Study objectives: To relate parentally reported sleep concerns in autism spectrum disorders (ASD) to polysomnographic (PSG) findings and measures of daytime behavior and autism symptomatology.

Design: Cross-sectional study involving validated questionnaires, sleep histories and diaries, 2 nights of PSG, and the Autism Diagnostic Observation Schedule (ADOS).

Setting: Vanderbilt University General Clinical Research Center Sleep Core.

Participants: 21 children with ASD and 10 typically developing (TD) children, aged 4-10 years. Children were free of psychotropic medications, with no history of mental retardation or epileptic seizures.

Interventions: N/A.

Measurements and results: Children with ASD were defined as "good sleepers" (10 children) and "poor sleepers" (11 children) on the basis of parental report; the age-comparable TD children were all reported by their parents to be good sleepers. Poor sleepers with ASD showed prolonged sleep latency and decreased sleep efficiency on night 1 of PSG and differed on insomnia-related subscales of the Children's Sleep Habits Questionnaire (CSHQ; increased sleep onset delay and decreased sleep duration). The good sleepers with ASD did not differ from the TD children in sleep architecture or on CSHQ domains. As compared with ASD good sleepers, the ASD poor sleepers also had higher scores related to affective problems on the Child Behavior Checklist and more problems with reciprocal social interaction on the ADOS.

Conclusions: Parentally reported sleep concerns of insomnia in children with ASD are substantiated by validated sleep questionnaires and by PSG. Furthermore, good sleepers with ASD showed fewer affective problems and better social interactions than ASD poor sleepers.