Abstract
Purpose: Recent studies demonstrate that shape discrimination sensitivity is significantly reduced in early macular degeneration. This could be due to the structural changes in the cone mosaic caused by the dysfunction/death of photoreceptors. In this study the impact of undersampling, irregular sampling and parafoveal scotoma on the performance of shape discrimination was evaluated. Methods: A series of simulation experiments were conducted. The stimuli were the outputs of a retinal sampling model with a normalized, balanced DOG (difference of Gaussian) as the postreceptoral component. The inputs to this sampling model were radial frequency (RF) patterns. The sampling array used in the model was based on a Cartesian hexagonal lattice. Undersampling was implemented by reducing sampling density. Irregular sampling was introduced by adding Gaussian noise to the location of sample points. Parafoveal scotomas were simulated by removing patches of samples in the paracentral area of the sampling array to cover parts of RF patterns. A temporal 2AFC staircase paradigm was employed. The task was to identify which interval in a trial contained the RF pattern. Results: When the resolution limit is more than 2 times above the peak spatial frequency of RF patterns, changing of sampling density has little effect on the detection threshold. With the increase of sampling irregularity, threshold increases even for well-sampled RF patterns. For a given sampling density, the thresholds obtained with irregular sampling are within 2 times of that obtained with regular sampling mosaic. The larger the scotoma, the higher the threshold required for detecting RF patterns, even for regularly well-sampled RF patterns. The presence of multiple scotomas further disrupts the performance. Conclusion: While irregular/under-sampling can affect the ability of patients to detect RF patterns, parafoveal scotomas appear to be a major factor responsible for the loss of shape discrimination sensitivity in early macular degeneration.