Morning coffee science

"Thought"- controlled robotic arms


Buz Chmielewski has not been able to voluntarily move his hands and feet for over three decades. When he was young, Mr Chmielevski got into a surfing accident that left him paralyzed. In January of 2019, this man was able to perform his first voluntary movement in quite some time - he cut a piece of cake using a fork and a knife. Interestingly, this latest achievement was not done due to extensive kinesitherapy but thanks to electrodes implanted in his brain. These implants can sense Mr Chmielewski's thoughts and turn them into electric impulses that can move prosthetic robotic arms.

The clinical trial in question was carried out by John Hopkins University’s scientists, who are world-acclaimed in the field of "brain-machine interface" (BMI). The brain-machine interface field is a branch of neuroscience that investigates devices capable of measuring, analysing, and transmitting central nervous system impulses to commands that could be understood by computer software.

To illustrate what BMI is one could look into Mr Chmielewski's case.

The first step is "sensation" - electrodes implanted in the patient's brain sense changes in electric potential when he thinks that he wants to cut a piece of cake. Almost at that same moment, cutting-edge artificial intelligence algorithms decode the electric impulses sensed by the electrodes. These decoded signals are then sent to prosthetic robotic arms which end up performing Mr Chmielewski's desired command.

The fact that a paralyzed patient is able to control robotic arms using only their "thoughts" is, however, not a new discovery for the field of brain-machine interface. The significance of these latest findings, though, lies in the difficulty of simultaneously controlling two robotic arms at the same time. Dr Gabriela Canterero was a lead contributor to the project and here is what she has to say about the significance of solving this engineering problem: "Simultaneous brain-machine interface control of two limbs is a particular challenge because it’s not a simple 1+1 summation of what the left arm is doing plus what the right arm is doing in the brain, but more like trying to calculate the sum of the two arms as 1 plus 1 equals 3.8,” In other words, Dr Cantarero makes a point that decoding the electric impulse sensed by the electrodes in the patient's brain is a hard problem because simultaneous use of hands creates neuronal synergy (and, thus, a more complex electric pattern to decipher than if both hands were used separately).

Solving the problem of simultaneously controlling two robotic hands opens the door to next-generation BMI applications. Everyday applications of BMI require coordination between two hands: making food, tying one's shoes, or buttoning up one's shirt. It is important to note that there are still various unresolved questions and problems with the commercial application of this technology. For example, the electrode implants in the patient's head can create discomfort and be esthetically undesirable. Neuralink, Elon Musk's neuro-engineering company, tries to address this problem. Neuralink is working to develop its own minuscule electrode implants that would not be visible from the "outside". The company is also focussing on engineering a solution to streamline the implantation surgery (30-min surgery at your local hospital) as well as being able to charge itself overnight by induction (the chargers would be located at your bedside). As futuristic and exciting as this sounds, it is still too early to cheer for Neuralink. Neuralink has been conducting their trials in pigs and only recently had their product be designated by the FDA as a breakthrough medical device.

Article was prepared by Matas Vitkauskas on behalf of INA

Watch the clinical trial with Buz Chmielewski.

Watch Neuralink's product demo.