Abstract
Purpose :
To assess to what extent touch- and gaze-contingent feedback improve reading speed using a low-vision eye movement training program.
Methods :
Subjects with central scotomas as a result of retinal disease were recruited for this pilot study. All participants underwent training to improve reading eye movements using the protocol of Seiple et al., 2011. Training was done by requiring subjects to move their gaze between sequentially presented visual targets. In the current study, two feedback conditions were tested in a randomized, cross-over design. In Condition 1, using a touch screen, subjects were required to point to and touch each training target as it appeared. In Condition 2, eye position was tracked using a Tobii 4C head/eye tracker. A gaze-contingent dot was displayed during training. Subjects were asked to move their eyes and subsequently, the gaze-contingent dot to each new target. Subjects were trained for one hour per week for four weeks in each condition, with the order of condition counterbalanced. Reading performance, gaze path, fixation stability (BCEA), and reaction time were measured pre- and post-training.
Results :
Six females and four males, with an average age of 70.1 years, participated. BCEA decreased from 2.51 (±1.2) degrees square pre-training to 1.60 (±1.1) degrees square post-training (P=0.002). The time between appearance of a training target and fixation on that target decreased from 1.40 (±0.4) seconds pre-training to 1.17 (±0.4) seconds post-training (P=0.031). The efficiency of the scan path also improved from a mean deviation of 1.46 (±0.4) degrees at baseline to 1.05 (±0.5) degrees post-training (P=0.006). Average reading speed to IREST sentences increased by 35.8% in the touch feedback condition and by 101% in the gaze-contingent feedback condition.
Conclusions :
Touch feedback increased reading speed by an average of 36%. This value is similar to previously reported increases in reading speed without feedback (Seiple et al., 2005 - 28%; 2011 - 46%). In contrast, gaze-contingent-feedback training produced a much larger increase in reading speed, as compared to our previous findings, and therefore, should be futher tested in low-vison rehabilitation settings.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.