was chosen based on the Brusini system 3 owing to concerns the Hodapp-Parish-Anderson criteria may underestimate early glaucomatous damage. Diagnosis based on glaucoma progression could only be achieved by a longitudinal prospective design and was outside the scope of our study. Generally, in our clinic, open-angle glaucoma is diagnosed if a person has nonoccludable angles on gonioscopy, glaucomatous optic nerve appearance on fundus exam, and a corresponding defect on imaging (OCT or Heidelberg retinal tomography [HRT]) and/or visual field. When in doubt, patients are followed to determine whether progression occurs. As per standard of care, visual fields are repeated until patients achieve acceptable reliability indices to rule out a learning effect. Although not part of the data collection procedures for this study, patients in the clinic repeatedly undergo ophthalmic imaging in the form of OCT or HRT as part of their follow-up. Third, using a cutoff of e4d B allowed us to include glaucoma patients with all stages of disease: early, moderate, and severe. As Sen and Saxena requested, when we examined whether those with more severe glaucoma had worse cognitive scores, there does seem to be a trend such that the cognitive scores on the verbal digit span tests and the immediate and delayed story tests are worse with worsening glaucoma severity. Finally, the point of using the Mini-Mental State Examination Blind score was to exclude participants for whom we thought their self-reported data might not be reliable or who could not by themselves give informed consent owing to cognitive impairment. We did not attempt to identify mild cognitive impairment in our participants, although it would be likely that some of our participants had it given the low scores on some of the cognitive tests.