Neuroscientists find way to decode text from brain speech signals

Neuroscientists find way to decode text from brain speech signals
A digital illustration of brain neurons. Photo: Pixabay

A study funded by Facebook aims to improve communication with paralyzed patients

Doctors have found a way to turn brain speech signals into written text. The breakthrough was made possible by a Facebook-funded research project that aims to improve communication amongst paralyzed and severely disabled patients in the coming years.

The experiment is the first to demonstrate the extraction of words from brain activity as it is converted into text fast enough to keep pace in a natural conversation.

Currently, the brain-reading software only works for specific sentences it has been programmed with, but neuroscientists are hopeful that this will pave the way towards a far more powerful software that can decode brain signals in real-time.

Credit for the work goes to doctors at the University of California in San Francisco lead by Dr. Edward Chang. In the study published by the science journal Nature, Dr. Chang says, “To date there is no speech prosthetic system that allows users to have interactions on the rapid timescale of a human conversation.”

The three epilepsy patients who agreed to partake in the study are also worth mentioning. The patients were about to undergo neurosurgery when they each had a small patch of tiny electrodes placed on the brain to record brain activity while they answered a set of 9 questions. The response to these questions was then recorded and programmed into the software.

“This is the first time this approach has been used to identify spoken words and phrases,” said David Moses, one of the researchers on the team. “It’s important to keep in mind that we achieved this using a very limited vocabulary, but in future studies we hope to increase the flexibility as well as the accuracy of what we can translate.”