Sleep disturbances and sensory sensitivities are common disabling features of autism, but their underlying causes are not clear. We hypothesize that both of these difficulties reflect disrupted communication between a deep brain structure, the thalamus, and the brain's outer layer, the cortex. This communication is mediated by the thalamic reticular nucleus (TRN). Due to its small size and location deep in the brain, we cannot assess TRN function without invasive techniques. Fortunately, sleep spindles, a specific brain rhythm provide a noninvasive read-out of TRN function. In Aim 1 we will examine whether reduced sleep spindles are related to worse sleep quality, impaired selective attention during wake, and sensory sensitivities in children with autism. In Aim 2, we will determine whether melatonin, which is commonly used to improve sleep, also increases sleep spindles in autism. If successful, this study will introduce TRN as a target for treatment of sleep disruption and guide larger home-based sleep studies.