Sleep as a window into thalamocortical pathology: generative modeling implicates NMDA receptor hypofunction in 22q11.2 deletion syndrome

Background22q11.2 deletion syndrome (22q11.2DS) is a strong genetic risk factor for neuropsychiatric conditions, including autism, ADHD, intellectual disability, and schizophrenia. Up to 25% of adults with 22q11.2DS get diagnosed with schizophrenia and 40% develop psychotic conditions more broadly. Sleep disturbances and altered sleep EEG in 22q11.2DS suggest underlying thalamocortical dysfunction. However, EEG alone lacks mechanistic resolution, limiting insight into causal mechanisms and treatment targets. Computational modelling can bridge this gap by inferring synaptic dynamics from EEG data. MethodsWe applied conductance-based Dynamic Causal Modelling to sleep-wake EEG from children with 22q11.2DS (n = 28, Mage = 14.6, (SD = 3.4) and their neurotypical siblings (n = 17, Mage = 13.7, (SD = 3.4). We estimated synaptic parameters reflecting AMPA, NMDA, GABAA and GABAB receptor conductances across vigilance states. To identify candidate intervention targets, we performed "in silico pharmacology", simulating increases and decreases in receptor gain to assess which adjustments best aligned 22q11.2DS EEG with control patterns. We also examined associations with psychiatric symptoms. ResultsIncreasing NMDA receptor (NMDA-R) gain consistently improved alignment between 22q11.2DS and sibling neural spectra, especially during N2 (ES = 0.32) and N3 sleep (ES = 0.45). Key contributors were local NMDA-R-mediated recurrent-connections among superficial pyramidal cells, and NMDA-driven excitation from spiny stellate to superficial pyramidal populations. We also found that stronger thalamo-cortical connectivity during N3 was associated with greater sleep problems ({beta} = 0.34, PFDR = 0.006), while stronger AMPA-mediated excitation of superficial interneurons during wakefulness predicted anxiety symptoms ({beta} = -0.32, PFDR = 0.032). ConclusionModelling across sleep-wake states revealed state-dependent synaptic abnormalities in 22q11.2DS, identifying NMDA-R hypofunction as a critical mechanism across sleep stages. These findings highlight the potential of virtual pharmacology for predicting receptor-level targets in neurodevelopmental disorders.