Every year some 15 million people around the world suffer from stroke after which some five million become disabled for the rest of their lives. iBrain developed a neural interface and computer games which, according to the startup founders, will help an immobilized person after a stroke or a brain injury to train in order to be able to control a game app, a prosthetic limb, a manipulator or a robot using signals from brain. This will only be the first stage in recovering motor functions. Invest Foresight interviewed co-owner and CEO Konstantin Sonkin about the project.
From scientist to businessman
Konstantin Sonkin is a cyberneticist who graduated from the Polytechnical University in St. Petersburg. For eight years, he worked on artificial intelligence and wrote a dissertation.
“If artificial intelligence can recognize speech and images why can’t it recognize signals from natural intelligence?” he contemplated.
He started looking for ways AI could help a human brain. This is how the idea for iBrain came about. The company was registered in 2017 in St. Petersburg. The project team spent eight years developing AI for recovering brain segments that are responsible for motor functions. In 2018, they started selling the technology and earned several hundred thousand rubles.
Looking for smart investor
The iBrain neural interface is currently used at several large research and medical facilities, as well as the Medsi network of clinics, the 7 Doctors rehabilitation center, the Sechenov Institute of Evolutionary Physiology and Biochemistry of the Russian Academy of Sciences, a Spanish clinic, and others. Plans call for employing the device at French and Israeli clinics as well. Sonkin believes that researchers should not bother selling their developments, so he created his own sales channel. In order to increase sales, the company needs distributors and, therefore, money.
Patients pay some RUR 2-3.5K ($30-53) for a one-hour treatment session with the neural interface. Some patients need 20 sessions plus massage, physical therapy and medicines to fully recover; others need about 2-3 months. For some fifteen percent of patients rehabilitation does not work. At the moment, about 200 million people are immobilized because of a stroke or head injury. The iBrain neural interface improves muscle movement twofold.
The iBrain license will cost RUR 1 mio ($15K). The equipment – an EEG cap and a computer – will cost RUR 500K. The EEG cap is made by the Russian company Mitsar, which sells is equipment in 60 countries and is assisting iBrain in developing its business. On average, a medium-sized rehabilitation center can service six patients per day 20 days a month. The annual revenue will be some RUR 2.9 mio given that the cost of one session is RUR 2K. The volume of investments has not been revealed.
“First of all, we need not money but a smart investor with who we will change the lives of millions of people. We need a person who will help us enter the market with its numerous obstacles and excessive regulations,” Sonkin explained.
At the moment, iBrain is listed in the top-30 startups chosen by Russia’s Sberbank to work at the 500 Startups seed accelerator in the Silicon Valley. iBrain hopes to find investors there to make the distribution of their product easier. Plans call for creating a cloud access to the iBrain neural interface.
EEG cap and the power of thought
The existing rehabilitation methods such as massage and physiotherapy do not solve the main problem, which is restoring the brain function. iBrain is trying to turn the situation around using a neural interface and computer games that allow a patient to interact with the environment by the power of thought. How does it work? A special EEG cap reads the patient’s electric brain activity. The wireless cap does not manipulate with the patient’s brain in any way. Hands will appear on the screen and the patient is asked to imagine moving them.
The brain signals are deciphered and reflected on the screen. iBrain software allows detecting movements based on brain signals and recognize what the patient wants to do. For post-stroke patients it is very important to train damaged brain areas. It is not the muscles that suffer from stroke but brain cells as some areas die and the patient loses ability to move his muscles. iBrain is working to restore this function. The neural interface recognizes motor impulses in real-time and transforms them into an avatar’s movements on the screen. The patient can’t move his right hand but he can imagine doing it and will see that the avatar on the screen does the same thing. By imagining the movement multiple times the patient can eventually restore the neuronic chains in his brain which are responsible for moving muscles. If several neurons died the neighboring neurons can take over provided that the patient continues to train. The next stage is a computer game in which the patient is to walk and move arms in order to continue playing. He can play by imagining the movement and hence training respective brain areas.
iBrain trains the artificial intelligence to read brain signals using machine learning. AI that has learned to detect brain signals will detect them according to certain patterns and teach patients to move.
“We can see changes in the brain before they transform into muscle movements. We often bring back hope to the patients who lost it after ordinary rehabilitation,” Sonkin said.
iBrain is not the only company that uses AI for brain recovery. While it studies brain signals from the head surface, Elon Musk’s Neuralink announced in 2018 that it plans to do the same by implanting chips inside the brain. Direct connection to brain neurons provides for more precise signal detection. The technology could help in epilepsy and paralysis treatment as well as to improve a person’s efficiency by accelerating their interaction with information – which should “help the mankind to avoid becoming slaves to artificial intelligence.”
By Natalia Kuznetsova