Date of Award

Spring 5-28-2015

Degree Type


Degree Name

Master of Science in Vision Science

Committee Chair

Yu-Chi Tai, PhD.

Second Advisor

John Hayes, PhD.

Third Advisor

James Kundart, O.D., M.Ed., F.A.A.O.


PURPOSE: To determine how various viewing conditions affect the accuracy and response time of the perception of stereopsis. Based on prior research findings, three specific hypotheses have been formulated relative to the main research question: (1) Stereopsis threshold (accuracy) and efficiency (response time in step-vergence) will be best for 3D-viewing conditions containing only local cues to fusion, intermediate for viewing conditions containing a combination of global and local cues, and worst for the viewing conditions containing only global cues. (2) Subjects with a lower (better) Titmus Stereoacuity threshold at 40cm will have faster response times and better accuracy on all three types of stereo tests than subjects with a worse stereoacuity. (3) Subjects with a lower Distance Randot stereoacuity threshold at 3m will have faster response times and better accuracy on all three types of stereo tests than subjects with a worse stereoacuity. METHODS: Ninety adult subjects (ages 19-43) with 120 arc seconds of near stereoacuity or better were recruited. Individual stereoscopic threshold and step-vergence efficiency were measured under three different testing conditions: Blank background Stereogram (BBS), Random Dot Stereogram (RDS), and Random Dot with Reference Frame Stereogram (RDRFS) for each subject. These tests were developed at the Vision Performance Institute and will hereafter by referred as the VPI Stereo Tests. The derived stereoacuity thresholds were also compared to the traditional clinical standard, the Titmus Stereo Test (for near viewing) and the Distance Randot Stereo Test (for far viewing). RESULTS: Mean stereoacuity threshold performance achieved was lowest (best) for the Titmus Stereo Test, intermediate for the VPI Stereo Tests (BBS < RDS < RDRFS), and highest (worst) for the Distance Randot Stereo Test. The majority of the subjects provided verbal feedback that RDRFS was easier to perform than RDS on both the stereoacuity and the step-vergence tasks. Step-vergence performance resulted in the fastest response time for BBS, intermediate response time for RDRFS, and the slowest response time for RDS. Response time was found to be faster when responding to visual stimuli exhibiting crossed disparity, and slower with uncrossed disparity. The majority of subjects responded fastest to the smallest amounts of disparity demand (20 pixels, 651 arc sec) and slowest to the largest amounts of disparity demand (80 pixels, 2604 arc sec). Threshold stereoacuity performance on both the Titmus Stereo Test and the Distance Randot Stereo Test was not predictive of accuracy or response time performance on any of the VPI Stereo Tests. CONCLUSIONS: Using the VPI Stereo Tests, the effects of local and global cues on stereopsis were compared directly. Threshold stereo acuity and vergence efficiency with local cues were better than with global cues. The distance measures of stereoacuity with VPI Stereo Tests are comparable with the near-viewing stereoacuity measured with the Titmus Stereo Test, which is considered the traditional clinical standard. The VPI Stereo Tests are more sensitive at assessing fusion ability than the Distance Randot Stereo Test, which is one distance-viewing stereo test currently on the market. Future studies should explore the comparison between the Titmus Stereo Test and the VPI Stereo Tests using manipulated viewing distance to lower the threshold scale via a population with normal binocular vision.

Included in

Optometry Commons