Search

The transition from wakefulness to sleep occurs when the core body temperature decreases. The latter is facilitated by an increase in the cutaneous blood flow, which dissipates internal heat into the micro-environment surrounding the sleeper’s body. The rise in cutaneous blood flow near sleep onset causes the distal (hands and feet) and proximal (abdomen) temperatures to increase by about 1 °C and 0.5 °C, respectively. Characterizing the dynamics of skin temperature changes throughout sleep phases and understanding its relationship with sleep quality requires a means to unobtrusively and longitudinally estimate the skin temperature. Leveraging the data from a temperature sensor strip (TSS) with five individual temperature sensors embedded near the surface of a smart bed’s mattress, we have developed an algorithm to estimate the distal skin temperature with a minute-long temporal resolution. The data from 18 participants who recorded TSS and ground-truth temperature data from sleep during 14 nights at home and 2 nights in a lab were used to develop an algorithm that uses a two-stage regression model (gradient boosted tree followed by a random forest) to estimate the distal skin temperature. A five-fold cross-validation procedure was applied to train and validate the model such that the data from a participant could only be either in the training or validation set but not in both. The algorithm verification was performed with the in-lab data. The algorithm presented in this research can estimate the distal skin temperature at a minute-level resolution, with accuracy characterized by the mean limits of agreement [−0.79 to +0.79 °C] and mean coefficient of determination R2=0.87. This method may enable the unobtrusive, longitudinal and ecologically valid collection of distal skin temperature values during sleep. Therelatively small sample size motivates the need for further validation efforts.

IntroductionInsomnia causes serious adverse health effects and is estimated to affect 10–30% of the worldwide population. This study leverages personalized fine-tuned machine learning algorithms to detect insomnia risk based on questionnaire and longitudinal objective sleep data collected by a smart bed platform.MethodsUsers of the Sleep Number smart bed were invited to participate in an IRB approved study which required them to respond to four questionnaires (which included the Insomnia Severity Index; ISI) administered 6 weeks apart from each other in the period from November 2021 to March 2022. For 1,489 participants who completed at least 3 questionnaires, objective data (which includes sleep/wake and cardio-respiratory metrics) collected by the platform were queried for analysis. An incremental, passive-aggressive machine learning model was used to detect insomnia risk which was defined by the ISI exceeding a given threshold. Three ISI thresholds (8, 10, and 15) were considered. The incremental model is advantageous because it allows personalized fine-tuning by adding individual training data to a generic model.ResultsThe generic model, without personalizing, resulted in an area under the receiving-operating curve (AUC) of about 0.5 for each ISI threshold. The personalized fine-tuning with the data of just five sleep sessions from the individual for whom the model is being personalized resulted in AUCs exceeding 0.8 for all ISI thresholds. Interestingly, no further AUC enhancements resulted by adding personalized data exceeding ten sessions.DiscussionThese are encouraging results motivating further investigation into the application of personalized fine tuning machine learning to detect insomnia risk based on longitudinal sleep data and the extension of this paradigm to sleep medicine.