Quantitative gait analysis has been used to elucidate characteristic features of neurological gait disturbances. Although a number of studies compared single patient groups with controls, there are only a few studies comparing gait parameters between patients with different neurological disorders affecting gait. In the present study, gait parameters were compared between control subjects, patients with parkinsonian gait due to idiopathic Parkinson's disease, subjects suffering from cerebellar ataxia and patients with gait disturbance due to subcortical arteriosclerotic encephalopathy. In addition to recording of baseline parameters during preferred walking velocity, subjects were required to vary velocity from very slow to very fast. Values of velocity and stride length from each subject were then used for linear regression analysis. Whereas all patient groups showed slower walking velocity and reduced step length compared with healthy controls when assessed during preferred walking, patients with ataxia and subcortical arteriosclerotic encephalopathy had, in addition, increased variability of amplitude and timing of steps. Regression analysis showed that with changing velocity, subjects with Parkinson's disease changed their stride length in the same proportion as that measured in controls. In contrast, patients with ataxia and subcortical arteriosclerotic encephalopathy had a disproportionate contribution of stride length when velocity was increased. Whereas the findings in patients with Parkinson's disease can be explained as a reduction of force gain, the observations for patients with ataxia and subcortical arteriosclerotic encephalopathy reflect an altered spatiotemporal gait strategy in order to compensate for instability. The similarity of gait disturbance in subcortical arteriosclerotic encephalopathy and cerebellar ataxia suggests common mechanisms.