Purchase this article with an account.
A.J. Kelley, L. Yang, G. Dagnelie; The effects of stabilization, font scaling and practice on reading in simulated prosthetic vision . Invest. Ophthalmol. Vis. Sci. 2004;45(13):5436.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Purpose:To investigate possible effects of image stabilization provided by a visual prosthesis: Visual information presented through stationary retinal or cortical electrode arrays will not respond appropriately to eye movements, or to the movement of a head–mounted camera. Methods:Normally sighted volunteer subjects were included in this study. They wore a video headset with built–in pupil–tracking capability, and viewed pixelized text segments through a 10x10 to 25x25 dot raster with varying dot parameters (18–77 arcmin), font size (4.5–9.7 dots/char), and contrast (12.5% and 100%). They were instructed to scan the dot raster across the text using the computer mouse, and to read aloud, word by word. Each subject was tested in both free–viewing and fixation–controlled (stabilized) conditions. Accuracy (% of words read correctly) and reading speed (words/min read correctly) were recorded for 384 trials of 7–10 words each. Results:Reading speeds of up to 40 words/min and >99% accuracy were recorded in the free–viewing condition with 8 dot/char font scale at high contrast. Reading speed was affected by font scale (60–80% drop at 4.5 dot/char) and contrast (15–25% drop), while accuracy remained over 95% in most conditions, falling to 80% (high contrast) or 70% (low contrast) for the 4.5 dots/char font only. Retinal stabilization reduced reading speeds by 50–75%, but accuracy remained above 60% (4.5 dots/char) and 90% (other font sizes), even at low contrast. With practice, reading speeds (averaged across blocks of 22 trials at all stimulus conditions) more than doubled, for both stabilized and free–viewing conditions. Practice also resulted in increased accuracy, especially for the stabilized (from 40% to 70%) and low contrast (from 80% to 90%) conditions. Conclusions:These results have important implications for retinal and cortical prostheses currently being developed: 1. Optimizing stimulus magnification will be crucial to the feasibility of reading with a prosthesis, at least until electrode grids with dot resolution over 25x25 can be implanted. 2. Image control by the wearer’s eye movements should be an important goal in prosthesis development. 3. Until eye movement control becomes feasible, training in the avoidance of rapid eye, head, and camera movements will be an important part of a prosthetic rehabilitation program.
This PDF is available to Subscribers Only