Accuracy of the Estimated Core Temperature (ECTemp) Algorithm in Estimating Circadian Rhythm Indicators

Abstract : While circadian rhythm analysis provides information critical to physiological status monitoring, there are considerable challenges to core body temperature (CT) measurement outside of stringent laboratory environments. This study evaluated ECTemp, a heart rate-based CT estimation algorithm, in the assessment of circadian rhythm indicators. Eleven participants (age, 23 +/- 3 years; height, 173.8 +/- 7.7 cm; body mass, 70.12 +/- 8.94 kg) were assessed on two occasions in which they were confined to a calorimeter chamber for a 22.5-hr period. Circadian rhythm indicators (MESOR, amplitude, and acrophase) were determined using a mixed effects cosinor regression models. ECTemp provided reasonable estimates of MESOR (RMSE, 0.17) and amplitude (RMSE, 0.10). While ECTemp estimates of circadian rhythm indicators were lower than observed CT, the differences were smaller than heart-rate based models analyzed in previous studies. As such, ECTemp demonstrates strong potential for estimating circadian CT rhythm indicators, particularly if the algorithm is updated to fit additional data from periods of low CT and heart rate.