NEONATAL PREDICTORS OF LANGUAGE PERFORMANCE

Recent advances in evoked potential brain response techniques have shown
this method to be capable of predicting language skills with a high degree
of accuracy. In one study, factor scores based on AERs recorded at birth
were used as independent variables in regression models to predict language
scores obtained from the subject at age 3 years. Seventy-eight percent and
sixty-nine percent of the total variance were accounted for in predicting
language scores on the McCarthy Scales of Children's Abilities and the
Peabody Picture Vocabulary Test, respectively. This research proposal
outlines procedures for assessing the reliability and validity of these
findings with a larger subject population. The subjects will be normal
infants and infants "at risk" for abnormalities in development due to
prematurity, poor postnatal status and perinatal risk factors. Infants
selected as subjects will be those whose parents consent to (1) permit
medical records information to be opened (for birthweight, gestational age,
Apgar and obstetric complications information), (2) allow behavioral
(Bayley, McCarthy and Peabody Scales) and electrophysiological (AERs to
synthetic CV syllables) testing and (3) who are willing to participate for
the three year duration of the project. The results of this study will
permit us to determine if the original predictors continue their high
levels of accuracy in identifying later language skills and to determine
the extent to which electrophysiological measures obtained from more
heterogeneous subjects allow for early discriminations to be made
concerning later language performance. The contribution of AER measures
(factor scores, amplitude measures, and template correlations) will be
compared with measures of perinatal risk, socioeconomic status, home
environment, mother-infant interactions, developmental assessment scores to
determine the relative contribution of these variables to the prediction of
language performance measures.