ELVIS: Sight Restoration with Artificial Intelligence

We had the privilege of interviewing Denis Kuleshov, Director of Sensor-Tech, a Russian-based company that is working on a cortical visual prosthesis that hopes to restore vision to people who are blind. Called ELVIS (short for “electronic vision”), the device was first showcased at Moscow’s Skolkovo Innovation Center in June 2021. Looking like something out of Cyberpunk 2077, the device is worn on the head and designed to bypass any damage in the retina or optic nerve that blocks the signals from being transmitted to the vision center in the brain. We asked Dr. Kuleshov about the company's philosophies and future plans in the visual neuroprostheses field.

Dr. Kuleshov, please give us an overview of the ELVIS device and what sets it apart from competing products.

[DK] As with any other cortical visual prosthesis, we are trying to transfer the image from the camera directly to the brain. The ELVIS prototype has three parts: a band, a microcomputer, and an implant. The potential blind user will have to wear the band on the head, with a stylish red look, and the microcomputer on their belt. There is an implantable part that contains an antenna which receives energy from the wireless band, and within the implantable device is a box which controls and contains the electrodes. Our demo has a 10x10 array, and we are currently testing a different configuration on rats. Basically, the whole system is about 80% ready. The difference between ELVIS and other cortical implants is not because of the technology, but because of the patient experience. The technologies will be nearly the same on the surface, but with this approach the patient will have a better experience with limited vision, which is our philosophy.

What are the advantages of your approach compared to, say, Argus II?

[DK] Well, in 2017 we had the experience of implanting patients with the Argus II prosthesis in Russia in collaboration with Second Sight. It was difficult to find patients because they have to have late-stage retinitis pigmentosa to be implanted. After a few years, we discovered that Second Sight was going to stop production of their Argus II retinal implant. It was around the same time when Retina Implant AG, manufacturer of the Alpha AMS and IMS subretinal prostheses, also stopped production, due to lack of investments. So with the understanding that we wanted to make something local to Russians and that retinal implants were limited to clinical cases with specific retinal degeneration, we decided that a cortical implant (as difficult as it may be to produce) would be the way to go if we wanted to help masses of people.

The ELVIS device seems to rely on artificial intelligence (AI) for image processing, especially object recognition. Could you elaborate on the use of AI algorithms in the device?

[DK] We already have a product in Russia that is a smart camera for blind people, which can recognize and measure distance to objects, can tell you what's going on in front of you, etc., so we are familiar with AI. We have another AI feature in testing now where with a user's push of a button, the device recognizes objects and gives auditory feedback. For example, it can tell you when there is a person, chair, bottle, etc. near you. So, we are trying to combine bionic vision and AI in one system. We also have algorithms to detect currency builds, read text, and all that. Many of those functions will be available like they are on Tesla cars, where you install a special app and you have a new function available to you.

So it’s like installing an app on your device, except on your brain implant.

[DK] Yes, something like that. You can see that our algorithms right now are using AI to detect the contours or edges of the objects. To transmit the signals, not only black and white shapes, but directly translate the lines of the objects will definitely be better for the patients if they want to understand what's going on in front of them, if they want to walk along the road, if they want to find the door, etc. All this AI works inside the device without the need for internet connection. It’s all in real time.

What plans does Sensor-Tech have for the future?

[DK] We are currently polishing the electronics, algorithms, and AI of ELVIS, but the implantable parts still need to be researched and tested accurately. We are testing the system with rats now, and plan to start testing with monkeys soon. Hopefully in three or four years the Russian government will let us test it on individuals with blindness. We will need to test this with people for a minimum of 2 years. We hope to certify the ELVIS cortical implant by the year of 2027 in Russia. After that we will probably start to think about going to Europe, Asia, and possibly the US.

Where do you see the field of visual neuroprosthetics going in the next 5-10 years?

[DK] Well, I think that we know technological and medical breakthroughs are possible. We also know that these are safe procedures due to the well-trained neurosurgeons and ophthalmologists who perform them. What makes the field complicated is the marketing aspect. We are a non-profit company and supported by the Deaf-Blind foundation and other sponsors because they believe that we can help a substantial number of patients. When we talk to other investors though, they ask us why we are not making something like Neuralink, which is cost-affordable and eligible for a large number of people to obtain. Well, obviously it’s more complicated than that, but we need to face the marketing reality that we are at a disadvantage. With a cost of around 140,000$, there will be some people unable to purchase a cortical implant. So then the government insurance agencies and other stakeholders need to adjust the business model effectively, which creates complications. Additionally, once the product is approved in one country, there are more hoops to go through to get the product approved in other countries, which makes investors wary. Basically, the price is high, the time of production is long, and that is a drawback for investors. We are still searching for answers to this marketing dilemma because there is no point in stopping. We have to keep pushing forward.