In routine clinical electronystagmographic (ENG) tests (postrotatory, optokinetic, caloric and tracking tests), eye movement signals were analyzed and a multiple discriminant analysis was carried out with the aid of a microcomputer. Six parameters were selected and, based on these, two functions for discriminating between peripheral and central disorders were established. Discrimination between 35 patients with peripheral lesions and 15 patients with central lesions was made with a correct classification rate of 97.1 and 86.7 %, respectively. These rates are significantly higher than that of any single ENG test analysis. Our results indicate that the clinical application of ENG can be improved by searching for more sensitive ENG parameters and adopting the comprehensive analysis approach.