Everyday use of brain-controlled robots and neuroprostheses is the main promise of the brain-machine interface (BMI) for people with severe motor disabilities. A new study from the University of Texas at Austin takes a step forward for brain-machine interfaces — computer systems that translate brain activity into action.
In this study, several people with motor disabilities were able to operate a wheelchair that converts their thoughts into movement. The study is also important because of the non-invasive equipment used to operate the wheelchair.
José del R. Millán, a professor in the Chandra Family Department of Electrical and Computer Engineering at the Cockrell School of Engineering, who led the international research team, said: “We’ve shown that the people who will be the end users of these types of devices can navigate a natural environment using a brain-machine interface.”
The idea of a mind-powered wheelchair has been explored for years. Yet most efforts rely on non-disabled people or incentives that make the wheelchair control the user, rather than the other way around.
In this case, three people with tetraplegia – the inability to move the arms and legs due to spinal injuries – operated the wheelchair in a chaotic, natural environment with varying degrees of success. The interface captured their brain activity and a machine learning algorithm converted it into instructions for operating the wheelchair.
Scientists noted, “This is a sign of future commercial viability for mind-powered wheelchairs that can help people with limited motor function.”
“The study is also important because of the non-invasive equipment used to operate the wheelchair.”
Surprisingly, scientists did not implant any device into the participants or use any kind of stimulation on them. Participants had to wear a cap with electrodes that recorded the brain’s electrical activity, known as an electroencephalogram (EEG). These electrical signals were amplified and transferred to a computer, turning each participant’s thoughts into action.
Two key dynamics contributed significantly to the success of the study. The first concerns a training program for the users.
The techniques for visualizing chair movement were taught to users in the same way they would have learned to move their hands and feet. The brain activity of the study participants changed as they gave commands, and the scientists were able to track these changes.
The second contributor borrowed from robotics. To better understand their environment, the scientists equipped their wheelchairs with sensors. In addition, they used robotic intelligence software to drive the wheelchair accurately and safely by filling in the gaps in the users’ commands.
Millan said, “It works just like horseback riding. The rider can tell the horse to turn left or enter a gate. But the horse will ultimately have to figure out how best to execute those commands.”
Team members in the project are Luca Tonin from the University of Padova in Italy; Serafeim Perdikis of the University of Essex in the United Kingdom; Taylan Deniz Kuzu, Jorge Pardo, Thomas Armin Schildhauer, Mirko Aach and Ramón Martínez-Olivera from Ruhr-Universität Bochum in Germany; Bastien Orset of École polytechnique fédérale de Lausanne in Switzerland; and Kyuhwa Lee of the Wyss Center for Bio and Neuroengineering in Switzerland.
- Luca Tonin et al. Learning to drive a BMI-powered wheelchair for people with severe tetraplegia. iScience. DOI: 10.1016/j.isci.2022.105418