The research was published in Biomedical engineering of nature, the document being entitled “A stabilized brain-computer interface based on the alignment of multiple neurons. “
Brain-computer interfaces (BCI)
Brain-computer interfaces (BCIs) are devices that enable people with disabilities to control prosthetic limbs, computer curses or other interfaces using their minds.
One of the biggest challenges associated with the use of BCIs in the clinical setting is that neural recordings can be unstable. The individual controlling the BCI may eventually lose control due to variations in the signals picked up by the BCI.
Whenever this loss of control occurs, the individual must go through a recalibration process. The individual must reset the link between their mental commands and the tasks performed, and another human technician must often be present.
William Bishop is a fellow at the Janelia Farm Research Campus. He was previously a doctoral student and postdoctoral fellow in the Machine Learning Department of the CMU.
“Imagine that every time we wanted to use our cell phone, to make it work properly, we had to somehow calibrate the screen so that it knew which part of the screen we were pointing at,” says Bishop. “The current state of BCI technology is a bit like that. For these BCI devices to work, users must perform this frequent recalibration. It is therefore extremely inconvenient for users, as well as for the technicians who maintain the devices. ”
New machine learning algorithm
The researchers presented a new machine learning algorithm capable of taking into account different signals. The individual is able to maintain control of the BCI even when instabilities are present. The researchers developed it after discovering that the activity of the neuronal population takes place in a small “neural variety”.
Alan Degenhart is a postdoctoral researcher in electrical and computer engineering at the CMU.
“When we say ‘stabilization’, we mean that our neural signals are unstable, perhaps because we are recording from different neurons over time,” says Degenhart. “We have found a way to take different populations of neurons over time and use their information to essentially reveal a common picture of the arithmetic happening in the brain, keeping BCI calibrated despite neuronal instabilities. “
Previous approaches to self-recalibration methods have also faced challenges related to instability. Unlike other methods, this one does not rely on a good performance of the subject during the recalibration process.
Byron Yu is professor of electrical and computer engineering and biomedical engineering at CMU.
“Let’s say that the instability was so great that the subject was no longer able to control the BCI,” explains Yu. “The existing self-recalibration procedures are likely to encounter difficulties in this scenario, whereas in our method, we have shown that they can in many cases recover from these catastrophic instabilities. “
Emily Oby, a postdoctoral researcher in neurobiology at Pitt, also spoke about the issue of instability.
“The instabilities in neural recording are not well characterized, but it’s a very big problem,” says Oby. “There is not a lot of literature we can point to, but anecdotally, a lot of laboratories doing clinical research with BCI have to deal with this problem quite frequently. This work has the potential to significantly improve the clinical viability of BCIs and help stabilize other neural interfaces. “
The document also included authors Steve Chase, professor of biomedical engineering and the CMU Institute for Neuroscience, as well as Aaron Batists, associate professor of bioengineering at Pitt, and Elizabeth Tyler-Kabara, associate professor of neurological surgery at Pitt .
The research was funded by the Craig H Neilsen Foundation, the National Institutes of Health, the DSF Charitable Foundation, the National Science Foundation, the PA Dept of Health Research and the Simons Foundation.