Leveraging pediatric state-specific functional brain network dynamics to predict developmental outcomes

PROJECT SUMMARY Determining accurate models of the developing brain’s functional architecture during the first two years of life has potential prognostic value for maturation milestones and the onset of psychopathology. A growing body of literature suggests that infant and toddler brain functional connectivity (FC) networks (as defined using functional magnetic resonance imaging; fMRI) are less mature than adult networks. However, there are two distinct and potentially interacting factors present when estimating infant and toddler FC: 1) age and 2) state. Critically, infant and toddler fMRI data are collected during natural sleep, while fMRI data in older populations are typically collected during resting wakefulness. Research suggests that pediatric FC estimated from sleeping state fMRI appears more similar to adult sleeping state networks (SSN) than awake adult resting state networks. However, even within pediatric FC studies, comparisons regarding age are compounded by developmental differences in state as time spent in different sleep stages changes over the first two years of life. Crucially, individual variability in time spent in each sleep stage poses a challenge for reproducibility and prediction accuracy of extant pediatric, sleeping state fMRI studies such as the Early Life Adversity, Biological Embedding (eLABE) study, Baby Connectome Project (BCP), Developing Human Connectome Project, and the Healthy Brain and Child Development Study. The overarching goals of this Award are to 1) disentangle the relative contributions of sleep stage and age in early brain network development and 2) decode sleep stages in extant infant and toddler sleeping state fMRI. Completion of these goals will enable age and state-specific FC prediction of mental health and clinical outcomes. Towards these goals, infant and toddler sleep stages and FC networks will be characterized using concurrent electroencephalogram (EEG) - fMRI. Aim 1 will optimize pediatric EEG-fMRI acquisition and analysis. Aim 2 will determine pediatric FC SSN development by collecting cross-sectional EEG- fMRI data from 50 children at birth and 24 months of age. Aim 3 will decode sleep stages in out of sample fMRI data (eLABE & BCP). While EEG-fMRI in infants and toddlers is an ambitious and new research direction for the PI, our investigative team has the necessary expertise for success, including FC brain network analysis (PI: Wheelock), pediatric fMRI acquisition and analysis (Smyser), concurrent EEG-fMRI acquisition and analysis (Zempel, Palanca), pediatric EEG sleep staging (Rudock), and machine learning (Lahiri). The knowledge generated from this R01 will be transformative, providing a state-based understanding of early developmental brain networks and a framework for accurate developmental outcome predictions.