For people who struggle with movement and speaking, even attempting to talk requires immense effort. However, a new study has shown that Brain-Computer Interface (BCI) technology can recognize words that a person only pronounces in their mind.
BCI technology reads signals in the brain, for example, when a person attempts to control a prosthesis or speak. During this process, the brain signals are converted into text. Thanks to this technology, communication becomes much faster and simpler because no physical exertion is required. This means that people with paralysis will have a much more comfortable means of communication.
The Stanford Study
In the study, led by a team of scientists from Stanford University, microelectrodes were implanted into the motor cortex of four patients who had paralysis due to Amyotrophic Lateral Sclerosis (ALS) and a brainstem stroke.
The implant recorded brain activity when the patients were attempting to speak or only pronouncing words in their mind (inner speech). Although “inner speech” produced weaker signals than attempting to speak, similar regions of the brain were activated during both processes.
This allowed the Artificial Intelligence (AI) to decipher sentences from a vocabulary of up to 125,000 words. The system even managed to capture spontaneous mental activities, such as counting numbers while performing a visual task.
The researchers also built a password function into the device for confidentiality. For example, mentally envisioning the phrase, “chitty chitty bang bang,” activated the decoding with 98% accuracy.
Although fully and flawlessly deciphering “inner speech” is not yet possible, researchers believe that improving the technical equipment and algorithms will make the technology useful for those who cannot speak.
