Abstract
Purpose :
Contour interaction (crowding) is the substantial interference of neighboring contours on target identification, and has been studied primarily in the visual periphery, with conflicting results for foveal stimuli. While the cortical locus for peripheral crowding is well established (with a large extent up to half the eccentricity), disentangling the contributing factors in the fovea is more challenging due to optical limitations. Here, we used adaptive optics to overcome ocular aberrations and employed high-resolution stimuli to precisely characterize foveal contour interactions.
Methods :
Stimuli (white light) were displayed on a DLP projector and relayed to the eye via an adaptive optics system through a 6mm artificial pupil. Subjects’ pupils were dilated and accommodation was paralyzed. Relative vergence between the wavefront sensing beacon (840nm) and the projector was subjectively minimized using trial lenses. Accuracy in identifying the orientation of a Tumbling-E was measured with AO and with only sphero-cylindrical correction. Optotype sizes ranged from 20/8 to 20/20 and viewing time was unlimited. When flanked, Tumbling Es appeared on all 4 sides of the target, with edge-to-edge spacings from 0 (abutting) to 1.6 arc-min. Performance vs. flanker spacing was modeled with a modified logistic function (enhanced for flanker facilitation at close spacings) and fit with Markov chain Monte Carlo (2000 fits per condition). Critical spacing was computed as the flanker distance yielding 95% of asymptotic (unflanked) performance.
Results :
Correcting higher-order aberrations improved acuity for each subject by ~0.18 logMAR (Mean AO thresholds: -0.38±0.07; non-AO: -0.20±0.08), allowing the testing of small stimuli inaccessible without AO. Flanked performance vs. spacing curves were consistent across correction type and target size, with identical slopes and shape defined solely by per-condition asymptotic performance. Flankers within 0.4 arc-min improved performance for most (8 of 11) conditions. Critical spacings ranged between 0.5 and 1.2 arc-min (mean=0.75), anti-correlated with stimulus size (r=-0.7).
Conclusions :
Under optimal foveal viewing conditions, the zone of interference from surrounding contours is 0.5-1.2 arc-min, significantly smaller than earlier estimates (2-5 arc-min), but large enough to implicate non-optical factors. A single parametric template captures the complex shape of flanker interactions across a variety of conditions.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.