Careful screening of the near visual functions monocular acuity, stereopsis, suppression, and color are an important component of a comprehensive eye examination. Testing these visual skills is particularly challenging in children and patients who are averse to monocular occlusion1 or wearing of goggles or head-pieces to convey depth perception.2
An autostereoscopic parallax barrier screen has a layer of liquid crystal display pixels behind a second layer with parallel, vertical “windows and shields” so the right and left eye perceive a different set of pixels.3 Practically, this unique type of screen functions so it can test acuity monocularly without occluders and depth perception without the need for goggles.4 PDI Check (PDI Check LLC, Anchorage, AK) is a vision assessment game developed for the autostereoscopic upper screen on the Nintendo 3DS XL (Nintendo Co., Kyoto, Japan) game console.5 PDI Check version 0.2.5 emphasizes quick assessment of monocular acuity, stereopsis, and red-green color deficiency using a dynamic, forced multiple-choice format with the patient making selections on the Nintendo 3DS XL lower touch screen. This study was conducted to confirm clinical validity and comparative timing.
Patients and Methods
This research was approved by a prospective institutional review board from Pacific University (reference number 109-17, Project 10098158-2) and was compliant with the Health Insurance Portability and Accountability Act of 1996 and the tenets of the Declaration of Helsinki. Written consent was obtained from a responsible adult. De-identified data for this study can be accessed from http://www.abcd-vision.org/references/PDI%20Check.html.
Conventional testing included patched, monocular near visual acuity with a Rosenbaum pocket card held at 40 cm, Ishihara Concise color plates binocularly, Worth 4-dot, and Titmus Fly stereo with polarized spectacles. The time to perform all of these tests was monitored.
Sample size for a Spearman correlation was determined using power analysis. Given an alpha value of 0.05, a power of 0.80, and a medium effect size (ρ = 0.33) for a two-tailed test, the sample size was determined to be 70. Data were compared using the Mann–Whitney U test. Each participant played the PDI Check game on Nintendo 3DS XL development kits using version 0.2.5.
The game is viewed from 30 cm. It has an orientation phase for familiarization and determination of a minimum lag time from initial observation of a coarse image change on the upper three-dimensional color screen to correct pressing (with stylus or finger) on the lower touch screen. For the orientation phase, an abrupt, coarse presentation of binocular visual acuity (Landholt C rotates 90° to “U”), stereopsis (rings with crossed binocular separation to resemble 800 seconds of arc), and color (one of four quadrant identical hue discs commencing at a green axis of confusion on the Farsworth-Munsell D100 test rapid shifts 10 FM D100 units) is used so even patients with limited vision should be able to visualize.
The participant selects acuity, stereopsis, and color from the home menu to initiate a gradual presentation from fine resolution to coarse resolution scored when the participant selects the corresponding quadrant on the lower touch screen. For color, one randomly selected quadrant colored disc gradually shifts in hue away from the common axis of green cone confusion. For stereopsis, one randomly selected quadrant has gradual crossed separation of two rings such that a binocular person perceives ring levitation up to 800 seconds of arc. For acuity, all four quadrants commence with small Landholt C oriented as C in each eye. In one randomly selected quadrant, the right eye Landholt C orients to “U” while the non-tested left eye remains a rival “C.” Then all optotypes gradually increase in size until the participant scores by touching the corresponding lower screen touch screen quadrant. Because the acuity optotypes are presented with rivalry, ocular suppression can be estimated. Actual performance of version 0.2.5 on the PDI Check game is shown in https://vimeo.com/290176594.
Output from the PDI Check game includes clinical visual metrics derived from previous trials and appropriate regression formulas with acuity in 20/20 near format, stereopsis in seconds of arc, and color as percent green cone (deuteranopia) efficiency. The game also gives times for performance of the orientation phase and overall near vision testing of acuity plus stereopsis plus color. The participants were queried at the end of testing as to whether they preferred the game, conventional books, or both.
PDI Check selection times were compared to conventional testing; linear regression was used when appropriate. Poor stereopsis vision was defined as less than 200 seconds of arc on the Titmus Fly test. Poor color vision was defined as missing five or more plates on the Ishihara test. Ocular suppression was determined from Worth 4-dot testing. Validation testing was performed with appropriate instrument referral criteria preselected for PDI Check near visual parameters (Table 1).
Of 77 children and adults (aged 5 to 62 years, median: 11 years), 12 had suppression. PDI Check (x) estimated Titmus (y) by y = 0.61x−8, R2 = 0.64 and screened poor stereopsis (< 200 sec arc) with 90% sensitivity and 85% specificity. PDI Check green cone (x) estimated Ishihara plates missed by y = 0.01x2−0.38+4, R2 = 0.52 with a value less than 60% yielding 80% sensitivity and 99% specificity for deuteranopia. When PDI acuity in one eye lagged chart acuity by greater than 0.8 logarithm of the minimum angle of resolution, deep binocular suppression was screened with 67% sensitivity and 96% specificity. Monocular PDI acuity screened poor Rosenbaum near acuity (< 20/63) and suppression with 67% sensitivity and 91% specificity for the right eye and 55% sensitivity and 94% specificity for the left eye. Figure 1 shows the median time to completely screen with the PDI Check of 84 seconds (90% confidence interval [CI]: 42 to 173) was significantly less than for conventional screening (202 seconds; 95% CI: 43 to 361; Mann–Whitney z = 6.6; P < .01). For abnormal patients (n = 22), the PDI Check was quicker (median: 136 seconds, 95% CI: 55 to 195) than conventional screening (median: 234 seconds; 95% CI: 171 to 361; Mann–Whitney z = 29; P < .01). Normal patients (n = 55) were also quicker with PDI Check (median: 77 seconds; 95% CI: 42 to 157) than with conventional screening (median: 186 seconds; 95% CI: 43 to 360; Mann–Whitney z = 4.9; P < .01). Normal patients were quicker than abnormal patients for conventional screening (Mann–Whitney z = 2.3; P = .02) and especially for the PDI Check (Mann–Whitney z = 3.7; P < .05). A total of 87% preferred the game over conventional screening.
Box and whisker plot showing the time to screen near visual modalities stereopsis, color, suppression, and monocular visual acuity using conventional static tests and the Quick Version of PDI Check dynamic game (PDI Check LLC, Anchorage, AK) for the Nintendo 3DS XL platform (Nintendo Co., Kyoto, Japan). The patients are separated by their conventional test results as normal (n = 22) and abnormal (n = 55).
The PDI Check Quick Screening program was able to screen for poor near visual acuity, suppression, stereopsis, and red-green color deficiency 2 minutes quicker than conventional tests. Conventional monocular acuity testing of distance is preferred over near visual acuity, especially in children, for two reasons: myopia can be missed with near testing and children with subnormal acuity but ample accommodation can hold a near card so close as to falsely pass. Because PDI Check is on the autostereoscopic screen presenting a rivalry form of acuity, we have not yet fully determined its ability to screen for amblyopia.
A poor score on monocular acuity of PDI Check could imply poor conventional chart acuity, ocular suppression, or both. The ocular rivalry presentation of the Landholt C optotypes is not the only paradigm for presentation on the autostereoscopic screen; further developments include simple monocular presentation, which may be simpler to pass for some patients with ocular suppression. Stereopsis and deuteranopia were efficiently screened. The high levels of specificity for color, suppression, and monocular acuity indicate that the Quick Screening quickly identifies normal patients as normal, whereas abnormal patients might be subjected to repeated or more enhanced screen testing in future versions of the game. By modifying the presentation of monocular and binocular cues through the autostereoscopic barrier screen, different challenges to suppression and stereopsis may be repeatedly tested in the future. The PDI Check is also developing more refined games for color and stereopsis.
This study is an ongoing calibration of this relatively new near visual assessment game from which consistent correlations and dichotomous screening were confirmed. Mathematical formulas that produce clinically relevant values from game performance times of more participants are still evolving, but Bland–Altman plots were analyzed. There was better function for stereopsis (mean ± standard deviation: 172 ± 327 seconds of arc) and less for near acuity (mean ± standard deviation: 0.11 ± 0.37 logarithm of the minimum angle of resolution) and color (mean ± standard deviation: 63% ± 9% green cone deficient).
Visual acuity screening, especially in children, is recommended as a distance chart with the untested eye patched and approved optotypes crowded or surrounded to select for amblyopia. Current guidelines propose age-based referral cut-offs of 20/50 for age 3 years, 20/40 for pre-kindergarten, and 20/32 for older children.6 The PDI Check is a near test with a unique method of ensuring monocular testing. Instead of a 20/40 cut-off for near, we instead aimed for the low-vision classification of 20/70, but because that is not logarithm of the minimum angle of resolution, we instead chose the next smallest size of 20/63. Because the visual acuity test on the PDI Check is not directly comparable to static chart optotype identification/matching, but rather a dynamic, enlarging forced-choice process, we are not sure whether the PDI Check near task is better or worse in comparison with distance vision; further testing is underway.
Limitations of this study included a relatively small sample size of normal participants despite an ample number of patients with prior strabismus or ocular suppression. Only six patients with color deficiency were included. The same technician who presented and timed the conventional tests also presented the PDI Check game. The screen presentation-to-clinical correlation functions for this and former PDI Check version calibrations have been derived from just over 300 clinical cases,7 so further testing with high pre-screening probability cohorts is needed to better refine clinical data ranges. Further work with this near vision testing format also needs to be done in more patients with different age ranges with nystagmus, patients with aversion to patching or goggles, and those with developmental delays such as autism spectrum disorder.
The dynamic Quick Screening version of the PDI Check video game correlates with conventional static clinical tests for monocular visual acuity, suppression, stereopsis, and color, but it is substantially quicker.
- Coulter RA, Bade A, Tea Y, et al. Eye examination testability in children with autism and in typical peers. Optom Vis Sci. 2015;92:31–43. doi:10.1097/OPX.0000000000000442 [CrossRef]
- Bhaskaran S, Lawrence L, Flora J, Perumalsamy V. Functional and cognitive vision assessment in children with autism spectrum disorder. J AAPOS. 2018;22:304–308. doi:10.1016/j.jaapos.2018.03.010 [CrossRef]
- Kim SK, Yoon KH, Yoon SK, Ju H. Parallax barrier engineering for image quality improvement in autostereoscopic 3D display. Opt Express. 2015;23:13230–13244. doi:10.1364/OE.23.013230 [CrossRef]
- Kavsek M, Braun SK. Binocular vision in infancy: responsiveness to uncrossed horizontal disparity. Infant Behav Dev. 2016;44:219–226. doi:10.1016/j.infbeh.2016.07.001 [CrossRef]
- Arnold RW, Damarjian AG, Smith KA. Sensory testing and stereopsis with Nintendo 3DS game. J AAPOS. 2018;22:e16. doi:10.1016/j.jaapos.2018.07.053 [CrossRef]
- Donahue SP, Baker CN, Committee on Practice of Ambulatory MedicineAmerican Academy of Pediatrics et al. Procedures for the evaluation of the visual system by pediatricians. Pediatrics. 2016;137(1). doi:. doi:10.1542/peds.2015-3597 [CrossRef]
- Arnold RW, Damarjian AG, Molina A, Smith KA. Calibrated measurement of acuity, color and stereopsis on a Nintendo 3DS game console. Clin Optom. 2019;11:47–55. doi:10.2147/OPTO.S199992 [CrossRef]
|Visual Function||PDI Check Criteria||A||B||C||D||i||Sensitivity||Specificity|
|Near acuity R||logMAR > 0.5||6||6||3||61||1||67%||91%|
|Near acuity L||logMAR > 0.5||6||4||5||61||1||55%||94%|
|Color||green < 60%||4||1||1||67||4||80%||99%|
|Stereopsis||> 1,020 arc sec||13||3||2||58||1||87%||95%|
|Suppression||ΔlogMAR ≥ .8||7||1||5||55||9||58%||98%|
|exam +||exam −|