Researchers at the University of Texas, Austin, hit a milestone by introducing mind-controlled wheelchairs, proven to assist quadriplegics in acquiring new mobility. These wheelchairs were demonstrated to translate users’ thoughts into mechanical instructions, helping them navigate in a natural, cluttered environment. Additionally, the experimenters also showed that this mind-controlled wheelchair could be eligible to use with training for a prolonged period.

Image credit – Wikimedia Commons 

How do these mind – controlled wheelchairs function?

Jose del R. Milan, a corresponding author at the University of Texas, explains the significant factors which assist in the device’s functioning. These include the user and the brain-machine interface algorithm. It’s the mutual learning, interaction, and subsequent coordination of these elements which helps in the successful operation of this innovation. Milan also calls this “a potential pathway for improved clinical translation of noninvasive brain-machine interface technology. “

How did the research take place?What were the inferences drawn? 

The study for determining this potential pathway was conducted on three quadriplegic people. These recruited participants had to undergo training sessions three times per week for 2 to 5 months. To observe the progress of these participants, they were required to wear a skull cap. This skull cap would make use of electroencephalography to detect brain activities. These detected activities were further converted to mechanical commands for the wheelchairs through a brain-machine interface device.

The participants were then asked to envision moving their body parts to control the wheelchair’s direction. In particular, to turn left, they needed to think about moving their both hands and both feet to turn right.

Image credit – Pixabay

While similar levels of accuracy, i.e., around 43 percent to 55 percent, were observed in the alignment of the device’s responses with the three users’ thoughts in the first training session, a notable improvement in the accuracy of participant 1 was observed throughout the training. This accuracy was analysed by the brain-machine interface device team and reached over 95 percent in the 1st participant. A similar improvement was observed in the 3rd participant. 

This improvement was credited to enhanced feature discriminant, the algorithm’s ability to discriminate the brain activity pattern encoded for “go left” thoughts from that for “go right.”The team then found a connection between the machine learning of the device and the learning capacity of the participants. According to Millán, this happened due to cortical reorganisation that was determined to be the result of the participants’ learning process as observed through the EEG results, a test that measures the brain’s activity.  

Interestingly, participant 2 showed no significant changes in brain activity patterns, and his accuracy remained stable throughout the training. According to Milan, this indicates that machine learning alone is insufficient for successfully manoeuvring such a mind-controlled device. 

This claim was further consolidated when all participants were asked to drive their wheelchairs across a cluttered hospital room. This room with all its obstacles was a simulation of the real world. While participants 1 and 3 succeeded in finishing the task, participant 2 failed to complete it.

What is the significance of this study? 

This longitudinal study is one of the first to evaluate the clinical translation of non-invasive brain-machine technology in tetraplegic people. Milan with his team also aims to conduct a more detailed analysis of all participants’ brain signals to understand their differences and possible interventions for people struggling with the learning process in the future.

This study was published in the journal iScience.