Time-dependent interactions between pressure, flow and volume of cerebral blood and cerebrospinal fluid were mathematically modelled. The model was designed to simulate blood inflow and storage, arteriolar and capillary blood circulation controlled by cerebral autoregulation, venous blood outflow and storage modulated by intracranial pressure, and cerebrospinal fluid production, storage and reabsorption. The software implementation of the model was used to calculate the response to a gradual decrease in cerebral perfusion pressure corresponding to either systemic hypotension or intracranial hypertension. We computed flow pulsatility index (PI) and short range correlation coefficients between systolic, diastolic and mean flow velocity (FVs,d,m) and mean cerebral perfusion pressure (CPP). In simulation, the changes in cerebral flow produced by intracranial hypertension and systemic hypotension were practically indistinguishable. The relationship between PI and CPP was reciprocal, independent of the state of autoregulation. The short range running correlation coefficients between FVs, FVm and CPP indicated both combined safe CPP range and preserved autoregulation, promising a clear clinical detection of "non-worsening" blood supply conditions. A similar procedure was applied to selected clinical data to illustrate the theoretical considerations.