Amblyopia (lazy eye) is a developmental disorder of the visual cortex that causes deficits in visual acuity and shape perception. The loss of visual acuity is thought to originate from weakened cortical responses to stimuli. Here, we provide evidence for a similar mechanism to explain distortions in shape perception. We introduce a computational model that simulates perceptual distortions of grating patterns drawn by humans with amblyopia (Barrett et al., 2003). The model simulates a large variety of distortions by performing a weighted sum of rectified sinusoidal gratings (average, 3.3 gratings ∼6 times larger than foveal receptive fields in the primary visual cortex) with different dark/light duty cycles. The simulations accurately reproduce self-reported perceptions of amblyopic patients and decrease drawing-percept differences when ideal percepts (stimuli) are replaced with simulated percepts (9.03 ± 12.37%; p = 0.0002; Wilcoxon test comparing normalized Laplacian pyramid distances; Laparra et al., 2016). The simulations also reveal an increase in the number of stimulus orientations contributing to visual percepts in amblyopia and a strong correlation between contrast sensitivity deficits and both magnitude of perceived visual distortions (r = 0.96; p = 0.0007) and predicted spread of cortical activation (r = 0.82; p = 0.02). The results also demonstrate a compensatory shift in the spatial frequency distribution of cortical filters in amblyopia, which closely resembles the spatial frequency shift caused by contrast reduction in thalamocortical inputs of male cats. Taken together, our results indicate that amblyopia compensates weakened cortical responses by increasing the spread of cortical activation to include neurons with mismatched stimulus preferences that cause perceptual distortions.
Keywords: LGN; computational modeling; developmental disorder; shape perception; visual cortex.
Copyright © 2025 Olianezhad et al.