Purpose. Previous studies ha\e shown that reading speed is slower in peripheral vision, even when character si/e i.s not a limiting factor, and when oculomotor demands are minimiJ.ed with rapid serial visual presentation (RSVP). The visual span in reading is the nutnher of characters that can be recognized in a glance. We tested the hypothesis that slower reading speed m peripheral vision is due to a reduction in the visual span. Method. The visual-.span hypothesis predicts thai reading lime (msec/word) becomes increasingly dependent on word length (steeper slope) with increasing retinal eccentricity. Five normally sighted subjects were tested with RSVP in (he fovea and at 2.5. 5. 10 and 15° in the inferior visual field. Eye movements were monitored. In a trial, the subject read aloud 6 unrelated words of fixed length. A staircase method was used to estimate threshold reading times for word lengths ranging from 4 to 10 letters. At each eccentricity, the character siJ.e was made large enough to support maximum reading speed (i e , twice the critical print size estimated from previous research). Results- Consistent with the visual-span hypothesis, plots of reading time vs word length became steeper with increasing eccentricity, with a statistically significant interaction between word length and retinal eccentricity. Inferred values of the visual span-computed troin the reciprocal slopes and nominal fixation times of 250 msecranged Irom at least 10 chaiacters for foveal viewing to a mean of 2.78 characters at 15° eccentric] U Conclusions. In the normal visual field, fewer letters can be recognized at one time at increasing eccentricities (reduced visual span), despite enlarging letters to compensate for lower spatial resolution. This result helps to explain deficits encountered by lowvision readeis with central-field loss.
|Original language||English (US)|
|Journal||Investigative Ophthalmology and Visual Science|
|State||Published - Dec 1 1997|