A method of nystagmus analysis has been developed for clinical and research purposes. Optimum flexibility of analysis is indicated to cope with the problematic signals frequently encountered in pathology. Different modes of analysis may be obtained, varying from fully automatic to step-by-step interactive analysis. Artifact detection and rejection is performed by means of regression analysis with some additional condition tests. The default conditions may be modified or inactivated. A variety of output and display facilities may be used, including permanent storage of adequately labelled retrievable nystagmus data.