Aperiodic and periodic neural activity during sleep in autism spectrum disorders.

BACKGROUND: Influential theories proposed that the complex and heterogeneous clinical and behavioral manifestations of autism spectrum disorder (ASD) arise from dysregulation of neural circuits driven by an imbalance between excitatory (glutamatergic) and inhibitory (GABAergic) neurotransmission. Building on this framework, and considering the role of homeostatic regulation of neuronal excitability in shaping sleep stages, we hypothesized that dysregulation of neuronal network activity-including altered ratios of excitatory drive to interneuronal inhibitory control-might manifest in sleep architecture. An elevated E/I ratio in ASD is expected to render the sleep EEG noisier, less synchronized, and less precisely timed, thereby compromising NREM sleep quality and disrupting memory-related oscillatory coordination.

METHODS: To test this hypothesis, we examined sleep patterns in a homogeneous cohort of adults with ASD without intellectual disability and free from pharmacological treatment, compared with neurotypical (NT) controls. We specifically investigated electrophysiological sleep markers that may reflect alterations in cortical excitability and inhibition. Macro- and microstructural features of nocturnal sleep were assessed using in-home polysomnography (PSG), including analyses of periodic EEG components, such as slow oscillations and sleep spindles, and aperiodic metrics of EEG activity.

RESULTS: Our findings revealed sleep-stage-specific PSG differences in ASD, characterized by increased N3 sleep, decreased N2 sleep, and heightened slope and offset of aperiodic EEG activity during NREM sleep relative to NT controls. Moreover, ASD participants showed elevated alpha power during N2, which positively correlated with ADOS total scores. We also observed a steeper slow-oscillation slope, a reduced anterior-posterior gradient in spindle density, and diminished spindle-slow oscillation coupling, collectively indicating atypical thalamocortical network dynamics in ASD. Overall, these spatially distributed and sleep-stage-dependent alterations reflect dysregulated neuronal dynamics, potentially pointing to increased inhibitory activity arising from altered thalamocortical regulation and compensatory mechanisms related to E/I imbalance.

CONCLUSIONS: Our study provides novel electrophysiological evidence for a nuanced, sleep-related dysregulation in ASD that varies by sleep stage and cortical region, and subtly diverges from patterns observed in NT controls. Given the mechanistic relevance of sleep for neurodevelopment and circuit homeostasis, these findings offer valuable insights into sleep-related neurophysiological dysregulation in ASD.