A paralysed man has used a brain-computer to write down on a screen at speeds almost as fast as an able-bodied adult, consistent with a latest study.
Researchers have developed a way of communication for people with paralysis that uses a computer to show mental handwriting into on-screen words. A computer decodes attempted handwriting movements from brain signals, and should allow much faster communication than was previously possible, scientists say.
They report that a mixture of mental effort and state-of-the-art technology have allowed a person with immobilised limbs to speak by text at speeds rivalling those achieved by his able-bodied peers texting on a smartphone.
The team, based at Stanford University , coupled AI software with a device, called a brain-computer interface (BCI), implanted within the man’s brain.
The software was ready to decode information from the PC to quickly convert the man’s thoughts about handwriting into text on a display screen , consistent with the study published in Nature.
The man was ready to write using this approach more-than twice as quickly as he could employing a previous method developed by the Stanford researchers, who reported those findings in 2017 within the journal eLife.
“This approach allowed an individual with paralysis to compose sentences at speeds nearly like those of able-bodied adults of an equivalent age typing on a smartphone,” said Prof Jaimie Henderson of Stanford University . “The goal is to revive the power to speak by text.”
Researchers say the new findings could lead on to further advances benefiting many people globally who have lost the utilization of their upper limbs or their ability to talk thanks to spinal cord injuries, strokes or amyotrophic lateral sclerosis (ALS), also referred to as motor neurone disease (MND).
“Just believe what proportion of your day is spent on a computer or communicating with another person,” said study co-author Prof Krishna Shenoy. “Restoring the power of individuals who have lost their independence to interact with computers & others is extremely important, which is what’s bringing projects like this one front and centre.”
The participant was ready to reach writing speed of about 18 words per minute with 94.1% accuracy. By comparison, able-bodied people of an equivalent age can write about 23 words per minute on a smartphone.
The authors instructed him to write down sentences as if his hand wasn’t paralysed, by imagining that he was holding a pen on a standard size piece of ruled paper. During this exercise, the BCI used a neural network – a kind of AI – to translate attempted handwriting movements from neural activity into text in real time.
The participant, mentioned as T5, lost practically all movement below the neck due to a spinal cord injury in 2007.
9 years later, Henderson placed two brain-computer interface chips, each the dimensions of a baby aspirin, on the left side of his brain. Each chip has 100 electrodes that devour signals from neurons firing within the a part of the motor cortex – a area of the brain’s outermost surface – that governs hand movement. Those neural signals are sent via wires to a computer, where AI algorithms decode the signals & surmise T5’s intended hand and finger motion.
“This method may be a marked improvement over existing communication BCIs that depend on using the brain to maneuver a cursor to ‘type’ words on a screen,” said lead author Dr Frank Willett said.
“Attempting to write down each letter produces a unique pattern of activity within the brain, making it easier for the PC to spot what’s being written with much greater accuracy & speed.”
Experts say further demonstrations of longevity, safety & efficacy are required before the tactic are often put to widespread clinical use.
The authors suggest that these methods might be applied more generally to any sequential behaviour that can’t be observed directly, like translating speech from someone who can not speak.
This article originally published on BBC Science Focus