Power-law temporal auto-correlations in day-long records of human physical activity and their alteration with disease

TL;DR

This study reveals that human physical activity exhibits power-law auto-correlations over long durations, with variations during day and night, and altered patterns in individuals with chemical sensitivities, highlighting potential disease markers.

Contribution

It demonstrates the presence of power-law auto-correlations in human activity and how these patterns differ with routine conditions and disease states, providing new insights into activity dynamics.

Findings

Power-law decay in activity auto-correlations across conditions

Differences in diurnal and nocturnal activity correlations during routine

Altered auto-correlation patterns in patients with chemical sensitivities

Abstract

We investigate long-duration time series of human physical activity under three different conditions: healthy individuals in (i) a constant routine protocol and (ii) in regular daily routine, and (iii) individuals diagnosed with multiple chemical sensitivities. We find that in all cases human physical activity displays power law decaying temporal auto-correlations. Moreover, we find that under regular daily routine, time correlations of physical activity are significantly different during diurnal and nocturnal periods but that no difference exists under constant routine conditions. Finally, we find significantly different auto-correlations for diurnal records of patients with multiple chemical sensitivities.