A research team at the University of California, San Francisco (UCSF) has unveiled a built-in interface that converts brain signals into speech.
This study is the first step towards a system to aid conversation in patients with severe paralysis. Patients who have lost their ability to speak now require special effort to show their intentions. However, in the future, it may be improved through such technology.
The UCSF research team asked five patients who had already had surgery to say 100 phrases and recorded brain signals at this time. When this signal data was input into a computer model that reproduces the human speech system, half of the clear speech was synthesized.
The characteristic of this study is that it captures the neural signals that are fired when humans speak a specific voice, not abstract thoughts as before. In other words, it reproduces the movement of the brain signals and the vocal cords that move by them.
In principle, it can be said that it is an approach similar to the technology that captures signals generated from the brain to motor nerves such as limbs and controls robotic arms. The research team said that it uses the part of the brain that controls these speech movements, and that it is not a direct speech, but attempts to decode the movement.
As a result of testing with synthetic speech to see how much it can actually be used in everyday conversation, it is said that they understood an average of 50-70% of words. Of course, there is still room for improvement in this regard. In this experiment, a flexible electrode pad was used to collect the signal ECoG from the brain surface. Experts believe that placing the probe in brain tissue could potentially increase the accuracy.
Even now, there are artificial vocal cords that allow people who have lost the larynx and vocal cords to speak, but if the paralysis is severe or it is difficult to move the muscles, it cannot be said to be effective. In order to help patient communication and restore relations with society, research and development is expected to continue in the future. Related information can be found here .