According to The Scientist, a new program has been developed that converts brain activity into spoken words. It does so by tracking electrical messages passed to the mouth’s muscles to decode what the brain is attempting to communicate. Although the program is still in early stages, it could be used to help people who are unable to speak due to a stroke or brain disease.
“We want to create technologies that can reproduce speech directly from human brain activity,” Edward Chang, a neurosurgeon at the University of California, San Francisco, who led the research, said. “This study provides a proof of principle that this is possible.”
The program would place electrodes deep in the brain. Due to the invasive nature of the technique, it has only been tested on five people with epilepsy so far. During the tests, the people spoke, allowing the computers to figure out associated brain signals. Now, the researchers must see if the technology works for people who can’t speak, which will be more difficult.
To decode brain signals, researchers used the electrodes to track signals from the brain when the volunteers read sentences aloud. An algorithm analyzed the instructions with a model of how the vocal tract makes sounds. A processing stage then converted the predicted movements into sentences spoken out loud. 101 sentences were played to listeners, who were asked to identify the words from a 25-word list. 43 percent of words were transcribed accurately.
The signals in the brain that control speech are more complex than those to make basic movements with arms and legs. Thus, a synthetic speech system would likely be limited to a specific set of words.
Read the full story here.