centered image

centered image

New Brain-Computer Interface Translates Handwritten Thoughts Into Text For Paralysis Patients

Discussion in 'Neurology' started by Mahmoud Abudeif, May 20, 2021.

  1. Mahmoud Abudeif

    Mahmoud Abudeif Golden Member

    Mar 5, 2019
    Likes Received:
    Trophy Points:
    Practicing medicine in:

    Scientists have developed a brain-computer interface (BCI) that successfully recognizes the brain activity associated with hand writing individual letters, using this to generate text on a screen. Describing their work in the journal Nature, the study authors reveal that the system was tested on a single paralyzed patient, who was able to type 90 characters per minute simply by imagining that he was writing by hand. This set a new world record for speed of typing with a brain-computer interface.


    The study participant was a 65-year-old man who became paralyzed from the neck down following a spinal cord injury in 2007. A decade later, he had two electrodes inserted into a part of his brain called the motor cortex, which coordinates movement. Using this system, he was able to type 40 characters per minute by visualizing the hand movements required to maneuver a cursor and click on letters displayed on a screen.

    At the time, this represented the fastest rate of typing achieved by any BCI, yet the results of this latest study have more than doubled that record.

    "This approach allowed a person with paralysis to compose sentences at speeds nearly comparable to those of able-bodied adults of the same age typing on a smartphone," said senior study author Jamie Henderson in a statement. "The goal is to restore the ability to communicate by text."

    Despite being unable to actually move his hand, the test subject was instructed to imagine that he was holding a notepad and pen, and to concentrate on writing individual letters. The brain activity associated with these movements was picked up by the electrodes and fed into an algorithm, which was then able to learn the specific neural signature pertaining to each letter of the alphabet.

    While decoding the brain activity behind such fine movements might sound like more of a challenge than deciphering the activity associated with basic actions like moving a cursor in a straight line, the researchers report that the opposite is in fact the case.

    “We've learned that complicated intended motions involving changing speeds and curved trajectories, like handwriting, can be interpreted more easily and more rapidly by the artificial-intelligence algorithms we're using than can simpler intended motions like moving a cursor in a straight path at a steady speed,” said lead study author Francis Willett.

    This is because each letter elicits a highly unique pattern of neural activity, making them easily distinguishable. As Willett explains, “Alphabetical letters are different from one another, so they're easier to tell apart."

    Having learned to recognize the brain activity associated with each letter, the BCI was able to display these characters on a screen roughly half a second after the subject had imagined them. Initially, the system achieved an accuracy rate of 94.1 percent, but this was later increased to 99.1 percent when the researchers added in an autocorrect function, similar to those installed on most smarthphones.


    Add Reply

Share This Page