A novel machine learning unsupervised algorithm for sleep/wake identification using actigraphy
Xinyue Li1,2, Yunting Zhang2,3, Fan Jiang3,4
1School of Data Science, City University of Hong Kong, Hong Kong, China.
Chronobiology International
|April 29, 2020
Summary
A new unsupervised Hidden Markov Model (HMM) algorithm accurately identifies sleep/wake states from actigraphy data, outperforming existing methods and characterizing individual activity patterns for broader research applications.
Area of Science:
- Sleep science and biomedical engineering.
- Machine learning applications in healthcare.
Background:
- Actigraphy is a common tool for sleep studies, but lacks a universal unsupervised algorithm for sleep/wake identification.
- Unsupervised algorithms are crucial for large-scale studies and when polysomnography (PSG) is unavailable, as they don't require pre-labeled data.
Purpose of the Study:
- To propose and evaluate a novel unsupervised machine learning algorithm based on the Hidden Markov Model (HMM) for individualized sleep/wake identification using actigraphy.
- To compare the performance of the HMM algorithm against existing unsupervised (Actiwatch Software) and supervised (UCSD) algorithms using PSG as the reference standard.
Main Methods:
- Developed an individualized, unsupervised Hidden Markov Model (HMM) algorithm utilizing actigraphy data to infer sleep and wake states.
- Evaluated the HMM algorithm using actigraphy and PSG data from 43 participants in the Multi-Ethnic Study of Atherosclerosis.
- Compared epoch-by-epoch and sleep variable estimates against Actiwatch Software (AS) and UCSD algorithms using PSG as the ground truth.
Main Results:
- The HMM algorithm achieved an accuracy of 85.7%, comparable to AS (84.7%) and UCSD (85.0%), but with significantly higher specificity (36.4% vs. 30.0% and 31.7%).
- HMM demonstrated superior performance in Bland-Altman analysis for total sleep time, sleep latency, and sleep efficiency, showing closer agreement with PSG and narrower limits of agreement.
- The HMM approach successfully differentiated active and sedentary individuals based on activity count variability, offering insights into sedentary behavior patterns.
Conclusions:
- The proposed unsupervised HMM algorithm offers a robust and individualized method for sleep/wake identification from actigraphy, outperforming current software and supervised methods in key metrics.
- HMM enhances the utility of actigraphy in research settings where PSG is impractical or supervised training data is absent.
- The algorithm's ability to characterize individual activity patterns provides a valuable tool for downstream analyses of behavior and health.
Related Concept Videos
01:24Sleep-Wake Cycles
2.6K
Sleep is an essential physiological process vital to maintaining overall well-being. The reticular activating system (RAS), a network of neurons in the brainstem, regulates wakefulness and sleep. While it may seem passive, sleep consists of distinct cycles, each with its unique characteristics and functions. Two key sleep phases are non-rapid eye movement (NREM) and rapid eye movement (REM).
NREM Sleep
NREM sleep comprises four progressive stages that seamlessly merge:
NREM Sleep
NREM sleep comprises four progressive stages that seamlessly merge:
2.6K
01:11Understanding Sleep
1.3K
Sleep, an essential biological state, involves significant reductions in physical activity, sensory awareness, and interaction with the environment. This complex physiological process is primarily regulated by specific brain regions, notably the hypothalamus and pons, which govern the sleep-wake cycle or circadian rhythm.
The circadian rhythm, a nearly 24-hour cycle, is deeply influenced by environmental light cues. Light exposure directly affects the hypothalamus, which in turn regulates...
The circadian rhythm, a nearly 24-hour cycle, is deeply influenced by environmental light cues. Light exposure directly affects the hypothalamus, which in turn regulates...
1.3K
01:22Stages of Sleep
1.2K
Sleep progresses through distinct stages, each characterized by specific brain wave patterns and physiological responses ranging from wakefulness to stages of non-rapid eye movement, known as non-REM, to rapid eye movement, referred to as REM. Understanding these stages helps in recognizing how sleep supports various bodily and cognitive functions.
Before sleep begins, in wakefulness, the brain exhibits primarily beta waves, which are high in frequency and low in amplitude, indicating alertness...
Before sleep begins, in wakefulness, the brain exhibits primarily beta waves, which are high in frequency and low in amplitude, indicating alertness...
1.2K