"Communication Possible Even for Paralyzed Patients"…Device That Reads Thoughts Released

Interpreting the Brain with BCI Systems and Outputting to Computers

A research team in the United States has developed a device that measures brain activity of severely paralyzed patients who have lost the ability to speak due to amyotrophic lateral sclerosis (ALS) and stroke, and outputs the intended vocabulary.

On the 24th, the scientific journal Nature published articles about 'brain-computer interfaces' (BCI) developed respectively by Dr. Francis Willett's team at Stanford University and Professor Edward Chang's team at the University of California, San Francisco (UCSF). According to Nature, these devices interpreted brain activity of severely paralyzed patients and rapidly outputted it as speech or text.

"BCI enables people who cannot speak to connect with the world"

Brainstem stroke patient participating in the study and an avatar displaying the interpretation [Photo by Yonhap News]

Dr. Willett's team at Stanford University developed a BCI device that reads brain commands of Pat Bennett (68), who was diagnosed with ALS in 2012.

Mr. Bennett's brain could process commands to produce phonemes, but his muscles such as lips and tongue were paralyzed and could not execute them. Therefore, the research team implanted sensors in the brain areas involved in Mr. Bennett's language generation. The sensors measured and interpreted brain activity and outputted the words Mr. Bennett intended onto a computer screen.

According to the research team, Mr. Bennett is recently approaching a natural conversational speed (160 words per minute). Also, this BCI system has been confirmed to reduce the error rate to less than half compared to previously disclosed systems.

Mr. Bennett stated, "This result demonstrates the concept of BCI, and ultimately people who cannot speak will be able to use this technology. This means that people who cannot speak can continue to connect with the larger world, work, and maintain relationships with friends and family."

"Enabling paralyzed patients to communicate more naturally and richly"

ALS patient Pat Bennett and the research team participating in the BCI study
[Photo by Yonhap News]

Professor Edward Chang's team at UCSF revealed a device that measures and interprets brain activity of a woman, Ms. A (47), who lost the ability to speak due to a brainstem stroke 18 years ago.

The research team implanted a rectangular array of electrodes on the surface of the brain region. They then measured brain signals that would have been transmitted to the face, tongue, jaw, and larynx during speech if the stroke had not occurred, interpreted them using a deep learning model, and outputted the content as text, speech, or a speaking avatar.

According to the research team, Ms. A trained the AI algorithm by repeating a conversational vocabulary consisting of 1,025 words over several days. As a result, a BCI system was developed that interprets and outputs an average of 78 words per minute. The error rate was significantly lower compared to existing speech BCIs.

The research team said, "This BCI system can output speech in the patient's pre-injury voice and express brain signals through avatar facial expressions. It will provide the possibility for paralyzed patients to communicate in a more natural and richer manner."

This content was produced with the assistance of AI translation services.