Associations between 24-hour Behavior Compositions, Memory and Hippocampal Volume in Preschoolers

• In Event: Poster Session 02
  In Poster Session: PS 02 Topic - Attention, Learning, Memory

Abstract

Introduction: Early childhood serves as an important time for brain and cognitive development and includes critical years for developing healthy activity profiles and sleep habits. Although sedentary behavior, physical activity, and sleep (referred to as 24-hour behaviors) have been independently associated with cognitive and brain outcomes in older children and adults, evidence is limited in early childhood. Additionally, most studies have examined sedentary behavior and physical activity independently of sleep and these behaviors may interact to influence neurocognitive outcomes. The proportion of time spent in each 24-hour behavior is inherently co-dependent with the other behaviors given that they occur within a fixed period. Traditional analyses using absolute values for variables that are components of a composition do not account for the closed nature of the data. The emerging field of time-use epidemiology has promoted a framework that identifies the need to collectively examine 24-hour behaviors with physical and mental health outcomes with appropriate statistical analyses that account for the co-dependent nature of these behaviors. A recent approach that has gained support is the application of both compositional data analysis (i.e., converting each 24-hour behavior into a set of isometric log-ratio coordinates that can then be used in conventional statistical models) and isotemporal substitution (i.e., estimating effects on an outcome from substituting time spent in one 24-hour behavior with other behaviors). Purpose: To determine if 24-hour behaviors are associated with memory and hippocampal volume in early childhood. Hypotheses: It is hypothesized that 1) the composition of 24-hour behaviors will be positively associated with memory and hippocampal volume, and 2) time reallocations that either increase moderate-to-vigorous phyiscal activity or sleep, or decrease sedentary time will be postively associated with memory and hippocampal volume. Methods: Cross-sectional analyses will be conducted with data from a study examining the benefits of napping on memory in early childhood (already available data: n = 56 children ages 33-71 months). Measurements assessed at baseline of the parent study include time spent in each of the 24-hour behaviors from 3-16 days of actigraphy (i.e., accelerometry via a wrist monitor), declarative memory performance from a visuospatial assessment, and hippocampal volume (total, hemispheric, head, body, and tail volumes) from magnetic resonance imaging. Planned Analyses: Compositional data analyses using linear regression models will be used to examine each relation of interest. Twenty-four hour behaviors will be transformed into isometric-log ratio coordinates to address concerns of multicollinearity. Models will include the isometric log-ratio coordinates as the independent variables and each of the memory performance and hippocampal measures as dependent variables. Each model will be adjusted for age, gender, and intracranial volume (for hippocampal models only). Isotemporal substitution will be used to compare estimated effects of 10-minute time reallocations between 24-hour behaviors on the outcome measures. Data collection has recently concluded and data is currently being cleaned for the proposed analyses. Implications: Understanding how 24-hour behaviors are related to memory performance and brain development in early childhood will provide insight into the optimal composition of daily movement and sleep behavior patterns in early childhood.

Authors

• Christine Woodward St. Laurent, University of Massachusetts Amherst
  Presenting Author

• Sanna Lokhandwala, University of Massachusetts Amherst
  Non-Presenting Author

• Tamara Lynn Allard, University of Maryland - College Park
  Non-Presenting Author

• Angela Ji, University of Maryland - College Park
  Non-Presenting Author

• Tracy Riggins, University of Maryland - College Park
  Non-Presenting Author

• Rebecca Spencer, University of Massachusetts Amherst
  Non-Presenting Author