Visual Performance assessment using a mobile platform

Abstract

Purpose: To propose and validate a new tablet-enabled test for screening and evaluation of visual function in glaucoma, the PERformance CEntered Portable Test (PERCEPT)

Methods: This cross-sectional study involved 71 patients with glaucomatous visual field defects on standard automated perimetry (SAP) and 59 control subjects. The PERCEPT was based on the concept of increasing visual task difficulty to improve detection of central visual field losses in glaucoma patients. Subjects had to perform a foveal 8-alternative-forced-choice orientation discrimination task, while detecting a simultaneously presented peripheral stimulus within a limited presentation time. The test was performed at low contrast. Subjects also underwent testing with the Useful Field of View (UFOV) divided attention test. The ability to predict history of motor vehicle crashes (MVC) and falls was investigated. Driving records were obtained from the California Department of Motor Vehicles. History of falls was evaluated by a standard questionnaire. Analyses were adjusted for age, binocular visual acuity, Pelli-Robson contrast sensitivity and score of cognitive impairment. Main Outcome Measures: Areas under the receiver operating characteristic (ROC) curves to discriminate glaucomatous from control subjects; and ability to predict history of MVC and falls, as assessed by odds ratios (ORs) and incident-rate ratios (IRR), respectively.

Results: When adjusted for confounding variables, only the PERCEPT processing speed (PS) parameter showed significantly larger values in glaucoma compared to controls (difference: 194ms; P<0.001). For SAP 24-2 binocular mean sensitivities of 26dB, 23dB, and 20dB (corresponding to mild, moderate and severe visual field losses), ROC curve areas for the PERCEPT PS parameter were 0.77, 0.92, and 0.98, respectively. PERCEPT PS performed better than UFOV in predicting MVC and falls. Each 1SD increase in PERCEPT PS was associated with an OR of 2.69 (P=0.037) for predicting MVC and with an IRR of 2.05 for predicting falls (P=0.006).

Conclusions: A portable platform for testing visual function was able to successfully detect visual field losses in glaucoma and predict risk of falls and MVCs. Due to its simplicity, low cost, and diagnostic value, the proposed platform could be an effective tool for screening of functional impairment.