[Statistical study of gaze displacement calibration]

The value of quantitative nystagmus analyses depends strongly on the accuracy of the calibration of gaze displacements. Therefore, 120 nystagmograms with one or two controls of the first calibrations (total: 300 calibrations) were evaluated to get a view over the characteristic properties, mainly the variations, of the calibration factors. Systematic and stochastic deviations were taken into account as potential sources of errors. These are mainly ascribed to variations of the corneo-retinal potential and inaccurate visual angles, e.g. as a result of head movements. Statistic evaluation shows no time-dependent variations of the mean and of the standard deviation of the calibration factors. For successive calibrations, relative differences (e2-e1)/e1 were additionally computed. These too, proved to be time-independent, yielding a mean and standard deviation of 2,37 +/- 16,4 %. It can be shown that in 76% of the nystagmograms positive differences are followed by negative ones, and vice versa. In 19% only positive differences and merely in 5% only negative differences were found. It is concluded, therefore, that stochastic variations predominate over the systematic ones. This means that changes in potential are by far of less consequence than patient-induced errors.