Identifying Clinical Subtypes of Patients with Post-Concussion Syndrome: Evaluating Brain-Behaviour Relationship with a Computational Approach

To better capture the heterogeneity of the symptoms of patients with Post-Concussion Syndrome (PCS), we hypothesized that brain changes could be used to separate the patients into clinical subtypes. Subject specific structural and functional connectomes were created based on Diffusion Weighted Imaging (DWI) and Resting State Functional Magnetic Resonance Imaging (rs-fMRI), respectively. We identified the structural and functional connectivity differences between the PCS patients (N=83) and the healthy controls (HC, N=13) with a random forest algorithm; then identified subtypes of PCS, using a gaussian mixture model. Based on structural and functional connectivity, we observed 2 subtypes; a sub-group that had similar connectivity and behavioural profiles (i.e. neuropsychiatry, symptoms) as the HCs and a subgroup that showed great deviation from the first subgroup and HCs. Defining PCS subtypes will allow us to move towards precision medicine in PCS patients and a better understanding of the heterogeneity in presentation and recovery.M.Sc