Figure 2.

Statistical analysis and data analysis plan. We first analysed the behavioural data arising from the explicit and implicit agency tasks. Behavioural outcome measures showing significant differences between groups were considered for neuroimaging analysis. We assessed associations between behavioural markers of sense of agency and structural metrics in regions of interest (ROIs), which show alterations of grey matter and diffusion weighted imaging measures in patients. We then characterized the functional coupling (functional connectivity) between the ROIs which showed a degree of structural impairment related to altered sense of agency. Lastly using dynamic causal modelling (DCM) with Parametric Empirical Bayes (PEB), we evaluated the direction of the coupling (i.e. which area drives the activity of other ROIs).