Human Circadian Phase Estimation from Signals Collected in Ambulatory Conditions Using an Autoregressive Model

Abstract

Phase estimation of the human circadian rhythm is a topic that has been explored using various modeling approaches. The current models range from physiological to mathematical, all attempting to estimate the circadian phase from different physiological or behavioral signals. Here, we have focused on estimation of the circadian phase from unobtrusively collected signals in ambulatory conditions using a statistically trained autoregressive moving average with exogenous inputs (ARMAX) model. Special attention has been given to the evaluation of heart rate interbeat intervals (RR intervals) as a potential circadian phase predictor. Prediction models were trained using all possible combinations of RR intervals, activity levels, and light exposures, each collected over a period of 24 hours. The signals were measured without any behavioral constraints, aside from the collection of saliva in the evening to determine melatonin concentration, which was measured in dim-light conditions. The model was trained and evaluated using 2 completely independent datasets, with 11 and 19 participants, respectively. The output was compared to the gold standard of circadian phase: dim-light melatonin onset (DLMO). The most accurate model that we found made use of RR intervals and light and was able to yield phase estimates with a prediction error of 2 ± 39 minutes (mean ± SD) from the DLMO reference value.

Effects of Two 15-min Naps on the Subjective Sleepiness, Fatigue and Heart Rate Variability of Night Shift Nurses

“…Humans have biological clocks that follow a circadian rhythm of approximately 24 h. The circadian rhythm may be seen in patterns of sleeping and waking 8 ) , secretion of hormones such as melatonin 9 , 10 ) and in body temperature 8 ) . Normally, patterns of sleeping and waking are synchronized with internal rhythms such as body temperature.…”

An Optimal Time for Treatment—Predicting Circadian Time by Machine Learning and Mathematical Modelling

“…To ensure a sufficient amount of information, two different types of data can be used, low-dimensional data with a high temporal resolution or high-dimensional data with a lower temporal resolution. High temporal resolution can be gained by wearable technical devices that continuously measure physiological data [ 123 , 124 , 125 , 126 , 127 , 128 ]. The high temporal resolution seems to balance the low number of features measured, for example, for temperature and light exposure data, in contrast to RNA sequencing data with thousands of transcripts measured.…”

“…Though these approaches result in relatively precise prediction of the circadian time [ 127 , 128 ], the need of special devices has so far limited their applicability in a clinical context. The availability of high-resolution data appropriate for predictions of circadian time will dramatically increase now that patients can acquire their own devices that track physiological parameters, such as smart watches.…”

“…Besides genomic data, many studies rely on indirect measurements like ambulatory recordings of wrist actigraphy and DLMO to predict circadian time and evaluate misalignments [ 175 , 176 ]. Several studies applied indirect and non-invasive measurements to develop a predictive model of phase [ 123 , 124 , 125 , 126 , 127 , 128 , 177 ]. Autoregressive moving average with exogenous inputs (ARMAX) model was developed using a supervised learning approach and predicted the DLMO phase using 24 h data obtained from ambulatory recordings [ 127 ].…”

“…Several studies applied indirect and non-invasive measurements to develop a predictive model of phase [ 123 , 124 , 125 , 126 , 127 , 128 , 177 ]. Autoregressive moving average with exogenous inputs (ARMAX) model was developed using a supervised learning approach and predicted the DLMO phase using 24 h data obtained from ambulatory recordings [ 127 ]. Such recordings included wrist actigraphy, light exposure, electrocardiograms (ECG) and proximal/distal skin temperature data from healthy subjects.…”