Medindia LOGIN REGISTER
Medindia
Novel ‘Mind-Reading’ Implant Uses Brain Signals to Write

Novel ‘Mind-Reading’ Implant Uses Brain Signals to Write

by Dr. Trupti Shirole on Nov 14 2022 11:50 PM
Listen to this article
0:00/0:00

Highlights:
  • A person may lose the ability to speak due to neurological conditions, including amyotrophic lateral sclerosis (ALS) or brain damage
  • Using a neuroprosthetic device that translates brain waves into complete sentences, a paralyzed man, unable to speak or type, has been able to spell out more than 1,000 words
A new study uses technology to enable patients with speech paralysis to silently spell out messages. Although it has only been tried on one person so far, it could open the door for future strategies that could change people’s lives who have trouble communicating because of paralysis.

Advertisement

What is Anarthia

Anarthria is the loss of the ability to speak verbally. Numerous neurological conditions, including amyotrophic lateral sclerosis (ALS) or brain damage, may be to blame for this. Anarthria significantly lowers patients’ quality of life because they may find it difficult to communicate with family, friends, and caregivers and may experience limb paralysis that prevents them from using assistive devices. Therefore, it is evident that new technological approaches in this field are needed.


Advertisement

What are Neuroprostheses

Neuroprostheses are broadly defined as devices that attempt to replace lost nervous system functions. The 36-year-old man who took part in this study communicates with the help of a neuroprosthesis that is moved around by tiny head movements. After a brain stem stroke, he experienced anarthria and severe muscle paralysis in all four limbs.


Advertisement

Brain-Computer Interface Helps Paralyzed Patients to Communicate

The novel technology formed a brain-computer interface (BCI) using implanted electrodes connected to a computer rather than depending on head motions. The words he was trying to say were translated from his brain signals and shown on a screen as a result.

BCIs themselves are not new; they have successfully been used to enable a paralyzed man to type by simulating the motions of writing by hand. Patients with paralyzed limbs have also been helped by similar techniques. Even with the same individual, the researchers behind the new study had utilized this technology to decode speech when he attempted to vocalize out loud in the past. Since not all paralysis patients find it easy or possible to speak out loud, they instead looked into the possibility of using quiet attempts to talk to control a communication brain-computer interface.

When the participant silently tried to utter the relevant word from the NATO phonetic alphabet, the deep-learning algorithm employed in the BCI was trained to decipher individual letters of the English alphabet. The participant would have needed to ‘speak’ India/Foxtrot/Lima/Sierra to spell out ‘IFLS’, for instance. The participant was instructed to try to squeeze his right hand to indicate that he had finished spelling his message.

The machine could generate sentences at a rate of 29.4 characters per minute with an average mistake rate of 6.13% using a vocabulary of 1,152 words. The mistake rate was 8.23% when this was increased in subsequent experiments to include a vocabulary of nearly 9,000 terms.

To try to replicate this strategy in more people, the authors emphasize the necessity for additional research. They are upbeat about the possibilities of these findings though, saying that "future communication neuroprostheses could enable users with severe paralysis and anarthria to control assistive technology and personal devices using naturalistic silent-speech attempts to generate intended messages."

Brain-computer interface research is advancing in a variety of fields. Investigating these new technologies may hold the key to restoring voice function for patients with anarthria.

Source-Medindia


Advertisement