BACKGROUND Conventional kinetic perimetry is of especial use in case of advanced scotoma. However, examiner dependency is a major drawback. Purpose of this study was to evaluate the clinical feasibility and intraindividual scatter of computer-assisted kinetic perimetry in monitoring advanced visual field defects of various origins. METHODS Examinations were carried out with the Tuebingen Computer Campimeter (background lumincance 10 cd/m2). In an initial session, the localization of the scotoma border is estimated with conventional manual kinetic perimetry. In the subsequent computer assisted kinetic mode, an individually adjusted set of vectors is designed. Each vector crosses the manually assessed visual field defect border almost perpendicularly, starting approximately 3 degrees within the scotoma. Each individual set of vectors can be stored and recalled for follow-up. Stimuli move along these vectors with identical characteristics as in manual kinetic perimetry. Stimulus presentations are repeated six times in a randomized order. Patients' responses are recorded and additionally corrected for mean individual reaction time. A "local kinetic threshold" (mean) and a related parameter for dispersion (standard deviation) are assessed. RESULTS Four subjects with advanced visual field loss of various origin (retinitis pigmentosa, vigabatrin-associated visual field defect, glaucomatous nerve fibre layer defect, and postgeniculate visual pathway defect) participated in this study. Maximal difference between manual-kinetic and automated kinetic thresholds reaches from 1.7 degrees to 5 degrees. Local scatter (standard deviation) of kinetic threshold, assessed by computer-assisted perimetry, varies between 0.1 degree and 3.0 degrees. CONCLUSION Computer assisted kinetic perimetry is a new, useful, examiner-independent, reliable method for effective evaluation and monitoring of advanced visual field loss.