Abstract
Purpose :
Eye movement studies with low vision readers using onscreen magnification have mostly focused on horizontal scrolling text (HST). Relatively little attention has been given to screen magnification, a technology that is available in any modern operating system. We present an analysis of eye movements measured for low vision subjects while reading short passages with a full screen magnifier.
Methods :
Gaze points on screen were recorded using a Tobii Fusion gaze tracker at 120 Hz from 18 subjects with low vision (age: 32 – 95; acuity in better eye: 0.24 – 1.04 logMAR; 9 with central vision loss). Subjects were asked to read short passages on a MacBook laptop using a full screen magnifier at their desired magnification level. We calculated the “compensated gaze point” — the location of the point in the document currently gazed at, defined in the reference frame of the unmagnified document (see figure). “Compensated fixations” thus include both actual fixations and smooth motions, which occur when gaze is pursuing a point in the moving magnified text (similar to HST).
Results :
Measured preferred angular print size (PAPS) ranged from 0.64 to 1.77 logMAR (mean: 1.03). Reading speed varied from 21 to 175 words/min (mean: 73). Within text lines, the average duration of compensated fixations for each subject ranged from 0.2 to 0.53 s (mean: 0.34), and the fixation rate varied from 1.59 to 3.93 fixations/sec (mean: 2.46). Forward compensated saccades accounted for 66% to 97% of total saccades (mean: 81%), and their length varied between 0.97 and 5.04 characters (mean: 3.14). No significant difference was found between the lengths of forward and backward saccades (paired t-test, p=0.71). Compensated forward saccade lengths were significantly negatively correlated with PAPS (r=-0.79; p<0.001). Stepwise linear regression shows that log reading speed is best explained by forward saccade length and fixation duration (p<0.0001).
Conclusions :
Operating a full screen magnifier is similar to using HST, with the important difference that users may vary the scrolling speed as desired, stop scrolling, and even temporarily scroll backward. When using a full screen magnifier, reading speed is predicted by fixation duration and saccade length, where the latter is negatively correlated with PAPS.
This abstract was presented at the 2024 ARVO Annual Meeting, held in Seattle, WA, May 5-9, 2024.