Abstract
Purpose :
Visual function is primarily assessed with visual acuity, which is used to infer performance for other activities of daily living. While visual acuity is limited by the ability to resolve fine details, other critical tasks that may be impaired in patients with visual impairment depend on multiple spatial scales, including face perception. To examine whether visual acuity is a reliable predictor of face perception, we measured the effect of simulated dioptric blur (kernel widths: 0, 0.1, 0.2, 0.4, and 0.8°) on these tasks.
Methods :
Participants (N=22) completed an AIM acuity task followed by a FInD face discrimination task. In the acuity task, participants reported the orientation of three grids of 16 Landolt Cs at an adaptive range of sizes, and threshold was estimated from the orientation report error as a function of optotype size. In the face discrimination task, participants were shown four grids of 16 boxes, each of which contained two faces that differed by adaptive distances in the Basel Face Model space. Participants were instructed to click on boxes where they perceived the faces as having different identities and d-prime analysis was used to estimate face discrimination thresholds.
Results :
We found that while acuity thresholds exponentially increased with blur, face identification thresholds were relatively unaffected by blur until larger levels of blur (~0.5°) were applied. The slopes of normalized thresholds as a function of blur were significantly different between acuity and face perception tasks (paired t-test: t(21) = -8.403, p < .001, d = 2.586).
Conclusions :
This demonstrates the resiliency of face discrimination in the presence of blur and suggests that visual acuity is a poor predictor of the face processing performance in visual impairment.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.