Visuocortical Foundations of Development

PROJECT SUMMARY/ABSTRACT The integrity of the visual system during infancy has been associated with later developmental and cognitive outcomes, suggesting that visuocortical function plays a foundational role for cognitive development. However, the development of the visual system is complex, and it is unlikely that a single measure reliably predicts cognitive function. The proposed research will assess the development and predictive value of infant visual function using a battery of tasks and machine learning techniques in a longitudinal sample of infants tested at 6-, 10-, 14- and 18- months of age. A vision screener and electrophysiological measures will be examined in relation to concurrent and subsequent cognitive performance measured by the NIH Infant and Toddler (Baby) Toolbox (NBT). Aim 1 of the proposed work is to characterize infant visuocortical function and developmental trajectories using a battery of tasks that measure (1) Visual Evoked Potentials (VEP), (2) frequency-tagged steady state Visual Evoked Potential (ssVEP), (3) occipitotemporal oscillatory EEG activity (dominant rhythm (alpha oscillation), (4) Event-related potentials (ERPs) in response to faces and objects, (5) refractive errors and ocular misalignments associated with poor vision using the Welch Allyn Spot Vision Screener, and (6) the development of pupil size, pupil distance, and eye gaze deviations. Models of development will be examined for each measure to estimate trajectories of developmental change as well as the internal consistency and trial by trial reliability of the EEG measures using Bayesian Hierarchical Linear Models (BHLM). Aim 2 is to determine the extent to which reliable indices of visuocortical function predict concurrent and later developmental outcomes using two different innovative data integration models. To address Aim 2, visual processing measures (VEP, ssVEP, EEG, ERP, and metrics from the Spot Vision Screener) will be used as predictors in statistical models that link them to concurrent and future results from the NBT (Cognition, Executive Function, Language, Numeracy, Self-Regulation, Habituation, and Reaction time). A combination of the BHLM methods and Variational Autoencoders will be used to reduce the data into a subset of reliable predictors. The short-term goal of the proposed work is to delineate the developmental trajectory of visuocortical function across infancy and determine the predictive value of a battery of visual measures in relation to concurrent and later cognitive processes. The long-term goal of the proposed work is to establish a reliable and informative index of healthy visual development that captures developmental complexity and can be widely applied in the future.