Developing a clinical probability density function for automated perimetry

Abstract

Background: Automated perimetry is associated with lengthy test times, but Baysean predictions can be applied to speed up testing. A critical component of such methods is the starting probability density function (PDF). Methods/Results: In the present study we show that a unimodal PDF, suggested in the literature as adequate for clinical data, fails to describe the thresholds of diseased eyes and we develop a bi-modal PDF representative of a clinical population. Conclusion: We suggest that the implementation of a bi- modal PDF will save test time and retain test accuracy.