Validity of Actigraphy Compared to Polysomnography for Sleep Assessment in Children With Autism Spectrum Disorder

Abstract

Actigraphy (ACT) is a non-invasive objective assessment tool for the study of sleep–wake rhythms. It is of particular interest in children with autism spectrum disorder (ASD), as sleep disorders are highly prevalent and have a significant impact on both cognitive and behavioral functions. As polysomnography (PSG), the gold standard for the assessment of sleep, is difficult to perform in children with ASD, ACT has become a tool of choice but has not yet been validated against PSG using state-of-the-art methodology. The main objective of this study was to assess, for the first time, the validity of ACT compared to PSG for the measurement of sleep in children with ASD. During the same night of hospitalization, PSG and ACT were conducted in 26 children (6 girls and 20 boys; mean age 5.4 years ± 1.6) diagnosed with ASD according to DSM-5 criteria and standardized diagnostic scales. Sleep parameters were total sleep time (TST), sleep latency (SL), wake after sleep onset (WASO), and sleep efficiency (SE). To compare PSG and ACT, we conducted sleep parameter agreement analyses including: intraclass correlation coefficient (ICC), Bland-Altman plots, and equivalence tests. The comparison also included an epoch-by-epoch (EBE) agreement analysis to determine sensitivity (ability to detect sleep) and specificity (ability to detect wake). According to equivalence tests, the difference between ACT and PSG measures was clinically acceptable for TST (<30 min, p < 0.01), SL (<15 min, p < 0.001), and SE (10%, p < 0.01), but not for WASO (<15 min, p = 0.13). There was a good agreement between methods for SL (ICC = 0.79) and TST (ICC = 0.85) and a moderate agreement for WASO (ICC = 0.73) and SE (ICC = 0.68). The EBE agreement analysis revealed a high sensitivity (0.94 ± 0.06) and moderate specificity (0.5 ± 0.2). Since sleep disorders are one of the most common comorbidities within the ASD population and are highly prevalent, it is essential to validate objective tools of assessment. To our knowledge, our study is the first to validate ACT compared to PSG, using a state-of-the-art methodology, in children with ASD. The results suggest ACT to be a valid method to evaluate sleep within this population, with a good reliability for most sleep parameters.

Disturbances of Continuous Sleep and Circadian Rhythms Account for Behavioral Difficulties in Children with Autism Spectrum Disorder

“…Children's sleep-wake cycles depend on their parents rather than their internal rhythms, many still nap and thus display a biphasic circadian rhythm and they present more movement during sleep. While actigraphy-derived measures of sleep have been validated against its gold standard, polysomnography [17], this is yet to be done for NPCRA measures, which have not yet been validated against a gold standard measurement of circadian rhythms such as melatonin or core body temperature.…”

“…Actigraphy data were scored automatically for sleep/wake using the Actiwatch Activity and Sleep Analysis 7 ® software algorithm, version 7.23. As our previous actigraphy validation study in children with ASD indicated that the low sensitivity-threshold setting shows the best fit when compared to polysomnography (PSG) [17], this setting is thus reported here. The low sensitivity-threshold is defined as a threshold level of 80 counts, i.e., an activity score of 80 counts or more during an epoch of 1 min is necessary for that epoch to be scored as wake.…”

“…As such, LSE was defined here as the longest uninterrupted sleep episode or interrupted by less than five consecutive minutes of scored wake on ACT. This modification was necessary due to the brief episodes of sleep that may be identified incorrectly as a wake episode [17]. The definition of LSE was based on a comparison of ACT and PSG in this sample (unpublished data), which identified a tolerance margin of 5 min as the most fitting one, when compared to LSE measured by PSG.…”

“…Objective sleep was measured either with the gold standard polysomnography or with actigraphy (ACT). The latter is being more and more used to assess sleep in children with ASD, as it is more feasible and allows ecological recordings over longer periods of time, and was recently validated against polysomnography [17].…”

Sleep Regularity Index in Patients with Alcohol Dependence: Daytime Napping and Mood Disorders as Correlates of Interest

“…For this study, the Philips Actiwatch-2 model was used to collect objective data on sleep quality and duration as well as daytime activity patterns. Sleep parameters assessed via actigraphy demonstrate high sensitivity and moderate accuracy for detecting sleep in populations with normal and disturbed sleep, when compared with polysomnography [25,26]. Participants wore the watch on their non-dominant wrist from enrollment in our study (day 2 of the inpatient stay) up until discharge.…”