Abstract
Wearable computing devices allow collection of densely sampled real-time information on movement enabling researchers and medical experts to obtain objective and non-obtrusive records of actual activity of a subject in the real world over many days. Our interest here is motivated by the use of activity data for evaluating and monitoring the circadian rhythmicity of subjects for research in chronobiology and chronotherapeutic healthcare. In order to translate the information from such high-volume data arising we propose the use of a Markov modelling approach which (i) naturally captures the notable square wave form observed in activity data along with heterogeneous ultradian variances over the circadian cycle of human activity, (ii) thresholds activity into different states in a probabilistic way while respecting time dependence and (iii) gives rise to circadian rhythm parameter estimates, based on probabilities of transitions between rest and activity, that are interpretable and of interest to circadian research.
Highlights
Questions of interest regarding the research of sleep –wake cycles in humans and mammals are commonly studied by measuring activity through gross motor movement where accelerometers have become a feasible and affordable way to obtain objective non-obtrusive recordings of rest–activity rhythms of free living individuals over many days [1,2,3]
In this paper we propose the use of a hidden Markov models (HMMs) approach which can address the challenges of modelling activity data, and provides a natural framework for extracting information from them
The model can capture the characteristic features discernible in time series of activity measured over days, such as the notable square wave form with heterogeneous ultradian variances over the circadian cycle of human activity
Summary
Questions of interest regarding the research of sleep –wake cycles in humans and mammals are commonly studied by measuring activity through gross motor movement where accelerometers have become a feasible and affordable way to obtain objective non-obtrusive recordings of rest–activity rhythms of free living individuals over many days [1,2,3]. HMMs have been used, among other methods such as Gaussian mixture models and K-means clustering, as an unsupervised machine learning technique for the task of recognition and classification of specific human activity modes such as standing, walking, running, ascending/descending, sitting, cycling etc This scenario is different to ours, in particular as data acquisition in these cases is much denser at epoch lengths in units of small fractions of a second over a short time span of typically less than an hour, and the research questions is aimed at being able to recognize a specific kind of activity We shall show that HMMs naturally provide the necessary tools to model the features observed in the type of data we are interested in, and can be further extended toward a dynamic Markov process which is influenced by a circadian oscillator, the strength and nature of which can be inferred from an individual’s PA data from a sensing device, that can be unobtrusively worn over many days/weeks
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
