation, RMSE: 0.25 ˚C vs. 0.28 ˚C, acrophase estimation, RMSE: 1.11 h vs. 1.59 h). Although the usefulness of heart rate monitoring in circadian body temperature rhythm estimation was verified in healthy subjects, the validity of the model for clinical populations in which heart rate responses and thermoregulation are different from the healthy people should be verified. The noninvasive temperature monitoring techniques developed in this thesis are expected to contribute to obtaining various health-related information (e.g., infection, circadian rhythm, sleep diseases, female menstrual cycle) during daily life in a convenient, economical, and safe way.