Utilising predictive modelling to identify resting state networks predictive of spatial ability and comparing them to other cognitive measures

• Main Author: Ngieng, Shih Yang
• Other Authors: Darren Yeo
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Social sciences::Psychology
• Online Access: https://hdl.handle.net/10356/171897
• Institution: Nanyang Technological University

Spatial ability has been linked to success in future Science, Technology, Engineering, Mathematics roles, and contributes to mathematical achievement, especially in children. Reviews revealed that mental rotation (commonly studied spatial ability) and arithmetic overlap in brain activation. While brain regions activated during mental rotation tests have been identified, studies have not explored if these regions communicate as a network and little is known if these networks are consistent in predicting mental rotation performance. This study utilises the Adolescent Brain Cognitive Development study; a longitudinal study of 9-year-old children across 10 years involving a wide range of measures, to identify networks most predictive of mental rotation, and if the networks predictive of mental rotation concurrently at baseline and prospectively at two time points are stable, and if the networks predictive of mental rotation are also predictive of arithmetic ability. Results from ridge regression analyses on resting-state functional magnetic resonance imaging (rsfMRI) data revealed that connections within the default mode network were the most predictive feature of mental rotation performance, with little involvement of the frontoparietal networks. Moreover, networks predictive of mental rotation concurrently at baseline remain somewhat stable prospectively only at the second time point, but not the third. Aside from predicting mental rotation, the default mode network was predictive of arithmetic and oral reading performance, providing evidence of a domain-general network. This study highlights the possibility of utilising rsfMRI as an indication of future cognitive performance, and that task-based fMRI and rsfMRI may provide different information about cognitive processes.