Technological Breakthrough Can Change Stroke Patients' Brain Waves Into Sentences

Brainwave decodification demonstration (YouTube @UC San Francisco)

JAKARTA - For the first time in the world, researchers from the United States have succeeded in developing a neuroprosthetic device that is able to translate brain waves from comatose patients into whole sentences.

"This is a technological milestone for someone who can't communicate naturally," said David Moses, a post-doctoral technician at the University of California San Francisco (UCSF) and co-author of the study published in The New England Journal of Medicine, Thursday, July 15.

David said his findings demonstrate the potential of a technological approach to giving voice to people or patients who have lost consciousness or voice.

This breakthrough involved 36 elderly people who had a stroke at the age of 20 that left him suffering from anarthria. This is a condition in which a person is unable to speak clearly, even though his cognitive function remains intact.

During this time, thousands of people lose the ability to speak due to stroke, accident, or disease every year. According to previous research in this area, research has focused on reading brain waves via electrodes to develop mobility prosthetics to enable users to spell letters.

The new approach aims to provide more natural and faster communication. In a previous study, also conducted by scientists from UCSF, researchers placed electrode arrays on patients with normal speech abilities but undergoing brain surgery.

The goal is to decodify the signal that controls the vocal tract expressing vowels and consonants, and analyze patterns to predict words. However, this concept has not been tested on paralyzed patients to prove its clinical function.

Important Achievements of Neuroengineering

David Moses and his team then launched a new study entitled 'Brain-Computer Interface Restoration of Arm and Voice'. The first participant to join the study was referred to as BRAVO1.

Since suffering a stroke that attacks the brainstem, BRAVO1's head, neck, and limb movements have been severely restricted. As for communicating, the BRAVO1 only uses a pointing rod attached to a baseball cap to poke the letters on the screen.

Together with BRAVO1, the scientists developed a 50-word dictionary. Most of the words are important types of words that are often used in daily activities, such as 'water', 'family', and 'good'. Then, the researchers carried out the operation of implanting high-density electrodes above the motor cortex.

Over the next several months, the researchers recorded neural activity as BRAVO1 attempted to pronounce the prepared 50 words. Then, it uses Artificial Intelligence to distinguish patterns in the data that tie it to words.

As a result, the system that David and his team developed was able to decode more than 18 words per minute with an average accuracy of 75 percent. And the AutoCorrect function, as we can find on mobile phones, contributed greatly to that success.

"To our knowledge, this is the first successful demonstration of fully decoding words directly from the brain activity of a person who is paralyzed and unable to speak," said neurosurgeon on the BRAVO1 patient, Edward Chang, who also co-authored the study.