Abstract
Purpose: :
Low vision patients with macular scotomas usually use a preferred retinal location (PRL) of their peripheral retina to read. Some authors have suggested that different PRLs can be used during reading to extract different kinds of word informations. For instance, a PRL below or above the scotoma (vertical strategy) would alternate with a PRL to the left or right of the scotoma (horizontal strategy). We investigated this controversial idea in a population of low vision patients with an absolute macular scotoma.
Methods: :
We used the MP-1 microperimeter to assess the fixation PRL of 13 AMD patients. Static microperimetry maps were measured with the MP-1 to determine the scotomas’ characteristics. Each patient read 14 sentences successively displayed on a monitor while their eye movements were recorded with a video eyetracker (Eyelink II). Included patients had an ETDRS decimal acuity larger than 0.1. Print size was 3x ETDRS acuity threshold. For each correctly read sentence, we analyzed the distribution of vertical fixation coordinates (y) to estimate changes in PRL (vertical versus horizontal). K-means clustering analysis was used with a v-fold cross validation method to extract the number of clusters (1, 2 or 3), on the y dimension, maximizing between-cluster distance and minimizing within-cluster distance. For each sentence and patient, we measured the proportion (P) of fixations in the largest cluster and the ratio (R) of the distance in degrees between clusters (0 deg for one cluster) divided by the radius of each patient’s scotoma. We used the combination of P and R values to define whether patients changed PRL strategies.
Results: :
11 patients used a single PRL strategy during reading while 2 patients used a combination of the horizontal and vertical strategies.
Conclusions: :
It seems that a small proportion of patients does use different PRL strategies. A future question is to investigate whether reading performance of these patients is improved compared to other patients. A larger number of patients is required in order to statistically assess this hypothesis.
Keywords: age-related macular degeneration • reading • eye movements