Sleep staging based on polysomnography (PSG) performed by human experts is the de facto “gold standard” for the objective measurement of sleep. PSG and manual sleep staging is, however, personnel-intensive and time-consuming and it is thus impractical to monitor a person’s sleep architecture over extended periods. Here, we present a novel, low-cost, automatized, deep learning alternative to PSG sleep staging that provides a reliable epoch-by-epoch four-class sleep staging approach (Wake, Light [N1 + N2], Deep, REM) based solely on inter-beat-interval (IBI) data. Having trained a multi-resolution convolutional neural network (MCNN) on the IBIs of 8898 full-night manually sleep-staged recordings, we tested the MCNN on sleep classification using the IBIs of two low-cost (<EUR 100) consumer wearables: an optical heart rate sensor (VS) and a breast belt (H10), both produced by POLAR®. The overall classification accuracy reached levels comparable to expert inter-rater reliability for both devices (VS: 81%, $\kappa$ = 0.69; H10: 80.3%, $\kappa$ = 0.69). In addition, we used the H10 and recorded daily ECG data from 49 participants with sleep complaints over the course of a digital CBT-I-based sleep training program implemented in the App NUKKUAA™. As proof of principle, we classified the IBIs extracted from H10 using the MCNN over the course of the training program and captured sleep-related changes. At the end of the program, participants reported significant improvements in subjective sleep quality and sleep onset latency. Similarly, objective sleep onset latency showed a trend toward improvement. Weekly sleep onset latency, wake time during sleep, and total sleep time also correlated significantly with the subjective reports. The combination of state-of-the-art machine learning with suitable wearables allows continuous and accurate monitoring of sleep in naturalistic settings with profound implications for answering basic and clinical research questions.

More and more people quantify their sleep using wearables and are becoming obsessed in their pursuit of optimal sleep (“orthosomnia”). However, it is criticized that many of these wearables are giving inaccurate feedback and can even lead to negative daytime consequences. Acknowledging these facts, we here optimize our previously suggested sleep classification procedure in a new sample of 136 self-reported poor sleepers to minimize erroneous classification during ambulatory sleep sensing. Firstly, we introduce an advanced interbeat-interval (IBI) quality control using a random forest method to account for wearable recordings in naturalistic and more noisy settings. We further aim to improve sleep classification by opting for a loss function model instead of the overall epoch-by-epoch accuracy to avoid model biases towards the majority class (i.e., “light sleep”). Using these implementations, we compare the classification performance between the optimized (loss function model) and the accuracy model. We use signals derived from PSG, one-channel ECG, and two consumer wearables: the ECG breast belt Polar® H10 (H10) and the Polar® Verity Sense (VS), an optical Photoplethysmography (PPG) heart-rate sensor. The results reveal a high overall accuracy for the loss function in ECG (86.3 %, $\kappa$ = 0.79), as well as the H10 (84.4%, $\kappa$ = 0.76), and VS (84.2%, $\kappa$ = 0.75) sensors, with improvements in deep sleep and wake. In addition, the new optimized model displays moderate to high correlations and agreement with PSG on primary sleep parameters, while measures of reliability, expressed in intra-class correlations, suggest excellent reliability for most sleep parameters. Finally, it is demonstrated that the new model is still classifying sleep accurately in 4-classes in users taking heart-affecting and/or psychoactive medication, which can be considered a prerequisite in older individuals with or without common disorders. Further improving and validating automatic sleep stage classification algorithms based on signals from affordable wearables may resolve existing scepticism and open the door for such approaches in clinical practice.

This study investigates the relationship between objective sleep indicators derived from heart rate variability (HRV) and sleep perception in young adults using an intensive longitudinal design over 14 consecutive nights. The sample included 178 participants aged 18-29 years from two separate studies, who provided data via daily sleep diaries. Using a multi-resolution convolutional neural network model, heart rate variability measured via ambulatory electrocardiography was used for sleep stage classification. Within-subject analyses revealed that longer total sleep time, higher sleep efficiency, more slow-wave sleep, and more rapid-eye-movement sleep were associated with better sleep perception, while longer wake after sleep onset was linked to poorer sleep perception. Notably, no significant associations were found on the between-subject level. The objective sleep indicators explained nearly five times as much variance in sleep perception at the within-subject level than at the between-subject level. Additionally, gender, as well as depressive symptoms and insomnia symptoms measured at baseline did not moderate the within-subject associations between objective sleep indicators and sleep perception. These findings underscore the importance of sufficient sleep duration, quality of sleep architecture, and sleep continuity for individuals' perceptions of their sleep on a nightly basis. The study's use of HRV-derived sleep staging over multiple nights represents a methodological strength, providing detailed and less intrusive assessment compared to traditional polysomnography. Furthermore, these results are particularly important for clinical applications, as they can be basis for individualized interventions to improve sleep perception.

More and more people quantify their sleep using wearables and are getting obsessed in their pursuit of optimal sleep (“orthosomnia”). However, it is criticized that many of these wearables are giving inaccurate feedback and even can lead to negative daytime consequences. Acknowledging these facts, we here aim to extend previous findings [] in a new sample of 136 self-reported poor sleepers by implementing optimization procedures to minimize erroneous classification when ambulatory sensing sleep. Firstly, here, we introduce an advanced interbeat-interval (IBI) quality control using a random forest method to account for wearable recordings in naturalistic and more noisy settings. We further aim to improve sleep classification by opting for a loss function model instead of overall epoch-by-epoch accuracy to avoid model biases towards the majority class (i.e., “light sleep”). Using these implementations, we compare the classification performance between the optimized (loss function model) and the accuracy model. We use signals derived from PSG, one-channel ECG, and two consumer wearables: the ECG breast belt Polar® H10 (H10) and the Polar® Verity Sense (VS), an optical Photoplethysmography (PPG) heart-rate sensor. Results reveal high overall accuracy for the new model for ECG (86.3 %, κ=.79), H10 (84.4%, κ=.76), and VS (84.2%, κ=.75) with intended improvements in deep sleep and wake. In addition, the new optimized model displays moderate to high correlations and agreement with PSG on primary sleep parameters, while measures of reliability, expressed in intra-class correlations, suggest excellent reliability for most sleep parameters. Finally, it is demonstrated that the new model is still classifying sleep accurately in 4-classes in users taking heart-affecting and/or psychoactive medication which can be considered a prerequisite for use also in high age groups and/or with common disorders. Further improving and validating sleep stage classification algorithms with affordable wearables may resolve existing scepticism and open the door for such approaches in clinical practice.

This study aimed to investigate the bidirectional relationships between affective wellbeing and sleep, and to explore the moderating effects of depressive and insomnia symptoms. Data from two studies, involving 178 participants in total aged 18–29 years, were analyzed. Over 14 days, participants completed daily surveys assessing affective wellbeing and wore ECG chest belts for heart rate variability monitoring to derive objective sleep indices. Multilevel models were used to examine within-subject associations between affective states and subsequent sleep, and vice versa, while considering depressive and insomnia symptoms as moderators. Higher positive affect than usual in the evening was associated with shorter total sleep time, shorter rapid eye movement sleep, and lower number of awakenings and stage shifts than usual in the following night. Regarding sleep predicting morning’s affective wellbeing, longer slow-wave sleep was linked to higher positive affect than usual, while longer rapid eye movement sleep and higher sleep efficiency than usual predicted lower negative affect than usual. Also, higher subjective sleep quality than usual was associated with higher positive and lower negative affect than usual the next morning. No evidence for moderation effects of insomnia and depressive symptoms for the bidirectional within-subject associations between sleep and affect was found. The findings underscore the complex interplay between sleep and affective wellbeing, particularly highlighting that the subjective perception of one’s sleep appears to be relevant for the next morning’s affective states.