Eyeworld

EW NEWS & OPINION 24 September 2015 by Michelle Dalton EyeWorld Contributing Writer walked through the course with and without the device; collisions were reduced by about 37%, and none of the patients had more collisions when using the device. The reduction in collisions was "substantial" in both the HH and tunnel vision groups, Dr. Luo said. In the tunnel vision group (the retinitis pigmentosa and optic nerve damage patients), the reduction in collisions was 43%; Dr. Luo said there was an expected difference between the 2 groups and within the groups. "Collisions without the device ranged from 1 to 54 in the tunnel vision group and from 0 to 23 in the HH group," the researchers said, 1 noting multiple human factors also account for collisions, although visual field is a significant factor. The researchers said that al- though the reduction in collisions is far from 100%, patients achieved the effect with only minimal training. Dr. Luo thinks that the more people wear the device and gain experience with it, the greater the reduction would be. The researchers plan to evalu- ate the device in a rigorous clinical study, Dr. Luo said. The Department of Defense supported the device development and the first study, and Dr. Luo hopes the agency will support ongoing research as well. "The current prototype does not provide information about the direction of the predicted collision, but that will be addressed in our next study," he said, where he hopes to have recruited patients use the de- vice in real-world environments. EW Reference 1. Pundlik S, Tomasi M, Luo G. Evaluation of a portable collision warning device for patients with peripheral vision loss in an obstacle course. Invest Ophthalmol Vis Sci. 2015 Apr 1;56(4):2571–9. Editors' note: Dr. Luo has financial interests through a pending patent related to the collision detection method. Contact information Luo: gang_luo@meei.harvard.edu enrollment). Twelve of the remain- ing 13 patients had retinitis pigmen- tosa that resulted in tunnel vision, and the final patient had optic nerve damage. Based on time-to-collision computation, Dr. Luo's device automatically accounts for walking speed—meaning the device is able to detect potential collisions regardless of the speed of the wearer. "If you're walking slowly, it may take up to 2 seconds to make contact with an object 1 meter away. If you walk fast the device needs to warn you much farther out and still give you 2 seconds to avoid collision." Proximity-based systems cannot differentiate when contact will occur as easily as time-to-collision devices can. Study findings In the initial study, the obstacle course was a loop about 41 me- ters long; there were 46 stationary objects at different heights serving as obstacles to avoid. These included decoration "trees," hanging obsta- cles, and floor level obstacles. The device is worn chest-high (about 1.4 meters) and "is able to cover a range of 0.75 to 2 meters from 1.2 meters away," Dr. Luo explained. Subjects warning, you're going to bump into something, so it's very straightfor- ward," Dr. Luo said. "You don't need to interpret distance, direction, or anything, just listen." The device is pocket-sized and equipped with a video camera. Wearable warnings The device gives warnings only when users approach obstacles, not when users stand close to objects and not when moving objects just pass by, Dr. Luo said. The auditory collision warnings given by the device "are simple and intuitive- ly understandable. We tested the device in a density obstacle course to evaluate its effect on collision avoidance in people with peripheral vision loss. To show its beneficial effect, we compared the patients' mobility performance with the de- vice and without it. Just demonstrat- ing the device can give warnings for obstacles while walking would not prove the device is useful. We have to compare with a baseline, which is walking without the device in this case." The group evaluated 25 people with significant peripheral vision loss (12 patients with HH, some as long as 25 years before study Researchers at Schepens have developed a wearable device that may help patients with advanced peripheral vision loss regain some mobility P eople with severe periph- eral vision loss (as from retinitis pigmentosa, homonymous hemianopia [HH], or glaucoma) have a much higher probability of colli- sion than people without such loss, which can lead to increased falls and accidents. Working similarly to car back-up sensors, previous assistive technologies have trended toward detecting obstacles based on prox- imity to objects rather than a time to collision, and these technologies usually have a high false alarm rate. Researchers in Boston think they may have a viable solution that can reduce the number of collisions (and, therefore, accidents) in this patient population. "If you are standing in front of a person talking to him, the proximity devices can detect existence of the person. When you're using these devices, the user has to interpret where the obstacle is—whether it's directly in front or off to the side— and whether the location of the obstacle would prevent forward mo- tion," said Gang Luo, PhD, assistant professor, Harvard Medical School, and associate scientist, Schepens Eye Research Institute, Massachusetts Eye and Ear, Boston. He and his colleagues used an obstacle course to evaluate a wear- able collision-warning device they developed based on time-to-colli- sion detection. He said the study showed that the device may help patients with a wide range of vision loss avoid collisions with high-level obstacles, 1 even people with a rela- tively large residual visual field, such as those with hemianopia. Residual vision is "very valuable" for patients, and they can spot some obstacles by themselves, therefore avoiding some collisions. However, unforeseen collisions still occur at a high rate. That is where assistive technologies should be able to help, he said. "Our device provides an audible warning. Every time you hear that Avoiding collisions for vision-impaired patients Patient uses the wearable device; the transmitter is in his shirt pocket. Source: Gang Luo, PhD

• Cover