American researchers have created the interface between brain and computer able to translate what you are thinking and instantly display it as text with an accuracy rate of 90%.
The working group from the University of California has published an article in Journal of Neural Engineering, in which the light was shed on the successful development of interface that may read thoughts – convert signals from the human brain into text. Developed hardware and software complex is operating in real time and may boast of an accuracy rate of 90%. For the machine to be able to read mind, it shall first be taught how to do this (experiment included less than 7 minutes of training data). The training involved recording and analyzing of signals from auditory cortex of the brain used by test person when constructing a sentence in the brain. Neural signals of vowels and consonants are than translated into text that may be displayed or reproduced by using the speech synthesizer. Moreover, this machine may recognize words that were not taught by it before – by simply creating new word structures from vowels and consonants.
Study leader David Moses said that until today there were no solution of this type, and no one could resolve the problem of automatic interpretation of neural signals from auditory cortex of the brain.
This machine will help to solve severe communication problem of people with reduced capabilities, not able to speak for various reasons. Thus, it may significantly speed up text entry, interpretation of statistical data from machines, or may establish communication with person, who is unable to speak due to challenging environment, but shall transmit some textual information, while being in protective clothing, noisy places, etc. At the same time, there are ethical aspects of mind-reading issue – “last bastion” of privacy for a person.