Neonatal Electrophysiological Biomarkers for Neurodevelopmental Disorder Risk

Background: Presently, the factors leading to ASD are unknown, with no specific biological, genetic, or environmental marker. But, there is abundant evidence of structural, functional, and morphological brain abnormalities in individuals with ASD. Although ASD can be reliably diagnosed as early as 2, the median age of diagnosis is approximately 4 years of age. Prior research has demonstrated that early intervention services from birth to 36 months of age can drastically improve neurodevelopmental outcome. Therefore, the identification of an objective neonatal marker of ASD has the potential to aid in detecting at-risk children who may benefit from early intervention services.

Objectives: Prior research has demonstrated that neonatal electroencephalography (EEG) during sleep is predictive of neurodevelopmental outcome in clinical and neurotypical populations. Although abnormal EEG power and coherence has been demonstrated concurrently in both children and adults with ASD, to our knowledge this is the first study to examine EEG power in neonates and subsequent autism risk.

Methods: Neonatal EEG was collected from healthy term newborns enrolled in the Prenatal Alcohol and SIDS and Stillbirth Network in the Western Cape province of South Africa. Average power for 10 frequency bands were computed for multiple 30-second epochs for a 10-minute period prior to a physiological challenge. We examined neonatal EEG power in active sleep in the low-frequency theta (2-9Hz), mid-frequency Alpha (10-12Hz), and higher-frequencies Beta (13-21Hz) and Low Gamma (22-36Hz). Neurocognitive outcome and autism risk were assessed between 30 and 38 months of age by The Modified Checklist for Autism in Toddlers, Revised with Follow-Up (M-CHAT-R/F) (n=198), the Brief Infant-Toddler Social and Emotional Assessment (BITSEA) (n=253), and the Bayley Scales of Infant Development III Screener (Bayley)(n=117).

Results: When controlling for gestational age at birth, sex, chronological age in months at the time of assessment, and multiple comparisons (Benjamini & Hochberg 5% False Discovery Rate), significant associations were found between neonatal EEG power during active sleep, autism risk, and expressive language ability. BITSEA ASD Risk scores were correlated with EEG power in the Theta frequency in the right temporal region (r(86)=.279, p<.005) where lower EEG power was indicative of higher autism risk. Although there were no significant correlations between the receptive language, gross motor, fine motor, or the overall cognitive scores, there were moderate correlations between the Bayley expressive language subdomain and EEG power in the Alpha, Beta, and Low Gamma frequencies in the left central, occipital, and frontal polar regions (Alpha left-central: (r(86)=-.453, p<.001); Alpha left-occipital (r(86) = -.450, p<.001); Beta left-central: (r(43)=-.434, p< .001); Low gamma left-central (r(43)=-.483, p<.001); Low gamma left-frontalpolar (r(44)=-.457, p< .001)). Specifically, higher EEG power in these frequency bands was indicative of superior expressive language ability.

Conclusions: Our preliminary results demonstrate robust associations for expressive language ability and modest associations between early neural biomarkers for autism risk. Further research to examine neonatal neural oscillations and subsequent developmental trajectories may contribute to our comprehension of atypical neurodevelopment. Data collection is ongoing.