Investigating the Relationship Between Physical Activity and Cognition in Children

Date
2018
Authors
McPherson, Adrian
Supervisor
Duncan, Scott
MacKay, Lisa
Item type
Thesis
Degree name
Master of Health Science
Journal Title
Journal ISSN
Volume Title
Publisher
Auckland University of Technology
Abstract

Background A positive relationship between physical activity (PA) and cognitive ability in children has been documented consistently over the last 50 years. Cross-sectional studies show a relationship between PA and cognition in many different groups, ages and in both sexes, and longitudinal studies give promising indications that PA has a positive influence on cognition. Despite that knowledge, experts are still trying to understand the complexities of the PA-cognition relationship [1, 2]. For example, are there other variables that have a stronger influence on cognition? Researchers no longer need to prove the relationship between PA and cognition but to explain the many specific interactions. Is the PA relationship with cognition and academic ability equal or independent? And most importantly: does PA have a causal effect on cognition? The overarching purpose of this thesis was to attempt to answer these questions via two related studies. The aim of the first study was to develop and test a conceptual model that explains the cross-sectional associations among PA, cognition, academic performance, and potential mediating factors in children. Expanding on the first study, the aim of the second study was to develop and test a longitudinal model that examines potential causal relationships between PA, cognition and academic performance in children.

Methods This thesis represents secondary analyses of an existing dataset that comprised information collected from 675 New Zealand school children aged 6-11 years. The data included demographic details, body measures, weekly pedometer step readings, cognitive assessment from the CNS Vital Signs questionnaire, and formal school reading and maths assessment from on the New Zealand Ministry of Education electronic Assessment Tools for Teaching and Learning (e-asTTle). In Study 1, structured equation model (SEM) analyses was used to explain the cross-sectional relationship among three latent variables of PA, cognition, and academic performance. In Study 2, generalised linear mixed models were used to assess changes in PA, cognitive ability, and academic performance at baseline, two month, and six month time points, thereby examining potential causative pathways in the PA-cognition-academic performance relationships.

Results In Study 1, the initial model identified a significant association between PA and academic performance (r = 0.225). The direct association weakened (r = 0.121) when cognition was included in the model, showing the partial mediating effect of cognition. While cognition was strongly associated with academic performance (r = 0.750), PA was also associated with cognition (r = 0.138). Subgroups showed similar patterns to the full sample, but the smaller group sizes limited the strength of the conclusions. Study 2 identified significant relationships between PA change at two months and executive functioning change at six months, reading proficiency change at six months, and maths proficiency change at two months (P < 0.05). No significant relationships were identified for the remaining cognitive domains. Results were adjusted for age, sex, and school socioeconomic decile.

Conclusion The first study used SEM to characterise the direct cross-sectional association between PA and academic performance after controlling for cognition. That is important as it demonstrates PA has independent and cumulative relationships with cognition and academic performance. The second study showed that children who increased PA at two-months displayed small improvements in executive function and reading at six months, and maths proficiency. The findings support the theory that increased PA leads to improvements in cognitive function and academic performance. Future cross-sectional studies should include larger samples to assess patterns for groups of subject including sex, age and SES, and longitudinal studies consider time frames longer than six months to assess longer term effects and patterns.

Description
Keywords
Physical activity , Cognition , Academic performance , School , Children , Mediation , Causation , SEM , Mulitigroup analysis
Source
DOI
Publisher's version
Rights statement
Collections