Abstract

Recent advances in cognitive neuroscience emphasise the importance of healthy white matter (WM) in optimal behavioural functioning. It is now widely accepted that brain connectivity via WM contributes to the emergence of behaviour. However, the association between the microstructure of these fibre and behaviour is poorly understood. This is due to indirect and overlapping methods of assessing microstructure, and the oversimplification approaches in assessing behaviour. Here, we used the Mahalanobis Distance (D2) to integrate 10 metrics of WM derived from multimodal neuroimaging that have strong ties to microstructure. The D2 was chosen because it measures the voxelwise distance between every subject and the average, while also accounting for the overlap between the metrics. To examine WM-behaviour associations, we used multivariate regression to examine the voxelwise correlates of 2 cognitive and 2 motor tasks, which allowed us to compare within and across domains in white matter. We observed that behaviour is organised in cognitive, motor, and integrative variables that are widespread in their associations with WM, from frontal to parietal regions. Our results highlight the complex nature of microstructure and behaviour, and show the need for multivariate modelling when examining brain-behaviour associations.