Assessing Sleep with Wrist and Hip Actigraphy in young adults-Comparison to Simultaneous Laboratory Polysomnography
- Added on August 20, 2013
Laboratory-based polysomnograpy (PSG) is the gold standard technique for assessment and quantification of sleep. Wrist and hip actigraphy have been widely used to quantify sleep and physical activity, respectively. It remains unclear whether hip actigraphy is also a valid measure of sleep.
Purpose: To compare sleep varialbes simultaneously derived from PSG, wrist actigraphy and hip actigraphy in healthy, young adults.
72 healthy individuals (49% females) aged 20.0 +/- 0.2 years (mean +/- SD) were recruited from Raine Study, a population-based pregnancy cohort study in Western Australia. Each underwent in-laboratory PSG while simultaneously wearing one GT3X+ actigraph (ActiGraph, Florida, USA) on their non-dominant wrist and another on an elastic belt on the right hip. Estimates of sleep onset latency (SOL), total sleep time (TST), wake after sleep onset (WASO), and sleep efficiency (SE) were derived using standard proprietary equations for count based actigraphy and standard PSG definitions. The sensitivity and specificity of actigraphic estimates of sleep versus polysomnography-defined sleep were also assessed on a 60 second epoch-by-epoch basis.
Relative to PSG measures: SOL was underestimated by wrist and hip actigraphy; WASO was overestimated by wrist but underestimated by hip actigraphy; and TST and SE were similar by wrist actigraphy but overestimated by hip actigraphy (see Table). Compared to wrist actigraphy, SOL and WASO were less but TEST and SE more when measure by hip actigraphy. The sensitivity, specificity and accuracy of wrist actigraphy were 91%, 44% and 85%, respectively and of hip actigraphy were 98%, 13% and 86%, respectively.
A hip-worn GT3X+ actigraph can not be used to accurately measure sleep variablers in young adults using the standard equeations due to its very poor ability to detect wakefulness during sleep. Although wrist actigraphy is widely used in both clinical and research settings the present study shows it to have a limited capacity to detect periods of wake during sleep when using standard equations.