Brain-Computer Interfaces can restore speech for the speechless

in Popular STEMlast year (edited)

Researchers in neurosurgery are making strides towards brain implants that can restore speech for people with brain injuries and neurological illnesses.


Introduction

image.png

Pixabay license from PatoLenin at source

Last Wednesday, MIT Technology Review posted the article, Brain implants helped create a digital avatar of a stroke survivor’s face describing recent advances in brain computer interface (BCI) technology. Although the headline only mentions a stroke survivor, the article itself tells us about two patients who experienced advances in their ability to communicate through the use of brain implant technologies.

In the February 3 Nature article, A high-performance neuroprosthesis for speech decoding and avatar control, a team of researchers led by Edward F. Chang describe their progress with a study participant who had a stroke nearly two decades ago.

In the August 23 paper, A high-performance speech neuroprosthesis, a team led by Stanford researchers described their work with a research participant who suffers from ALS. The Stanford cadre included Jaime Henderson,, Krishna Shenoy, and Shaul Druckmann.

Through the use of these neuroprosthetic devices, these two independent teams were able to restore speech capacity for both of their research participants.

Speech encoding and avatar control for a stroke survivor

The 47 year old study participant in A high-performance speech neuroprosthesis had suffered severe paralysis after a stoke that occurred more than 18 years ago. As a result of the paralysis, the patient was unable to speak, or even vocalize speech sounds. She also experienced paralysis of all four limbs which made her unable to write or type.

Chang and his team - who have been working on this problem for more than a decade - demonstrated in 2021 that they could capture brain signals from a stroke survivor and translate those signals into english words and phrases. In the 2023 paper, the team now demonstrated their progress with a larger device that doubled the number of electrons. This newer, larger implant is about the size of a credit card.

Interestingly, the device doesn't actually read the operator's thoughts, but rather it observes the signals that go to the muscles of the tongue, voice box, jaws, and lips. With these symbols it is able to decipher and report the phonetic sounds and facial expressions that are being formed.

With the higher capacity device, the team was able to increase the processing speed with a 1,024 word vocabulary to a rate of 78 words per minute. At this rate, the device encountered an error rate of 23%. Another advance that the team reported is the ability to translate the thoughts directly into audible speech (using a speech synthesizing computer). Further, the computer display also had an on-screen avatar that was able to express three emotions, each at three levels of intensity. The available emotions are sad, surprised, and happy. Presumably the intensity levels would be low, medium, and high, although I didn't see that mentioned. The team even used the patient's wedding video in order to make the speaking voice sound as authentic as possible.

Restoring speech for an ALS patient

The second paper - A high-performance speech neuroprosthesis - was first published as a pre-print in January. In this paper, the team reports on progress with a patient who had lost her ability to speak as a result of an 8+ year battle with ALS (aka "Lou Gehrig's Disease"). Using their technology, the research participant was able to use a smaller BCI to speak at a rate of 62 words per minute. According to the paper's Abstract, this is "3.4 times as fast as the previous record", and it also approaches the speed of natural human speech, which is around 160 words per minute.

In this study, the device achieved an error rate around 9% with a 50 word vocabulary and 23.8% with a 125,000 word vocabulary. The authors suggest that this may be the first successful demonstration of a BCI that could decode speech with such a large vocabulary. The BCI that this team used was composed of four separate implants, each about the size of an aspirin.

In this case, the participant trained the device by reading "syllables, words, and sentences" in 25 training sessions. When training was completed, the device was tested by having her read other sentences that had not been used during training.

Summary

These studies show remarkable advances towards being able to restore speech for survivors of brain injuries and neurological diseases, but there is still a long way to go to make a device that is useful outside of the research labs.

As noted above, speeds of 62 and 78 words per minute are still less than half of the naturally achieved 160 words per minute, but they are far ahead of previous capabilities. Additionally, error rates of 23-24% remain fairly high. Further, the BCIs must be cabled to large computer systems in the lab to perform their processing, so they can't be used outside of the study environment.

However, both teams believe that future advances will lead to practically useful devices that will restore the ability to speak for people who have lost it due to brain injury or illness.


Thank you for your time and attention.

As a general rule, I up-vote comments that demonstrate "proof of reading".




Steve Palmer is an IT professional with three decades of professional experience in data communications and information systems. He holds a bachelor's degree in mathematics, a master's degree in computer science, and a master's degree in information systems and technology management. He has been awarded 3 US patents.


image.png

Pixabay license, source

Reminder


Visit the /promoted page and #burnsteem25 to support the inflation-fighters who are helping to enable decentralized regulation of Steem token supply growth.

Sort:  

The image seems generated by Ai. I think I read about this before and it will be a big contribution of technology to people who can't speak their words. This is unbelievable invention, I must say!