Artificial intelligence makes it possible for people who have hindered limb movement or are paralyzed to communicate by text using data interpretation from devices placed at the brain’s surface.
The fusion of human brainpower and ultra-modern AI technology has allowed a man with paralyzed limbs to communicate using text on a smartphone at nearing speed achieved by his healthy body parts.
Researchers from Stanford University have integrated artificial intelligence software with an electronic device, called a brain-computer interface (BCI), rooted in the brain of a man with full-body paralysis. The robust AI software has decoded the BCI information and instantly transforms the man’s feelings about handwriting into text on a computer screen.
After this integration, the man wrote using this technology more than twice as fast as he could using a former system developed by Stanford researchers, which reported the findings in a 2017 journal.
The participant who was integrated with BCI produced text at a speed of 18 words, whereas ordinary non-disabled people produced 23 words per minute on a smartphone.
This participant lost all his movement below the neck due to a spinal cord injury. After nine years later Henderson implanted 2 BCI chips on the left side of the brain. Each BCI chip has 100 electrodes but is as tiny as the size of a baby asprin. These electrodes pick up signals generated from neurons that fire in the motor cortex part, a part of the brain’s outermost surface that controls hand movements.
Those neural signals are sent to a computer via wires, where an artificial intelligence algorithm decodes the neural signals and speculates the intended motion of hands and fingers.
These algorithms were devised in Stanford’s Neural Prosthetics Translation Lab, co-directed by Henderson and Krishna Shenoy, a Ph.D. and a professor of electrical engineering and the Hong Seh and Vivian W. M. Lim Professor of Engineering.
Light at the end of the tunnel
Some recent research shows a ray of hope to the millions of people who have lost their body parts, spinal cord injury which causes damage to their ability to speak or restricted the body part movements. These researches could help people communicate with others and live their lives without becoming a problem for others.
Researchers like Willett studied that the brain can preserve its capacity to command full-body movement even if the body lost its ability to control those movements. This can help people achieve complicated designed motion like curved trajectories and handwriting speed, which can be performed smoothly and rapidly with the help of AI algorithms.
Brain powered Handwriting Speed Record
The study showed that a participant set an all-time high record of copying presented sentences at 40 characters per minute, whereas another participant copied at 24.4 characters per minute.
In further tests, that participant was directed to copy sentences where he could generate 90 characters per minute. Later on, he was asked to answer some open-ended questions, which required some halts for thinking, and he generated 73,8 characters per minute. The sentence copying error for the same participant was around one mistake in every 18 or 19 attempted characters. Meanwhile, researchers tried to use autocorrect functions similar to the function included in smartphones, lowering the error rate below 1% for copying. Comparing with other BCI’s, this error was relatively low.
The BCI used in the study is not yet approved for commercial use, and it is limited by law to investigational.
Patent for this BCI and AI has been applied on intellectual property associated with Willett, Henderson & Shenoy by Stanford University’s office.
Source: https://hai.stanford.edu/news/software-turns-mental-handwriting-screen-words
Paper: https://www.nature.com/articles/s41586-021-03506-2
Prathamesh Ingle is a Mechanical Engineer and works as a Data Analyst. He is also an AI practitioner and certified Data Scientist with an interest in applications of AI. He is enthusiastic about exploring new technologies and advancements with their real-life applications
