Acquired brain injury (ABI) is the leading cause of death and disability worldwide. The degree of severity varies according to a combination of numerous demographics, etiological, clinical, cognitive, behavioral, psychosocial and environmental factors, which can interfere with the effectiveness of rehabilitation interventions and, therefore, with the final outcome. The most important goal of the modern clinic is to predict in time the progression of possible recovery after the brain injury event in order to provide more effective treatment, but the high heterogeneity and clinical variability and the unpredictability of the onset of comorbidities makes this a hard target to reach. In recent years, artificial intelligence algorithms have been applied to more precisely define the role of critical variables that can help clinical practice to predict the final outcome. The classical approach of these algorithms provides only probabilistic values on the final outcome, without considering the typology of clinical interventions and overall complications that may appear throughout the hospitalization period. The objective of this multicentric study is to define a new statistical approach that can describe the dynamics of individual clinical changes occuring during the inpatient intensive rehabilitation care period. The proposed approach combines a principal component analysis (PCA) for dimension reduction (capturing the maximum amount of information and reducing the dimensionality problem) and a nonlinear mathematical modeling for describing the evolution of the clinical course in terms of the resulting new PCA dimensions. By using this approach, we may determine the individual patient's temporal trajectories while examining particular clinical factors. The secondary objective of this study is to validate a new version of the Early Rehabilitation Barthel Index (ERBI), a well-known clinical scale used to measure functional changes in patients with severe acquired brain injury.