EEG Resting State Architecture Is Atypical in Children with ASD and Sleep Disturbances

Background: Sleep disturbance is common in Autism Spectrum Disorder (ASD), with between 40 and 80% of children with ASD reporting sleep difficulties, as compared to just 9 to 50% of typically developing (TD) children (Robinson-Shelton, 2016). Sleep disturbances and consequent sleep deprivation are reflected in increased absolute theta (4-8 Hz) power during resting state brain activity in both children and adults (Cajochen, 1995). While the resting EEG profile associated with sleep deprivation is established in the literature, relationships among resting state EEG profiles and sleep disturbances in children with ASD are poorly understood.

Objectives: To investigate relationships among sleep disturbances and resting state EEG power in children with ASD and TD. Study of sleep disturbances in the ASD population may provide important information about underlying neuropathology and potential subgroups of ASD.

Methods: Participants were children with ASD (n=31) and TD controls (n=16), aged 6-18. Parents and children completed measures of social functioning, anxiety, and problem behaviors. Parents additionally completed the Children’s Sleep Habits Questionnaire (CSHQ), a validated 33-item measure of sleep problems; impaired sleep is defined as a CSHQ score ≥ 41. A resting EEG was collected from all participants. Participants sat for 60 seconds with their eyes open, followed by 60 seconds with their eyes closed. EEG data was filtered, segmented into 2-s epochs, and hand-edited for movement artifacts. Participants with less than 30 seconds of artifact-free data were excluded, resulting in a final sample of 13 children with ASD and 11 TD children.

Results: Consistent with the literature, 52% of children with ASD met CSHQ criteria for sleep impairment, as compared to 25% of children with TD, approaching a statistically significant difference in this small sample (p=0.08). CSHQ score did not correlate with measures of behavioral dysregulation from the Child Behavior Checklist. All children demonstrated global reduction of absolute theta power with increasing age (frontal: p=0.003, central: p=0.01, posterior: p=0.009; temporal p=0.01), consistent with the expected developmental trajectory. Children with TD showed a positive correlation between CSHQ score and theta power along the midline (r=0.9, p<0.001), frontal (r=0.87, p=0.001), and central (r=0.65, p<0.001) regions of the scalp, whereas children with ASD showed no correlation between theta power and CSHQ score.

Conclusions: Findings demonstrate that children with ASD had higher rates of sleep disturbances, however do not exhibit the expected relationship between sleep problems and EEG spectral power that was seen in TD children. This atypical presentation of sleep problems suggests a potentially distinct neural mechanism of sleep disturbance in ASD. An EEG correlate of the disruption of the sleep/wake cycle in children with ASD provides an important area for future study.