From Breath to Brain: NICU respiratory interventions and bedside brain signal entropy predict later autism risk

Premature infants often experience hypoxia and require prolonged ventilation, which can trigger systemic inflammation, damage the developing brain, and increase the risk of neurodevelopmental disorders such as Autism Spectrum Disorder (ASD). Early intervention is key for ensuring optimal outcomes for those with ASD; thus emphasizing the critical importance of accurately identifying infants at risk as early as possible. Here, infants underwent electroencephalography during social (held) and nonsocial (not held) resting state conditions to assess brain signal variability, saliva collection to determine inflammation, calculation of a novel Prognostic Respiratory Intensity Scoring Metric (PRISM) to assess the burden of respiratory support, and ASD testing in toddlerhood. Higher PRISM scores were associated with increased brain signal entropy during the nonsocial resting state. However, this association was not observed in the social resting state condition - particularly for male babies. Interestingly in female infants, we saw that the relationship between brain signal entropy and PRISM scores were potentially mediated by cytokines. Notably, the interaction between nonsocial resting state brain signal entropy, sex, and PRISM scores predicted risk of developing ASD with 88 % accuracy. These non-invasive measures can identify infants at the highest risk for an ASD diagnosis before discharge.