Bionic Eyes May Become a Reality in the Near Future

Each year, 200 million people worldwide suffer from vision loss because of retinal degenerative diseases, and an estimated 255 billion dollars is spent in direct health care costs from these illnesses. Age-related macular degeneration (AMD) and retinitis pigmentosa (RP) are two that account for the leading cause of vision loss. Wet AMD causes leakages in the macula of the retina, severely impairing a patient’s ability to see with sharp vision. Further, RP destroys the rod cells in the photoreceptors of the eye, limiting the ability to see in dimly lit environments [2]. Moreover, the brain initially compensates for the dark spots in the patient’s vision, hindering early detection of the disease until it has reached its late stages. Currently, there are no effective treatments and cures to reverse the effects of RP or AMD. Therefore, doctors can only prescribe medication to reduce the swelling or recommend the use of sunglasses.

This is devastating for blind people, especially in areas with a lack of resources like developing countries. With the onset of the coronavirus pandemic and an estimated 50-70 percent increase in internet use, this problem will only be exacerbated. But how can we solve this problem?

Enter retinal prostheses. Retinal prostheses are bioelectric implants that provide patients with visual percepts through electrical stimulation. A typical epiretinal prosthetic contains three main external components: a video-camera, virtual processing unit (VPU), and an

external coil for transmission. These components work together to produce fast, real-time imaging of a patient’s surroundings. An image is first captured through a camera, which is then converted into electrical parameters conveying spatial- temporal information. This data is then sent to the chip through RF telemetry and then finally stimulates the retinal neurons through a microelectrode array. These devices have revolutionized this space, as the Argus II has become the world’s first FDA-approved technology capable of vision restoration. 

Although these retinal prostheses can successfully stimulate a patient’s retina, they have a major limiting factor. Because of the compact environment in which these prostheses need to be surgically placed in, the size constraint is severely limited. In addition, the heat released from the retinal prosthesis can damage the retina. Other factors include power consumption, limitations on micro fabrications, and incision size on the eyeball. The current image sizes of retinal prostheses range from 60 to over 1500 electrode arrays. Each electrode ranges from 260 um to 520 um in diameter. This is significantly smaller than the size of pictures people see , which reach up to around 65,000 pixels. Because of the limited resolution, patients will only be able to see a blurry, pixelated image of their surroundings. Further, recent clinical trials have shown that patients with the Argus II implant had a visual acuity of 20/1262, which is far below the limit for legal blindness. (20/200). Although recipients with such visual acuity had a significant improvement of activities of daily living and mobility by assessing a variety of daily visual tasks, such as letter and large geometric shapes recognition, word reading, object localization, and outdoor movement detection, it’s extremely difficult for them to perform more sophisticated visual tasks such as object and face recognition. Therefore, any image taken by the camera will have to be processed and downsampled by the VPU of the retinal prosthetic before being converted into electrical currents to ensure that the maximum spatial-temporal information is viewable by the patient.

However, recent research has been conducted attempting to solve this issue through improved computational frameworks that maximize the spatial and color information viewed by the patient. I have personally set out to combat this processing limitation by developing an architecture that includes a virtual magnifier, an optimization framework and a patient training library that can be implemented with virtual processing units of retinal prostheses.

Even though the Argus II technology is revolutionary and important towards our battle against blindness, the same cannot be said for its company, Second Sight. In 2020, the company went bankrupt, leaving hundreds of prosthesis users without answers or upgrades to their system. Even worse, if any complications arise, Second Sight would not be responsible in helping them, and removing the technology would be both expensive and painful. With that being said, I encourage, you, the future scientists and engineers in the world, to learn from Second Sight and not leave cutting-edge technology and its users in the dark.

Review Questions

1. What are the three components of retinal prostheses?
2. Given that retinal prostheses have limitations in the electrode array size and thus have a low image resolution, what are some ideas you have to improve the processing unit?

Answers

1. Retinal prostheses contain a camera, a virtual processing unit (VPU), and an external coil for transmission.
2. Ex.: A thermal image processing system to help patients see in the dark, a text-to-speech (TTS) algorithm that informs the user what objects they’re seeing, etc.

Image: Getty Images