Monkeys implanted with brain-computer interface technology have successfully transcribed passages from Hamlet and the New York Times at a rate of 12 words a minute, in a study that could lead to significantly improved assistive technologies.
The research, which was undertaken by scientists at Stanford University, involved implanting a multi-electrode array into the brains of rhesus macaques (Macaca mulatta), allowing it to read brain signals normally responsible for directing arm and hand movements. The technology is designed to interpret these brain signals into cursor movements, allowing individual letters of the alphabet to be selected from an on-screen keyboard.
The monkeys were trained to retype letters shown to them, working through the transcription of both news articles and Shakespearean prose.
Earlier versions of the technology have been implanted in humans, although testing the latest, fastest version with monkeys allowed the researchers to quantify the potential typing rate without human mind wandering getting in the way.
“The interface we tested is exactly what a human would use,” said study lead author and Stanford postdoctoral fellow Paul Nuyujukian. ”What we had never quantified before was the typing rate that could be achieved.”
The study, which is published today in the journal Proceedings of the IEEE, is significant for assistive technologies, as current communication solutions for the severely disabled, such as eye or muscle tracking, remain challenging and time-consuming.
“Our results demonstrate that this interface may have great promise for use in people,” said Nuyujukian. “It enables a typing rate sufficient for a meaningful conversation.”
The use of monkeys, however, is an important step in making the technology and everyday reality for users.
“Brain-machine interfaces for the purposes of neural prostheses have not yet been FDA approved for human use, and thus any research involving humans requires a clinical trial,” explained Nuyujukian in an email to Factor. “The natural progression of clinical research is starting with preclinical animal model studies, as this paper is, and demonstrating the promise for human benefit, at which point the technology slowly transitions.
“Brain-machine interfaces for the purposes of prostheses [are] in the middle of this transition phase, where there are a growing number of clinical trials with human participants, but there is still much research to do which still involves non-human primates.”
According to the scientists, people using the system may not match the monkey’s pace as they would need time to think about what they want to communicate and how – a challenge to model in such research.
Nevertheless, while 12 words a minute is by no means going to rival a master typist, it does represent a significant step up from other technologies, and could be further improved by the integration of auto complete technologies like those found in smartphones.
However, Nuyujukian was unwilling to speculate about potential future advances to the implant technology.
“We are always looking to improve the performance of brain-machine interfaces,” he said. “It is challenging work to try to understand the signals coming from the brain, and we continue to improve this as the research moves forward.
“This work represents the culmination of many years of research in improving communication rates relative to prior studies. We hope that continued improvements will be possible.”