Abstract
Interaction between objects in close proximity could occur spatially and/or temporally. The purpose of this study was to investigate whether peripheral reading speed could improve by reducing the spatio-temporal interaction between successive words of a sentence. We hypothesized that presenting words at alternating spatial locations would reduce the spatio-temporal interaction of successive words and thus enhance reading performance. Eight observers with normal vision (aged 20 to 24) read aloud single sentences, presented one word at a time, using rapid serial visual presentation (RSVP). Testing was conducted at the fovea and 10° eccentricity in the inferior visual field. Print sizes were larger than the critical print size at each eccentricity. Words of each sentence were presented at the same location on the display (0 separation: control), or alternated between two vertically separated locations (1×, 1.4×, 2×, 2.6×, 3.6×, 5.2× the x-height). For each condition, the number of words read correctly was measured for six RSVP word exposure durations that spanned approximately 1 log unit, with six sentences per duration. Reading speed was calculated based on the duration that yielded 80% accuracy. At the fovea, the geometric mean reading speed for the eight observers ranged between 676 and 799 wpm and did not vary with vertical separation. At 10° eccentricity, however, reading speed did depend on vertical separation (p<0.0005). The geometric mean reading speed was 242 wpm for the control (zero-separation) condition, but was fastest at a vertical separation of 1.4× x-height (14% improvement, p = 0.042) and slowest at a vertical separation of 5.2× x-height (15% reduction, p = 0.003). Our results suggest that presenting words at two slightly separated spatial locations can reduce the spatio-temporal interaction between successive words and enhance reading speed in the periphery. These findings have important rehabilitative implications for patients with central vision loss.
Supported by the Stimson-Duvall Fellowship and NIH research grant R01-EY012810.