Study shows that Brain Machine Interface (BMI) can help speech-impaired persons

 

A new study from Caltech shows that a device called a Brain-Machine Interface (BMI) implanted in a person's brain could one day help patients who have lost the ability to speak. In a new study presented at the Society for Neuroscience 2022 conference in San Diego, researchers used BMI to pinpoint the words quadriplegic participants just thought without speaking or mimicking.

"You may have seen videos where a quadriplegic person manipulated a robotic arm and hand using their BMI. or indulge in chocolate," says Sarah Wandelt, a graduate student at Caltech Richard Andersen, professor of neuroscience James G. Boswell, and Caltech's Amanohashi and Chrissy Chen Blaine, Director of the Machine Interface Center.
“These new results are promising in the field of language and communication, says Wandelt.”

Previous studies have had some success in predicting participants speech by analyzing brain signals recorded by the motor cortex when participants whisper or imitate words. But predicting what someone is thinking—the internal dialogue—is much more difficult due to the lack of movement, Wandelt explains. "Algorithms that attempt to predict internal speech have so far only been able to predict three or four words and have either been wrong or not real-time," he claims.

A new study is the most accurate ever in predicting inner words. In this case, brain signals were recorded from individual neurons in a brain region called the supramarginal gyrus, located in the posterior parietal cortex. In previous studies, researchers found that this area of ​​the brain represents spoken language.

Emerging Trends in Brain-Computer Interfaces (BCI)

The team is now extending their findings into an internal language. In this study, the researchers first trained the BMI device to recognize brain patterns generated when a quadriplegic participant spoke or thought of specific words internally. This training time lasted about 15 minutes. A word was then displayed on the screen and participants were asked to say the word in their minds. The results showed that the BMI algorithm could predict the eight words with an accuracy of up to 91%.
Although this research is preliminary, it may help patients with brain damage, paralysis, or diseases such as amyotrophic lateral sclerosis (ALS) that affect language. "Neuropathy can cause total paralysis of the voluntary muscles, rendering the patient speechless and immobile but nevertheless capable of thought and reasoning.” Wandelt asserts that a linguistic BMI would be highly beneficial.
"We have previously shown that we can decode a fictitious grasping hand shape from the human supramarginal gyrus," says Andersen. "The capacity to comprehend voice from this range also implies that the implant can restore speaking and grasping, two crucial human abilities.”

This study, which is in the process of being submitted to the journal but has not yet been peer-reviewed, is entitled 'Online Internal Speech Decoding of Single Neurons in Human Participants', funded by the Tianqiao and Chrissy Chen Brain-Machine Interface Center, the Boswell Foundation, and the National Institutes of Health