Summary: Training tetraplegic patients to compete in an online competition to control an avatar with BCI, researchers report the most dramatic improvements in computer augmented performance occurs when both humans and machines are able to learn.
People using brain-computer interface are more efficient when both human and machine are allowed to learn. EPFL researchers trained two tetraplegic users to compete in the international Cybathlon BCI race. Both incrementally learned how to control the BCI, and obtained the best performances at the competition, confirming researchers’ hypothesis that mutual learning plays a fundamental role in BCI training.
Brain-computer interfaces (BCIs) are seen as a potential means by which severely physically impaired individuals can regain control of their environment. BCIs use the electrical activity in the brain to control an external device. They have seen growing use in people with severe motor disabilities, for communication (by controlling a keyboard), mobility (by controlling a powered wheelchair), and daily activities (by controlling a mechanical arm or other robotic devices). But establishing such an interface is not trivial.
In a study published in the open-access journal PLOS Biology, a group of researchers at the École Polytechnique Fédérale de Lausanne in Geneva (Campus Biotech), led by José del R. Millán -Defitech Foundation Chair in Brain-Machine Interface, School of Engineering, suggests that letting humans adapt to machines improves their performance on a brain-computer interface.
The scientists trained two tetraplegic subjects to compete in the Cybathlon BCI race 2016, an international competition where competitors control an on-screen avatar with brain-computer interfaces. The results suggest that the most dramatic improvements in computer-augmented performance are likely to occur when both human and machine are allowed to learn.
How does it work?
With BCIs, the electrical activity is typically detected at one or more points of the surface of the skull, using non-invasive electroencephalographic electrodes, and fed to a computer program that, over time, could improve its responsiveness and accuracy through learning.
As machine-learning algorithms have become both faster and more powerful, researchers have largely focused on increasing decoding performance by identifying optimal pattern recognition algorithms. It was suggested that the performance could be improved if the operator and the machine both engaged in learning their mutual task. However, direct evidences of such a learning mechanism were rare and fragmented.
Winning the gold medal
At EPFL, two tetraplegic adult men were trained with a BCI system designed to detect multiple brain wave patterns. Training took place over several months, culminating in an international competition, called the Cybathlon, in which they competed against ten other teams. Each participant controlled an on-screen avatar in a multi-part race, requiring mastery of separate commands for spinning, jumping, sliding, and walking without stumbling. The two subjects marked the best three times overall in the competition, one of them winning the gold medal and the other holding the tournament record.
A specific training
Electroencephalography recording of the subjects during their training showed that the brain wave patterns related to imagined movements (called sensorimotor rhythms), which have been adopted to control the avatar, became stronger over time, indicating that the subjects were learning how to better control the BCI during the training.
The authors believe they have maximized the chances for human learning by infrequent recalibration of the computer, leaving time for the human to better learn how to control the sensorimotor rhythms that would most efficiently evoke the desired avatar movement. Training in preparation of a competition may also contribute to faster learning, the authors propose.
“This study is one of the few to provide multi-faceted evidence on the efficacy of subject learning during BCI training,” the authors said. “Contrary to the popular trend of focusing on the machine learning aspects of BCI training, a comprehensive mutual learning methodology could strongly promote acquisition of BCI skills.”
Publisher: Organized by NeuroscienceNews.com.
Image Source: NeuroscienceNews.com image is credited to Erik Tham.
Original Research: Open access research for “The Cybathlon BCI race: Successful longitudinal mutual learning with two tetraplegic users” by Serafeim Perdikis, Luca Tonin, Sareh Saeedi, Christoph Schneider, and José del R. Millán in PLOS Biology. Published May 10 2018.
The Cybathlon BCI race: Successful longitudinal mutual learning with two tetraplegic users
This work aims at corroborating the importance and efficacy of mutual learning in motor imagery (MI) brain–computer interface (BCI) by leveraging the insights obtained through our participation in the BCI race of the Cybathlon event. We hypothesized that, contrary to the popular trend of focusing mostly on the machine learning aspects of MI BCI training, a comprehensive mutual learning methodology that reinstates the three learning pillars (at the machine, subject, and application level) as equally significant could lead to a BCI–user symbiotic system able to succeed in real-world scenarios such as the Cybathlon event. Two severely impaired participants with chronic spinal cord injury (SCI), were trained following our mutual learning approach to control their avatar in a virtual BCI race game. The competition outcomes substantiate the effectiveness of this type of training. Most importantly, the present study is one among very few to provide multifaceted evidence on the efficacy of subject learning during BCI training. Learning correlates could be derived at all levels of the interface—application, BCI output, and electroencephalography (EEG) neuroimaging—with two end-users, sufficiently longitudinal evaluation, and, importantly, under real-world and even adverse conditions.