Seeking for a fingerprint: analysis of point processes in actigraphy recording

TL;DR

This study analyzes actigraphy recordings to identify scale-invariant patterns and fractional point process measures that can distinguish healthy individuals from those with sleep deprivation, highlighting potential biomarkers for sleep-related behavioral changes.

Contribution

It introduces the use of fractional point process analysis and survival function exponents as effective biomarkers for detecting sleep deprivation effects in actigraphy data.

Findings

Survival function exponents effectively differentiate sleep-deprived from healthy subjects.

Scale-invariant statistics remain consistent over various time scales.

Alterations in activity patterns correlate with sleep deprivation severity.

Abstract

Motor activity of humans displays complex temporal fluctuations which can be characterized by scale-invariant statistics, thus documenting that structure and fluctuations of such kinetics remain similar over a broad range of time scales. Former studies on humans regularly deprived of sleep or suffering from sleep disorders predicted change in the invariant scale parameters with respect to those representative for healthy subjects. In this study we investigate the signal patterns from actigraphy recordings by means of characteristic measures of fractional point processes. We analyse spontaneous locomotor activity of healthy individuals recorded during a week of regular sleep and a week of chronic partial sleep deprivation. Behavioural symptoms of lack of sleep can be evaluated by analysing statistics of duration times during active and resting states, and alteration of behavioural organization can be assessed by analysis of power laws detected in the event count distribution, distribution of waiting times between consecutive movements and detrended fluctuation analysis of recorded time series. We claim that among different measures characterizing complexity of the actigraphy recordings and their variations implied by chronic sleep distress, the exponents characterizing slopes of survival functions in resting states are the most effective biomarkers distinguishing between healthy and sleep-deprived groups.