A method of estimating the likelihood of brain concussion from neurophysiological data acquired from the brain of the subject is disclosed. The method comprises: identifying activity-related features in the data; constructing a subject- specific brain network activity (BNA) pattern having a plurality of nodes, wherein each node represents a feature of the activity-related features, and each pair of nodes is assigned with a connectivity weight; calculating a BNA pattern similarity describing a comparison between the constructed BNA pattern and a baseline BNA pattern being specific to the subject; and assessing the likelihood of brain concussion responsively to the BNA pattern similarity.
