Abstract
Different visual factors like character size, contrast or font have been reported in the clinical litterature to affect reading speed for subjects with macular scotomas. In this study, using a gaze-contingent display, we examined whether increasing line-spacing improves reading performance in subjects with an artificial central scotoma.
7 normally-sighted subjects (23– 43 years) had to read sentences as fast as possible and without making mistakes. On each trial, a sentence, covering 3 or 4 lines, was displayed on the monitor. We used a set of 1000 sentences extracted from French litterature with strict constraints inspired from the MNREAD chart (e.g. number of characters for each sentence and word, frequency of each word). Line spacing was 0.5×, 1× or 2× the standard spacing, and letter size (x-height) varied from 0.52 to 2 degrees (step: x1.4). Gaze location was recorded at 500Hz with an EyeLink II eye-tracker and was used to draw a square-shaped gaze-centered mask (6×6 or 10×10 degrees). Reading time for each sentence was measured to compute reading speed.
Our main finding is that interline spacing does not affect reading speed, contrary to a recent RSVP study (Chung, 2004). We also find that the modulation of reading speed as a function of character and scotoma size has the same characteristics as those reported in the clinical litterature, suggesting that our paradigm is a valid scotoma simulation tool. Altogether, our results suggest that vertical crowding is not a major limiting factor of reading speed when the eyes are free to move. From a practical point of view, the lack of effect of line spacing on reading speed suggests that an optimal aid for low-vision patients should magnify print size while keeping relatively small line spacing. This would allow to display more text information within a given area.