The Impact of Sleep Variability on Sleep Duration and Daytime Behavior in Children on the Autism Spectrum

Background: Children with autism spectrum disorder (ASD) have sleep difficulties at a rate of 40-80 percent, compared to 25-40 percent of typically developing (TD) children, affecting daytime behavior and quality of life. Sleep difficulties can be measured objectively through actigraphy or polysomnography or through subjective report, with both methods providing comparable accuracy (Baddam et al., 2018). In typical development, sleep duration decreases with increasing age, yet this trajectory is poorly understood in children with autism.

Objectives: To examine relationships among variability of sleep cycle with sleep duration and daytime behavior across development in children with ASD and TD.

Methods: Participants were children aged 7-18 with ASD (n=34, Mean IQ=96.3, Mean Age=12.4 years) and TD (n=16, Mean IQ=112.0, Mean Age=12.8 years). Parents completed the Yale Developmental Sleep Questionnaire, which asks for the earliest, latest, and usual time of their child’s bedtime routine (BR), lights out (LO), and wake time (WT) for the most recent weekdays and weekend. Clinician-administered and parent-reported measures provided information on intellectual function and symptomatology: Differential Ability Scales, Autism Diagnostic Observation Schedule, and Child Behavior Checklist. Children were given variability scores (VS) consisting of the sum of the differences in earliest and latest times for BR, LO and WT, and difference between weekend and weekday usual BR, LO, and WT. Children were placed in either a low variability (LV; VS≤7h, n=18: 12 ASD, 6 TD) or high variability (HV; VS> 7h, n=32: 22 ASD, 10 TD) variability group. The mean ages and IQs did not significantly differ based on sleep variability. For both ASD and TD children, variability scores were normally distributed.

Results: Consistent with the literature, a negative correlation between sleep duration and age was seen in the TD group (r=0.773, p=0.000). However, this trend was not observed overall in the ASD group (r=0.185, p=0.295). When contrasting LV and HV participants with ASD, the expected developmental pattern was observed among LV ASD participants (r=0.740, p=0.006) but not HV ASD participants (r=0.040, p=0.861). Within the ASD group, parents of children with HV reported significantly less inattention (p=0.052) and hyperactivity (p=0.040) than the LV group.

Conclusions: High and low sleep variation in ASD children were associated with distinct developmental patterns of sleep maturation and different daytime behaviors. These differences were not observed in TD children. Children with ASD and high sleep variation failed to display the expected association between sleep duration and age and, counter to predictions, were less likely to display inattention and hyperactivity. Future studies can build on these findings by using objective measures of sleep such as polysomnography or actigraphy (Hodge et. al., 2012) to further explore the relationship between sleep variability and daytime behaviors.