Hatem Barhoom

Abstract

Sloan letters are commonly used optotypes in clinical practice, but they exhibit different relative legibility, which may be attributed to response bias, sensitivity differences, and letter similarity. In this thesis, we employed Luce’s choice model and developed a new noisy template model to investigate the role of response bias, sensitivity differences, and letter similarity in letter identification of Sloan letters at central and paracentral locations. Results show that the best model was the one that accounted for the effects of bias, sensitivity, and similarity, with bias contributing more than sensitivity and similarity. However, when estimating the letter acuity from the pooled data across all letters, no significant effects of bias, sensitivity, or similarity were observed. The models incorporating similarity demonstrated a substantial increase in the spread of the underlying psychometric function (the percent correct as a function of letter size), particularly in the periphery and upper portion of the function. Given that most letter stimuli in clinical vision tests are presented at supra-threshold sizes, it is plausible to attribute any increase in test-retest variability, particularly in peripheral vision, to similarity alone. Furthermore, this thesis explored the use of letters as stimuli and speech as a response method to assess macular visual sensitivity in healthy observers and individuals with 6 Glaucoma. Dissimilar letters following Sloan’s design were used to estimate peripheral letter acuity within 4 degrees of fixation. A speech recognition algorithm was employed to enable participants to perform the task without supervision. The participants’ perceived task difficulty was assessed through a questionnaire. Results from this experiment show that in observers with Glaucoma, letter acuity exhibited a close correlation with conventional perimetry, and most observers found the task easy to perform. These results demonstrate that letter acuity perimetry with speech input is a viable method for capturing macular damage in Glaucoma. These approaches have the potential to facilitate more intuitive and patient-friendly tests for macular visual field assessments.

Keywords

paracentral vision, Letter acuity, Home-based test, letter identification Task, Luce's choice model, Noisy Template model, letter similarity, response bias

Document Type

Thesis

Publication Date

2023

Download

Share

COinS