Theses Doctoral

Learning Across Development: Social and Neural Associations in Toddlerhood, Middle Childhood, and Adolescence

Harmon, Chelsea Marie

Contemporary cognitive psychology and neuroscience contend that learning across development is a dynamic process; that learning is flexible, adaptive, and context-dependent. Moreover, particular learning demands, capacities, and sensitivities to external (i.e., social, environmental, etc.) and internal (i.e., biological) influences change over the course of development, depending on specific developmental niches, or periods of unique specialization. Given these theoretical contexts, this dissertation examines developmental niche-relevant influences on learning at three different stages of development, such as social modulators (i.e., parental proximity, and caregiving-related early adversity) and neural mechanisms (i.e., neural dynamic network flexibility).

Studies 1 and 2 consider the parent-child attachment relationship and tested associations between parental proximity and classical conditioning reward learning in toddlers utilizing a within-subject study design. Findings suggest parental proximity and presence had a significant effect on reward learning outcomes. However, effects varied as a function of age and order of the parental presence versus absence manipulation. Study 3 investigates the relationship between caregiving-related early adversity (crEA) in school-age children and implicit motor sequence learning (MSL), a form of basic learning. There was no difference in MSL between crEA-exposed and age-matched, non-crEA-exposed, comparison children. However, group differences in executive functioning (EF) were present.

Findings illuminate the selective impact of early stress on higher-order functioning (i.e., EF), but not on possible underlying lower-level processes (i.e., basic learning). Lastly, Study 4 employed a novel approach to quantify brain network changes over the course of learning, dynamic network flexibility (DNF), to examine associations with learning on a probabilistic reinforcement learning task in a sample of adolescents who perform better on the task compared to adults. This study provides evidence that average whole brain DNF, and DNF of the striatum in particular, is associated with reinforcement learning in adolescents, as was previously found in adults. However, DNF did not explain better learning outcomes observed in the adolescent sample.

Each study separately contributes to the growing knowledge of social and neural associations with learning abilities that are not exclusive to, but important for study-specific developmental niches. Taken together, this dissertation provides evidence in support of a dynamic systems theory of cognition and asserts that learning and behavior during development is a dynamic process that is dependent on social input, adaptive to early experiences, and is supported by flexible neural network architecture.


This item is currently under embargo. It will be available starting 2028-05-12.

More About This Work

Academic Units
Thesis Advisors
Tottenham, Nim
Ph.D., Columbia University
Published Here
May 24, 2023