TechDuke's brain implant project advances communication for the silent

Duke's brain implant project advances communication for the silent

A brain implant will allow the patient to communicate - illustrative picture
A brain implant will allow the patient to communicate - illustrative picture
Images source: © Wikimedia Commons | Harrypotterrrrr
ed. KMO

1:44 PM EST, November 9, 2023

A team of scientists at Duke University in the United States have developed a small device that can be implanted in the brain and read signals from the speech center. This innovation may pave the way for improved communication methods for individuals who, for various reasons, have lost their ability to speak.

The device's development was spearheaded by Dr. Gregory Cogan, who authored a paper discussing it in "Nature Communications". Dr. Cogan explains that serious motor disorders, such as Amyotrophic Lateral Sclerosis (ALS) or locked-in syndrome, often render individuals speechless. Current assistive communication technologies are frustratingly slow and difficult to use.

Brain implant aims to improve communication

The small implants currently being tested and employed by the Duke University team are placed on the surface of the brain. The team compares their operation to listening to an audiobook at half speed. Their slow functioning results from the limited number of sensors the device incorporates.

The newly developed implant, about the size of a postage stamp and fabricated from a flexible material, houses 256 microscopic sensors. The greater the number of sensors, the more accurately the brain's complex neuronal activity can be decoded.

The cutting-edge invention has already been tested on patients who were undergoing unrelated brain surgeries, such as treatments for Parkinson's Disease or tumor removals. During the tests, the patients were asked to listen and repeat nonsensical words like "ava", "kug", or "vip", as the implant recorded their brain's responses.

The researchers clarify that when we speak, the brain coordinates over a hundred muscles in our mouths, tongues, and larynxes. Leveraging the data accumulated, the scientists were able to train artificial intelligence to recognize sounds based on corresponding brain activity. The system currently holds an average success rate of 40% for syllable recognition.

Although the success rate may appear relatively low, the time constraints were severe. The total procedure, including speech data collection, only lasted a brief 15 minutes with 90 seconds dedicated solely for data collection. This greatly limits the amount of data captured, in contrast with older iterations of the device that collected data over several hours, or even days.

Thanks to a generous grant of nearly $2.5 million, the scientists are planning on developing a wireless version of the implant. Dr. Cogan shared that the focus of their current research is a device without external wires or power supply requirements.

Despite the need for further research and testing, the inventors are optimistic that their device will eventually begin to restore speech capabilities. Researcher Dr. Jonathan Viventi mentioned that although the mechanically engineered speech is slower than natural speech, they are progressing towards their goal.