Sleep disorders are among the most common comorbidities in children with Autism Spectrum Disorder (ASD), and subjectively defined sleep disturbances have been related to ASD symptom severity. However, no study has investigated the differential impact of objectively measured sleep and circadian rhythm disturbances on behavioral difficulties in this population. Fifty-two children with ASD aged 3–10 years underwent assessments of sleep and circadian rest–activity rhythms objectively with actigraphy and subjectively with the Children’s Sleep Habits Questionnaire. Behavioral difficulties were assessed using the ABC-C. Group comparison analyses were used to compare sleep and circadian rhythm parameters of children with higher and lower behavioral difficulties and dominance analysis to rank predictors and address multicollinearity. Children with high irritability had a shorter continuous sleep period compared to those with lower irritability (−60 min, p = 0.04), as well as those with high stereotypic behaviors compared to children with less stereotypies (−75 min, p = 0.006). Objective circadian and sleep disturbances accounted together for, respectively, 17%, 18% and 36% of the variance in social withdrawal, irritability and stereotypic behaviors. The identification of both sleep and circadian rhythm disturbances as explanatory factors for behavioral difficulties warrants their inclusion in the existing behavioral management strategies for children with ASD.

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.