Measuring Variability in Rest-Activity Rhythms from Actigraphy with Application to Characterizing Symptoms of Depression

Abstract

The twenty-four hour sleep-wake pattern known as the rest-activity rhythm (RAR) is associated with many aspects of health and well-being. Researchers have utilized a number of interpretable, person-specific RAR measures that can be estimated from actigraphy. Actigraphs are wearable devices that dynamically record acceleration and provide indirect measures of physical activity over time. One class of useful RAR measures are those that quantify variability around a mean circadian pattern. However, current parametric and nonparametric RAR measures used by applied researchers can only quantify variability from a limited or undefined number of rhythmic sources. The primary goal of this article is to consider a new measure of RAR variability: the log-power spectrum of stochastic error around a circadian mean. This functional measure quantifies the relative contributions of variability about a circadian mean from all possibly frequencies, including weekly, daily, and high-frequency sources of variation. It can be estimated through a two-stage procedure that smooths the log-periodogram of residuals after estimating a circadian mean. The development of this measure was motivated by a study of depression in older adults and revealed that slow, rhythmic variations in activity from a circadian pattern are correlated with depression symptoms.

Profiles of Accelerometry-Derived Physical Activity Are Related to Perceived Physical Fatigability in Older Adults

“…To date, no known studies have explored similar types of RAR "profiles" and related them to perceived physical fatigability. Nonetheless, these results are in line with what we might expect based on another study conducted by Smagula et al [35], which used a similar clustering technique on RAR parameters and found that later and irregular RARs were associated with depression symptoms in a sample of adults [36].…”

“…For example, low pseudo-F (or low stability) could be attributed to a participant waking up regularly throughout the night, or a participant who has systematic differences in their waking times (e.g., early riser during weekdays versus late riser on the weekend). Future studies could investigate the use of the residual circadian spectrum to quantify frequency domains of variability [36]. Additionally, there are limitations to the k-means clustering technique: it contains an implicit assumption of normality, or at least symmetry, of the data (through the use of means over other statistics of central tendency), and it is an unsupervised technique.…”

Large cognitive fluctuations surrounding sleep in daily living

“…Still, diurnal, infra-and ultra-radian rhythms are all visible in the movements logged using actigraphy (Fossion et al, 2017;Wong et al, 2013). Of interest, the inter-individual differences in the near-24-h rhythm of physical activity may be markers of clinical conditions (Germain and Kupfer, 2008;Krafty et al, 2019;Leng et al, 2020). The observation in daily living that diurnal rhythms for the left versus right arm differ further support the idea that these rhythms can be compartmentalized in the nervous system (Natale, 2002).…”

Nonparametric time series summary statistics for high-frequency accelerometry data from individuals with advanced dementia

“…This model is often a poor fit to accelerometry data, especially for individuals with weak circadian patterns [ 6 ]. Alternative methods include singular spectrum analysis (SSA) which employs periodic components with varied amplitude and phase to fit the data [ 6 ], or the use of multiparameter-extended cosine functions [ 7 ]. In general, parametric methods such as cosinor can be very useful for characterising circadian rhythms, however such methods are typically model-based and suffer the usual disadvantages of parametric methods: namely the bias and lack of robustness that results when model assumptions are not met by the application data source of interest.…”