Cerebral hemodynamics is greatly altered by severe head injury. The main variables involved in the system are Mean Arterial Pressure and Intracranial Pressure, which are the independent variables (inputs), and Cerebral Blood Flow, which is the dependent variable (output). The relationship between these signals is not completely understood in severe head injury, especially the contribution of the input signals and the linearity of the system. In this work, cerebral hemodynamics is modelled using Finite Impulse Response models with both linear and non-linear Support Vector Machine. Models from 31 patients showed that both input variables contribute similarly and alternatively to Cerebral Blood Flow. © Springer International Publishing Switzerland 2015.