Psychometrics and Predictive Validity of Infant Learning

PROJECT SUMMARY The ability to detect sensorimotor contingencies between an organism's actions and changes in the environment is essential to survival, impacting social, motor, affective, and cognitive domains. Operant conditioning paradigms provide a direct measure of contingency learning, and are amenable for use throughout infancy. The primary goal of this research is to determine whether individual variations in contingency learning are consistent or correlated across tasks and/or across development, and to determine whether these measures predict future standardized measures of cognition. To this end, we will longitudinally test infants from 6-18 months using two established infant operant conditioning tasks—the synchronous reinforcement paradigm and train task—each involving a different response system (looking vs. lever pressing) and reinforcer. Despite the fact that learning is universally regarded as a fundamental component of human behavior, after over a century of behavioral science, we simply do not know whether or to what degree individual differences in measures of learning in infancy are consistent psychometrically, or whether they are directly associated with broader cognitive and language outcomes. There is substantial demand for the inclusion of measures of learning in clinical trials as potential outcomes for early interventions, but these fundamental issues have never been previously established or investigated in infancy. This proposal features two investigators with direct expertise in operant learning paradigms for infants and in conducting longitudinal projects; here, they seek to establish the fundamental psychometric properties of individual differences in classic parameters of operant conditioning within ages across tasks and across ages, and to establish the developmental course of these parameters. We also seek to determine the predictive validity of these individual differences, whether the indicators themselves or their developmental functions are predictive of standardized 24-month outcomes. We propose to use two paradigms that utilize different operants (looking and motor actions) each of which have previously been established as valid and workable in published literature. Learning within traditional conditioning paradigms has long been neglected in the area of human infancy; this study has the potential to resurrect learning in infant studies and developmental science. Given that differences in contingency learning have been noted in at-risk infant samples and learning is often used in preclinical animal work, this work would serve to provide a connection between preclinical studies and clinical trials. Our findings will have critical implications for the use of learning in clinical trials and long-term outcome studies.