Purchase this article with an account.
Jennifer M. Scholz, Gordon E. Legge, Tingting Liu, Christopher S. Kallie; Comparing the Visual Span for Letters and Faces. Invest. Ophthalmol. Vis. Sci. 2012;53(14):4813.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
The visual span for reading is the number of adjacent letters that can be recognized reliably without moving the eyes. It is about 8 to 10 letters in normal central vision, depending on the accuracy criterion. Reduction of the size of the visual span is associated with decreased reading speed in low vision. Since sensory limitations are thought to limit the size of the visual span, it is likely that similar limitations exist for other types of pattern recognition. We tested this possibility by measuring the visual span for another important set of targets--faces--and compared the results with the visual span for letters. The results will provide a normal baseline for our future research on face recognition in low vision.
The stimuli were 26 letters of the English alphabet and 26 grayscale images of familiar celebrity faces; the center-to-center spacing of adjacent stimuli was 4.6º. Visual span profiles were measured for five normally sighted subjects by plotting recognition accuracy as a function of stimulus position (in units of the number of letter- or face-widths) to the left and right of the fixation point. There were four stimulus conditions: single letters, single faces, trigrams (random strings of three adjacent letters), and trifaces (random triples of three horizontally adjacent faces). During each computer-based trial, participants were asked to report the letters or faces that appeared for 250 ms at one of thirteen positions (-6 to +6, centered at 0).
For a recognition criterion of 80% correct, the visual spans based on the trigrams and trifaces were nearly identical--nine letters and nine faces. Outside the central nine positions, recognition accuracy dropped more rapidly for faces than for letters. For the single letters, performance remained high (above 90%) across all tested positions, but performance for the single faces declined substantially outside the central nine positions.
Our results demonstrate a surprising similarity in the visual spans for letters and faces. The pattern of results may be determined in large part by crowding between letters, and crowding within and between faces.
This PDF is available to Subscribers Only