skip to main content
US FlagAn official website of the United States government
dot gov icon
Official websites use .gov
A .gov website belongs to an official government organization in the United States.
https lock icon
Secure .gov websites use HTTPS
A lock ( lock ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

Explore Research Products in the PAR
It may take a few hours for recently added research products to appear in PAR search results.


  • NSF Public Access
  • Search Results
  • The Combination of Topological Data Analysis and Mathematical Modeling Improves Sleep Stage Prediction From Consumer-Grade Wearables
Title: The Combination of Topological Data Analysis and Mathematical Modeling Improves Sleep Stage Prediction From Consumer-Grade Wearables
Wearable devices have become commonplace tools for tracking behavioral and physiological parameters in real-world settings. Nonetheless, the practical utility of these data for clinical and research applications, such as sleep analysis, is hindered by their noisy, large-scale, and multidimensional characteristics. Here, we develop a neural network algorithm that predicts sleep stages by tracking topological features (TFs) of wearable data and model-driven clock proxies (CPs) reflecting the circadian propensity for sleep. To evaluate its accuracy, we apply it to motion and heart rate data from the Apple Watch worn by young subjects undergoing polysomnography (PSG) and compare the predicted sleep stages with the corresponding ground truth PSG records. The neural network that includes TFs and CPs along with raw wearable data as inputs shows improved performance in classifying Wake/REM/NREM sleep. For example, it shows significant improvements in identifying REM and NREM sleep (AUROC/AUPRC improvements >13% and REM/NREM accuracy improvement of 12%) compared with the neural network using only raw data inputs. We find that this improvement is mainly attributed to the heart rate TFs. To further validate our algorithm on a different population, we test it on elderly subjects from the Multi-ethnic Study of Atherosclerosis cohort. This confirms that TFs and CPs contribute to the improvements in Wake/REM/NREM classification. We next compare the performance of our algorithm with previous state-of-the-art wearable-based sleep scoring algorithms and find that our algorithm outperforms them within and across different populations. This study demonstrates the benefits of combining topological data analysis and mathematical modeling to extract hidden inputs of neural networks from puzzling wearable data.  more » « less
Award ID(s):
2052499
PAR ID:
10555635
Author(s) / Creator(s):
 ;  ;  ;  ;  
Publisher / Repository:
SAGE Publications
Date Published:
Journal Name:
Journal of Biological Rhythms
Volume:
39
Issue:
6
ISSN:
0748-7304
Format(s):
Medium: X Size: p. 535-553
Size(s):
p. 535-553
Sponsoring Org:
National Science Foundation
More Like this
  1. ; ; ; ; (, eLife)
    null (Ed.)
    Sleep has many roles, from strengthening new memories to regulating mood and appetite. While we might instinctively think of sleep as a uniform state of reduced brain activity, the reality is more complex. First, over the course of the night, we cycle between a number of different sleep stages, which reflect different levels of sleep depth. Second, the amount of sleep depth is not necessarily even across the brain but can vary between regions. These sleep stages consist of either rapid eye movement (REM) sleep or non-REM (NREM) sleep. REM sleep is when most dreaming occurs, whereas NREM sleep is particularly important for learning and memory and can vary in duration and depth. During NREM sleep, large groups of neurons synchronize their firing to create rhythmic waves of activity known as slow waves. The more synchronous the activity, the deeper the sleep. Vaidyanathan et al. now show that brain cells called astrocytes help regulate NREM sleep. Astrocytes are not neurons but belong to a group of specialized cells called glia. They are the largest glia cell type in the brain and display an array of proteins on their surfaces called G-protein-coupled receptors (GPCRs). These enable them to sense sleep-wake signals from other parts of the brain and to generate their own signals. In fact, each astrocyte can communicate with thousands of neurons at once. They are therefore well-poised to coordinate brain activity during NREM sleep. Using innovative tools, Vaidyanathan et al. visualized astrocyte activity in mice as the animals woke up or fell asleep. The results showed that astrocytes change their activity just before each sleep–wake transition. They also revealed that astrocytes control both the depth and duration of NREM sleep via two different types of GPCR signals. Increasing one of these signals (Gi-GPCR) made the mice sleep more deeply but did not change sleep duration. Decreasing the other (Gq-GPCR) made the mice sleep for longer but did not affect sleep depth. Sleep problems affect many people at some point in their lives, and often co-exist with other conditions such as mental health disorders. Understanding how the brain regulates different features of sleep could help us develop better – and perhaps more specific – treatments for sleep disorders. The current study suggests that manipulating GPCRs on astrocytes might increase sleep depth, for example. But before work to test this idea can begin, we must first determine whether findings from sleeping mice also apply to people. 
    more » « less
  2. ; ; ; ; ; ; ; ; ; ; et al (, Sleep)
    null (Ed.)
    Abstract Study Objectives The usage of wrist-worn wearables to detect sleep–wake states remains a formidable challenge, particularly among individuals with disordered sleep. We developed a novel and unbiased data-driven method for the detection of sleep–wake and compared its performance with the well-established Oakley algorithm (OA) relative to polysomnography (PSG) in elderly men with disordered sleep. Methods Overnight in-lab PSG from 102 participants was compared with accelerometry and photoplethysmography simultaneously collected with a wearable device (Empatica E4). A binary segmentation algorithm was used to detect change points in these signals. A model that estimates sleep or wake states given the changes in these signals was established (change point decoder, CPD). The CPD’s performance was compared with the performance of the OA in relation to PSG. Results On the testing set, OA provided sleep accuracy of 0.85, wake accuracy of 0.54, AUC of 0.67, and Kappa of 0.39. Comparable values for CPD were 0.70, 0.74, 0.78, and 0.40. The CPD method had sleep onset latency error of −22.9 min, sleep efficiency error of 2.09%, and underestimated the number of sleep–wake transitions with an error of 64.4. The OA method’s performance was 28.6 min, −0.03%, and −17.2, respectively. Conclusions The CPD aggregates information from both cardiac and motion signals for state determination as well as the cross-dimensional influences from these domains. Therefore, CPD classification achieved balanced performance and higher AUC, despite underestimating sleep–wake transitions. The CPD could be used as an alternate framework to investigate sleep–wake dynamics within the conventional time frame of 30-s epochs. 
    more » « less
  3. ; ; ; ; (, Bioengineering)
    A large number of human intracranial EEG (iEEG) recordings have been collected for clinical purposes, in institutions all over the world, but the vast majority of these are unaccompanied by EOG and EMG recordings which are required to separate Wake episodes from REM sleep using accepted methods. In order to make full use of this extremely valuable data, an accurate method of classifying sleep from iEEG recordings alone is required. Existing methods of sleep scoring using only iEEG recordings accurately classify all stages of sleep, with the exception that wake (W) and rapid-eye movement (REM) sleep are not well distinguished. A novel multitaper (Wake vs. REM) alpha-rhythm classifier is developed by generalizing K-means clustering for use with multitaper spectral eigencoefficients. The performance of this unsupervised method is assessed on eight subjects exhibiting normal sleep architecture in a hold-out analysis and is compared against a classical power detector. The proposed multitaper classifier correctly identifies 36±6 min of REM in one night of recorded sleep, while incorrectly labeling less than 10% of all labeled 30 s epochs for all but one subject (human rater reliability is estimated to be near 80%), and outperforms the equivalent statistical-power classical test. Hold-out analysis indicates that when using one night’s worth of data, an accurate generalization of the method on new data is likely. For the purpose of studying sleep, the introduced multitaper alpha-rhythm classifier further paves the way to making available a large quantity of otherwise unusable IEEG data. 
    more » « less
  4. ; ; ; ; ; ; ; (, SPIE)
    Kim, Jaehwan (Ed.)
    Measuring and analyzing local field potential (LFP) signals from basolateral amygdala (BLA), hippocampus (HPC) and medial prefrontal cortex (mPFC) may help understand how they communicate with each other during fear memory formation and extinction. In our research, we have formulated a computationally simple and noise immune instantaneous amplitude cross correlation technique which can deduce lead and lag of LFPs generated in BLA, HPC, and mPFC and the directionality of brain signals exchanged between regions. LFP signals are recorded using depth electrodes in the rat brain and cross correlation analysis is applied to theta wave signals after filtering. We found that rats resilient to traumatic conditions (based on post-stress rapid eye movement sleep (REM)) showed a decrease in LFP signal correlation in REM and non-REM (NREM) sleep cycles between BLA-HPC regions after shock training and one day post shock training compared to vulnerable rats that show stress-induced reductions in REM. It is presumed this difference in neural network behavior may be related to REM sleep differences between resilient and vulnerable rats and may provide clues to help understand how traumatic conditions are processed by the brain. 
    more » « less
  5. ; ; ; ; ; ; (, Sleep)
    null (Ed.)
    Abstract Study Objectives We determine if young people with narcolepsy type 1 (NT1), narcolepsy type 2 (NT2), and idiopathic hypersomnia (IH) have distinct nocturnal sleep stability phenotypes compared to subjectively sleepy controls. Methods Participants were 5- to 21-year old and drug-naïve or drug free: NT1 (n = 46), NT2 (n = 12), IH (n = 18), and subjectively sleepy controls (n = 48). We compared the following sleep stability measures from polysomnogram recording between each hypersomnolence disorder to subjectively sleepy controls: number of wake and sleep stage bouts, Kaplan–Meier survival curves for wake and sleep stages, and median bout durations. Results Compared to the subjectively sleepy control group, NT1 participants had more bouts of wake and all sleep stages (p ≤ .005) except stage N3. NT1 participants had worse survival of nocturnal wake, stage N2, and rapid eye movement (REM) bouts (p < .005). In the first 8 hours of sleep, NT1 participants had longer stage N1 bouts but shorter REM (all ps < .004). IH participants had a similar number of bouts but better survival of stage N2 bouts (p = .001), and shorter stage N3 bouts in the first 8 hours of sleep (p = .003). In contrast, NT2 participants showed better stage N1 bout survival (p = .006) and longer stage N1 bouts (p = .02). Conclusions NT1, NT2, and IH have unique sleep physiology compared to subjectively sleepy controls, with only NT1 demonstrating clear nocturnal wake and sleep instability. Overall, sleep stability measures may aid in diagnoses and management of these central nervous system disorders of hypersomnolence. 
    more » « less
  • Citation Formats
  • MLA
  • APA
  • Chicago
  • BibTeX
  • Save / Share this Record
  • Send to Email