Abstract
Purpose :
Stereopsis is the ability to extract depth information from retinal disparities. Although perception of depth and 3-D structure is possible with just monocular cues, the percept obtained using binocular disparity is faster and more vivid. Measuring the degree of stereoacuity from which a patient can extract useful information is a basic part of a visual exam. Several stereotesting booklets exist, varying in the disparities presented, method of inter-ocular separation, and principles of design. Theoretically, clinical measures of stereopsis should be method-independent and repeatable. The goal of this research study was to examine the inter- and intra- test repeatability of the traditional Randot (tRandot), Special Edition Randot (SERandot), Randot3 and TNO stereotests, the latter three of which are based on random dot stereograms and are devoid of monocular cues.
Methods :
Thirty normal participants (age: 26.93±6.98; 20 female) were tested using the four stereotests, three times, in random order and retested a week later. The means of the measurements were computed. Standard deviations of the means were used to determine precision. Repeated measures ANOVAs with post hoc testing were used to compare between the booklets and the first vs second session measurements. Bland and Altmann analysis evaluated the inter-session reliability.
Results :
Mean stereothreshold measurements were: 32.83±14.30, 24.83±9.08, 21.28±2.17, and 13.29±1.64 arcsec for the TNO, tRandot, SERandot, and Randot3 respectively. These were significantly different (F(df=1,3)=6.62, p=0.02) with post hoc indicating that all but the tRandot and SE Randot yielded different measurements. The mean inter-session differences were 0.94 ±4.58 [95% confidence intervals:9.93-(-8.04)], 13.33± 1.45 [16.17-10.48], 21.25± 2.00 [25.18-17.32], and 30.45± 12.71 [55.39-5.58] arcsec for the tRandot, Randot3, SERandot, and TNO booklets, respectively.
Conclusions :
Stereothresholds measured with varying stereotesting booklets yield different measurements. These could be due to the anaglyphic inter-ocular separation in the TNO potentially introducing an imbalanced view to the two eyes, a ceiling effect in the Randot3 measurements, and monocular cues in the tRandot. Based on this small sample, because the Randot3 test contains the lowest disparity, lowest standard deviation, and highest inter-session repeatability, it is the optimal testing booklet.
This is a 2020 ARVO Annual Meeting abstract.