Authors:
Francis W. Usher
1
;
Chiying Wang
1
;
Sergio A. Alvarez
2
;
Carolina Ruiz
1
and
Majaz Moonis
3
Affiliations:
1
Worcester Polytechnic Institute, United States
;
2
Boston College, United States
;
3
U. of Massachusetts Medical School, United States
Keyword(s):
Sleep, Bout duration, Sleep dynamics, Data mining, Clustering, Machine learning.
Related
Ontology
Subjects/Areas/Topics:
Artificial Intelligence
;
Biomedical Engineering
;
Business Analytics
;
Cardiovascular Technologies
;
Clinical Problems and Applications
;
Computing and Telecommunications in Cardiology
;
Data Engineering
;
Data Mining
;
Databases and Information Systems Integration
;
Datamining
;
Enterprise Information Systems
;
Health Engineering and Technology Applications
;
Health Information Systems
;
Medical and Nursing Informatics
;
Pattern Recognition and Machine Learning
;
Sensor Networks
;
Signal Processing
;
Soft Computing
;
Software Systems in Medicine
Abstract:
This paper explores the discovery of patterns in human sleep data based on the duration statistics of continuous bouts in individual sleep stages during a full night of sleep. Hypnograms from 244 patients are examined. Stage bout durations are described in terms of the quartiles of their stage bout duration distributions, yielding 15 descriptive variables corresponding to wake after sleep onset, NREM stage 1, NREM stage 2, slow wave sleep, and REM sleep. Unsupervised Expectation-Maximization clustering is employed to identify distinct groups of hypnograms based on stage bout durations. Each group is shown to be characterized by bout duration quartiles of specific sleep stages, the values of which differ significantly from those of other groups (p<0.05). Among other sleep-related and health-related variables, several are shown to be significantly different among the bout duration groups found through clustering, while multivariate linear regression fails to yield good predictive model
s based on the same bout duration variables used in the clustering analysis. This provides an example of the successful use of machine learning to uncover naturally occurring dynamical patterns in sleep data that can also provide sleep-based indicators of health.
(More)