Another Look at HMD Safety
by Tom Piantanida, Ph.D.

This article originally apeared in the November/December 1993 issue of CyberEdge Journal
© 1993 CyberEdge Journal

Recent reports in the popular press have raised public concern about the safety of low-cost home virtual-reality (VR) systems. The major safety issue focuses on whether head-mounted displays (HMDs), of the sort that will proliferate in home VR systems, can permanently injure the eyes. That's a tough question that has no simple answer. To answer in the affirmative is to imply that HMDs are inherently dangerous, which is not true; to answer in the negative ignores the fact that such innocuous activities as sustained reading can permanently injure the eyes of a small number of susceptible readers.

It is difficult to assess the safety of HMDs partially because of the number of factors that contribute to safety, and partially because HMD safety is poorly defined. For example, does "eye strain" constitute an unsafe condition? What about visual stress?

To circumvent some of the more difficult problems in understanding the safety issues, the remainder of this article will concentrate on one aspect of HMDs that can affect the visual function of most viewers, and can permanently affect a few. That aspect is the distance between the viewer's eyes relative to the distance between the lenses in the HMD. More specifically, it is the difference between the viewer's interpupillary distance (IPD) and the optical center distance of the HMD.

Interpulpilary distance

Viewers come in various sizes and shapes. Interpupillary distance, the distance between viewers' eyes, can vary over a considerable range. For example, some viewers have eyes that are set 56 mm apart, others have eyes that are set 66 mm apart, and many are in between. To adjust for IPD, a few HMDs have moveable lenses that can be positioned so that the optical centers are aligned with the viewer's pupils; most do not. Those HMDs that do not have adjustable IPD have the potential for introducing optical distortions to which the viewer must adapt. Likewise, HMDs with adjustments for IPD may introduce optical distortions if they are improperly adjusted.

In theory, the pair of lenses in the HMD should not need to be adjusted for each viewer. A perfect lens system would provide a collimated bundle of rays to the eye, so it would not matter where the eye intercepted that bundle of rays. Thus, with a perfect lens system, it would be pointless to make adjustments for IPD. Unfortunately, all lens systems are imperfect; some distort squares seen through them into pincushion shapes, and others into barrel shapes. These distortions are evidence that the bundle of rays reaching the eyes is not everywhere collimated, so it does make a difference where the eye is located with respect to the center of the bundle of rays, and hence, to the lens.

The lens systems of many common HMDs introduce pincushion distortions. Because the distortion becomes progressively greater the further the eye is from the optical center of the lens system, the best way to minimize this distortion is to position the viewer's eye at the optical center of the lens system. As most viewers resist having their IPD changed, minimizing optical distortion will require repositioning the HMD optics.

From a viewer's perspective, things seen through the optical center of the lens system are undeviated, but things seen off center are displaced as though seen through a prism. If the fixed lenses of the HMD have optical centers that are separated by more than the viewer's IPD, then the viewer will experience an outward deviation of the left and right eye images when looking through the lens system. In essence, for the viewer to look straight ahead through the displaced optics of the HMD, both eyes will have to be rotated inward, that is, to converged.

Similarly, for any viewing distance, the viewer will have to converge his eyes more in this HMD than he would have if he were viewing the same scene outside the HMD. Initially, this overconvergence for a given fixation distance is uncomfortable, but the visual system is very adaptive, and after only a few minutes, it recalibrates the extraocular muscles of the eyes so that vergence movements compensate for the optical distortions of the HMD. Thus, when fully adapted to an HMD with fixed optics that are separated by more than his IPD, a subject would become slightly exophoric, or cross-eyed.

If we were to examine the changes in binocular functioning that occur in persons who wore such an HMD, we would note that before donning the HMD, most of our subjects were either orthophoric-the eyes point directly at fused binocular targets-or slightly exophoric-the eyes point slightly outward. (it is not uncommon to measure a slight exophoria in many normal subjects.) If we again measured the binocular functioning of persons who had worn such an HMD for only a few minutes, we would find that many of the subjects show an esophoria, that is, a slightly cross- eyed condition when viewing fused binocular targets. Repeated measures of binocular functioning would be expected to reveal that esophoria develops fairly rapidly within the subjects, perhaps over a period of a few minutes, and then stabilizes as the vergence system fully compensates for the prismatic deviation introduced by the HMD. Thus, there would be very little change in the degree of esophoria in subjects who had worm the HMD for periods ranging from about ten minutes to several hours.

Upon removing the HMD, subjects would again experience visual discomfort because their vergence system would be adapted to the HMD optics, rather than to the unaided view of the world. Our fixation disparity measures would confirm that most subjects were still mildly esophoric – slightly cross-eyed – immediately after they removed the HMD. Fortunately, virtually all of the subjects would readapt to the real- world viewing conditions over a period of a few minutes, so follow-up measurements would reveal a return to the baseline orthophoria or exophoria within minutes of removing the HMD.

Possible long-term effect

There is, however, a small number of persons who may experience more long-term effects of wearing such an HMD. An example of a population who may be at risk in HMDs is persons with a condition called intermittent exophorpia. Persons with this condition usually maintain good ocular alignment, that is, they usually have good binocular fusion, but sometimes one of their eyes will deviate so that they loose binocular fusion and, therefore, see double images. Among the things that can trigger this deviation, also called a squint, is fusional stress.

A very common source of fusional stress is sustained close work, such as reading. For this reason, it is not uncommon to see the first episodes of exotropia (squint) among high school and college-age students. Unfortunately, HMDs of the type described above induce fusional stress in the wearers. Thus, there is some potential that the HMD may trigger an episode-perhaps even the first episode-of double vision among persons with intermittent exotropia. The important distinction to make here is that the HMD does not cause the condition, but merely triggers it, just as fatigue might. Once triggered, however, repeated episodes of double vision can be expected.

The visual system attempts to overcome the inconvenience of double vision by suppressing one of the images. This suppression, if it occurs in very young children and if it is sustained, can lead to permanent visual changes of the type commonly called amblyopia or "lazy eye." Thus, in a small number of very young children, HMDs of the type described above have the potential for triggering latent visual anomalies that can produce permanent visual changes.

In the final analysis, for the type of HMD described above, there is reason for limited concern about long-term effects on young children and on persons with conditions like intermittent exotropia. This should not be construed as a condemnation of HMDs or even of one type of HMD. Rather, it is a cautionary note that should enable us to better balance the anticipated benefits of VR systems with the potential dangers of challenging the visual system with novel display systems.

Tom Piantanida, Ph.D., was the Founder and Principal Scientist of the Virtual Perception Program at SRI International, 333 Ravenswood Avenue, Menlo Park, CA 94025 USA. He can be reached at TPP6VISION at aol.com

Questions? More Info? Email the Webmaster

407 M. L. King Jr. Way, Oakland, CA 94607 USA
+1 510 419-0800

Our privacy policy may be read here.