Discrimination of language skills at five years of age using event‐related potentials recorded at birth

Abstract

Limited success has been noted in prior attempts to predict language and cognitive performance during the toddler and preschool years from measures obtained at birth. Most perinatal measures, although demonstrating some success in predicting motor and early cognitive skills at 1 year of age, have generally failed to maintain their predictive effectiveness when used to predict abilities that emerge later in development. However, 1 electrophysiological technique that involves the recording of event‐related potentials (ERPs) to speech sounds may provide a means to improve such predictions. This article describes a successful extension of an earlier attempt to predict language performance during the preschool years from neonatal scalp recorded ERPs. Auditory ERPs were recorded from the frontal, temporal, and parietal scalp regions of 71 newborn infants in response to a series of 9 consonant‐vowel syllables. Following artifact rejection and averaging, these ERPs were input to a principal components analysis that isolated 7 factors to account for 89.02% of the total variance in the data set. Selected factor scores from 2 factors matching the latency configuration of those identified earlier by D. L. Molfese and V. J. Molfese (1985) were then used in a series of discriminant function analyses to separate these infants into 2 groups based on performance on the verbal subtest of the Stanford‐Binet at 5 years of age. Three different discriminant function analyses involving different variables showed high classification accuracy in predicting later developmental outcomes.