Abstract
Purpose :
To simulate visual impairment across the macula, we masked sentences with patterned noise. We previously observed that reading speed and error rate are improved by oscillating the sentences in noise, and now characterize reading as a function of noise contrast and age group. We compared older subjects to patients with macular disease.
Methods :
We tested a broad range of subjects with normal contrast sensitivity (Pelli-Robson) no macular scotoma. Control subjects (27 - 70 yr, 47.1 + 16.6 yr, 5 females and 5 males, and spherical error 3.5 to -7.5 D) were divided into groups of 5 younger (< 50 yr) and 5 older ( > 50 yr). We compared the older group to 2 patients with macular pathology, one with partial macular hole and the other with early age-related macular degeneration but no focal defect. Using a digital light ophthalmoscope (DLO), we presented sentences via Maxwellian view in a 28 x 18 deg field. We determined retinal locus by simultaneously capturing coplanar retinal images at 860 nm in a 28 deg field. Subjects read each IU Read sentence separately, which were displayed with a 1920 x 1080 pixel resolution, at 20/60 Helvetica font. The text contrast (delta I/I) was 50%. The added noise had a center spatial frequency of 5.8 c/deg, and Michelson contrast 0 - 40%. Each 10 – 14 word sentence was stationary or moved horizontally in a sinusoidal manner, with amplitude from 0 - 2 deg and velocity from 0 - 2 Hz.
Results :
When no motion was present, reading rate declined as a sigmoidal function of noise contrast (p = .002). The improvement of reading with text motion depended on noise contrast. With no text motion, younger subjects read faster than older ones (p=.039), improving with motion when contrast was .3 by 19 and 29 words per minute. Reading rate at 0 noise contrast was 171 +/- 37.2 words per minute for younger subjects vs. 143 +/- 15.5 for older ones. The patients required 100% contrast letters even with 0 contrast noise, missing .168, .025 of the words, respectively ( .00067 errors for controls). With motion, the patients’ errors reduced to .060 and .0085 words, vs. .0081 for controls, with slower reading rates, 47 and 88 vs. 135 +/ -8.5 words per minute. In noise subjects scanned inappropriately right to left.
Conclusions :
Improvement by controls in reading sentences in noise, and by patients with visual impairment, can occur with text motion.
This abstract was presented at the 2019 ARVO Annual Meeting, held in Vancouver, Canada, April 28 - May 2, 2019.