Abstract
Purpose :
Recent studies have highlighted the specific impact of linguistic word features, such as frequency and orthographic neighborhood size, on the reading performance of individuals with Central Field Loss (CFL). Here, we focus on the specific effect of word orthographic similarity, measured with the Levenstein's orthographic distance (OLD20). OLD20, defined as the distance between a word and its 20 closest neighbors, is a measure of the visual ambiguity yield by words composed of similar letters: the higher OLD20, the lower the orthographic similarity. We hypothesize that orthographic similarity is more detrimental to reading performance with CFL compared to normal vision, because it increases visual ambiguity.
Methods :
40 young normally sighted participants read short sentences presented sequentially on a screen, using the self-paced reading paradigm, while eye movements were recorded using an EyeLink1000. 26 of them performed the task while a 10°-diameter circular scotoma centered on their gaze was displayed. The other 14 participants read with no scotoma. Each participant read 304 sentences over two 2-hour periods. Reading time was recorded for each well-read word. Linear mixed-effects models were used to inspect the effect of OLD20 on reading time, taking into accound word frequency, length and in-context predictability. Importantly, we estimated how this effect changed throughout the adaptation to the presence of the scotoma.
Results :
In the scotoma group, initial reading time was 50% longer for lower OLD20 values than for higher values (p<.001). Towards the end of the study, the effect of OLD20 reversed, resulting in a 14% decrease in reading time for the highest OLD20 values (p<.001). In the control group, reading time was 17% shorter for low OLD20 values compared to higher values (p<.001), which persisted throughout the experiment.
Conclusions :
In normal readers, orthographic similarity has a facilitator effect on reading performance: the lower OLD20, the greater the orthographic similarity, the shorter the reading time. In the presence of a simulated scotoma, this effect is reversed, at least at the onset of scotoma exposure, suggesting that the visual ambiguity led by orthographic similarity may slow down reading processes. Interestingly, this effect changes as participants adapt to the presence of the scotoma, showing an effect of orthographic similarity comparable to what is observed in controls.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.