The cerebrospinal fluid (CSF) infusion test is used to evaluate the dynamics of CSF circulation in patients with communicating hydrocephalus and is based on constant-rate infusion of the normal saline into cerebrospinal fluid space. The aim of the study was to refine methods of the analysis of intracranial pressure (ICP) recorded during and after the infusion test. The mathematical model of cerebrospinal fluid circulation was extended by the equation describing ICP decrease after the infusion. The nonlinear least-squares method of Levenberg-Marquardt was used to estimate the parameters describing the CSF compensatory mechanisms. Twenty-seven infusion tests were studied. Both phases of the test-the increase and the decrease of ICP-were recorded and the compensatory parameters were calculated for each of them. ICP often does not return to the resting level after the infusion test within the period equivalent to the time of infusion in all cases. In 20 tests the differences between post- and pre-infusion resting ICP (DeltaICP) was higher than 1 mmHg, which was considered as significant. The mean value of DeltaICP for 20 infusion tests was 3.0 +/- 0.7 mmHg. The cerebral elasticity evaluated during the infusion was greater than the elasticity estimated from the decreasing phase after the infusion (0.24 +/- 0.07 ml(-1) versus 0.14 +/- 0.03 ml(-1); p < 0.01).