One of the things that makes us human is our ability to communicate. However, a stroke or other medical condition can reduce this ability without warning. Although Stephen Hawking has done great things with a computer-aided voice, it took a lot of patience and technology to get there. Creating an e-mail or utterance for a voice synthesizer with a tongue-and-thumb or blinking can be quite frustrating because most people only speak about ten words per minute. Conventional language is an average of 150 words per minute. However, scientists recently reported in the journal Nature that they have successfully decoded the signals of the brain directly into speech, opening up a whole new world to those in need of support.
The technician is still only ready for the lab, but claims to be able to produce mostly understandable sentences with the technique. In earlier experiments, it was only possible to create single syllables, not whole sentences.
The researchers collaborated with five individuals who were implanted with electrodes on the surface of their brains as part of an epilepsy treatment. Using the existing electrodes, the scientists recorded the activity while the subjects read aloud. They also included physiological tone generation data during speech. Deep learning correlated the brain signals with the movement of the vocal tract and was then able to reproduce brain patterns into sounds.
The results are not perfect, but 70% of the sentences produced were understood by test subjects ̵
It seems that the training is specific to the speaker. The article even states: "In patients with severe paralysis … training data may be very difficult to obtain." It may limit the benefit for people who can not speak, which in combination with the poor voice quality means that work is unlikely to change anyone's life in the immediate future. However, it is a great first step to helping people with language problems and, of course, the dystopian future with this possible brain interface that will replace the cell phone.
This is more scientific than the last time we looked at a brainwave device. If you want to influence your brainwaves instead of reading them, we've experimented with them.