Login (DCU Staff Only)
Login (DCU Staff Only)

DORAS | DCU Research Repository

Explore open access research and scholarly works from DCU

Advanced Search

Computational Approaches to the Assessment of Motor Competence in Children: Prediction, Interpretation and Markerless Measurement

O'Donaghue, Colm (2026) Computational Approaches to the Assessment of Motor Competence in Children: Prediction, Interpretation and Markerless Measurement. PhD thesis, Dublin City University.

Abstract
Motor competence (MC), considered key in supporting child development, is associated with physical activity (PA) and broader lifelong health outcomes. However, traditional assessment and analytical approaches remain constrained by labour-intensive procedures that may not fully capture the complexity of MC development. Consequently, there is growing interest in computational approaches to address such limitations. This thesis examined the application of computational methodologies within MC research across four related studies. Study 1 evaluated machine learning (ML) classification models for predicting actual motor competence based off remaining actual (AMC) and perceived motor competence (PMC) data. Results demonstrated that several MC outcomes could be accurately predicted, highlighting the potential of ML within MC research. Building on these findings, study 2 utilised post-hoc model interpretability and graph analytics to characterise interdependencies between AMC and PMC by age and gender. Findings highlighted influential skills and constructs within MC networks while reinforcing the importance of PMC as an important correlate of broader MC development. The third study, a position paper, critically evaluated the opportunities, limitations, and future directions of markerless motion-capture technologies for MC assessment. The review noted the potential of computer vision systems to enhance the scalability and objectivity of assessment while considering methodological and practical limitations requiring further investigation. In response, study 4 explored the feasibility of integrating an open-source markerless motion-capture pipeline (‘FreeMoCap’) with a heuristics- based rules engine to automate classification of TGMD-3 Short Form locomotor performance criteria. Findings suggested that classification performance varied between performance criteria but remained relatively stable across camera configurations. While reducing camera quantity influenced kinematic data quality, a three-camera configuration provided a practical balance between data quality, classification performance, and deployment considerations. Collectively, this thesis demonstrates the value of computational techniques for advancing MC research and provides a foundation for future technology-enhanced assessment and analysis.
Metadata
Item Type:Thesis (PhD)
Date of Award:27 August 2026
Refereed:No
Supervisor(s):Behan, Stephen, Scriney, Michael and Belton, Sarahjane
Subjects:Social Sciences > Health (social science)
School, Faculty or Unit:Faculty of Science & Health
Faculty of Science & Health > School of Health & Human Performance
Use License:This item is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 License. View License
Funders:Science Foundation Ireland CRT: ML-Labs
ID Code:33479
Deposited On:09 Nov 2026 15:09 by Stephen Behan . Last Modified 09 Nov 2026 15:09
Documents

Full text available as:

[thumbnail of Final Thesis Document for Submission 27-08-26.pdf]
Preview
PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Creative Commons: Attribution-Noncommercial-No Derivative Works 4.0
3MB
Downloads

Downloads

Downloads per month over past year

Archive Staff Only: edit this record