"Look Versus See": Does Varying Fellow Eye Contrast Affect Perception of the Amblyopic Eye?

PURPOSE: This study aimed to evaluate how modulating fellow eye (FE) contrast during dichoptic visual search impacts amblyopic eye (AE) perception and oculomotor behavior in amblyopia. By leveraging eye-tracking data, we investigated how FE contrast and sensory-motor factors influence perceptual outcomes, distinguishing successful attentional allocation and target identification ("look and see") from "look but failed to see" (LBFTS) and "no look, no see" errors.

METHODS

All participants performed visual search under monocular and dichoptic viewing with varied FE contrasts. High-resolution eye-tracking captured microsaccades (<1°), saccades (>1°), drift velocity, eye deviation via the density-based spatial clustering of applications with noise algorithm (DBSCAN) and attention patterns (heatmaps, scan paths). Search efficiency was measured as the time from image onset to the final fixation within the region of interest (look time), and perceptual accuracy was assessed via mouse-click localization of the target.

RESULTS

In amblyopic participants, full-contrast FE dichoptic viewing impaired AE target detection and prolonged look time. Reducing FE contrast improved AE performance-especially in those with mild suppression and good stereopsis-by shortening look time and reducing LBFTS errors. However, excessive FE contrast reduction (≤25%) impaired FE target detection in controls and amblyopia participants. The "look and see" trials showed higher microsaccade frequency, whereas LBFTS and "no look, no see" (i.e., incorrect) trials were linked to increased saccades, eye deviation, and increased FE drift.

CONCLUSIONS

Contrast rebalancing can enhance AE attentional and perceptual performance, but optimal FE contrast varies with individual sensory-motor profiles. Our findings underscore the value of integrating visual search paradigms with eye tracking to personalize dichoptic therapies and objectively monitor treatment efficacy.