Brain implants are becoming an interface between the brain and the computer. Several leading companies such as Facebook and Neuralink are developing these implants to enable the future of devices to be controlled with just thoughts. Now, a brain-computer interface (BCI) system has for the first time successfully read thoughts with 74% accuracy.
This BCI was developed by Stanford University and is installed in the motor cortex of the brain that controls speech. Four test participants are individuals with severe paralysis due to ALS disease and a stroke in the base of the brain. An AI algorithm was used to read the signals produced by the participants' motor cortex and it was trained to recognize 125,000 words.
This new system is different from other BCIs because it continues to read the participants' thoughts. Previous systems read the signals the brain sends to the mouth when they want to speak or the signals to the hand when they want to write. With this new BCI system, theoretically the thoughts of people who are completely paralyzed or in a coma can be read.
To prevent everything they think from being read by the algorithm, participants are given a password to activate the BCI before it starts reading thoughts. The researchers will further improve the study to improve the accuracy of mind reading.
The research paper was published in the journal Cell.