The temporal relationship between physical activity, mood, and sleep in older adults via a lead‐lag analysis

Abstract

AbstractBackgroundPhysical Activity, sleep, and mood are key modifiable lifestyle behaviours that are individually associated with lower cognitive decline [1]. Prior research suggests there may be a tridirectionally relationship among these three behaviours. The temporal relationship among these behaviours is unknown. Using an Ecological Momentary Assessment (ECA) and a time series analysis we aim to understand the temporal and longitudinal relationship among these factors.MethodWe performed a 10‐day lead‐lag analysis using Moderate‐to‐Vigorous Physical Activity (MVPA) and sleep as measured by a 24‐hour wrist‐worn accelerometer, and a self‐assessed daily mood scale on 30 community dwelling older adults who participated in a behavioural study to enhance physical activity in sedentary individuals (Clinicaltrials.gov identifier: NCT04315363). Participants were included if they did not exercise more than 150 minutes of MVPA per week (International Physical Activity Questionnaire) and were sitting more than 8 hours per day (Marshall Sitting Questionnaire). We examine each variable separately for auto‐correlative relationships for different timepoints of the same variable. The temporal dynamics are calculated from the rank‐order of the lead‐lag coefficients. The relationships between different timepoints of mood, MVPA, and sleep are examined through cross‐correlations.ResultsAutocorrelations showed that mood presented seasonality with baseline mood positively correlated to the subsequent day’s mood (r = +.571, p = .037) and negatively correlated with mood at day 7 (r = ‐.330, p = 0.27) and day 8 (r = ‐.248, p = .013). The temporal dynamics identified mood (B = +.602), preceding MVPA time (B = ‐.001), and sleep time (B = ‐.276) (Figure 1). Furthermore, cross‐correlations demonstrated that mood was positively correlated to MVPA at the same time point (r = ‐.605) and also that previous day mood predicted next day’s MVPA (r = ‐.629). Cross‐correlations confirmed that MVPA (r = ‐.507) and mood (r = ‐.504) predicted next day’s sleep time.ConclusionImproved mood precedes the increase in MVPA and sleep time in sedentary older adults. Understanding temporal dynamics between mood, sleep time, and MVPA, can help develop better interventions targeting increase in physical activity in older adults. Reference:[1] 2020 Report of the Lancet Commission. https://doi.org/10.1016/S0140‐6736(20)30367‐6.