“Baby steps. Baby steps. Baby steps through the office. Baby steps out the door. It works. It works! All I have to do is take one baby step at a time and I can do anything!”
Anyone who has seen my phone or been in my office knows that I’m a big fan of Bill Murray. While “What About Bob” does not share the same cult status as “Caddyshack” or the brilliance of “Rushmore” or the stop-motion adventure of “Fantastic Mr. Fox,” its mantra of getting through life taking one baby step at a time is applicable to all manner situations – including neurotech research. I’m not the only one who thinks so. Professor Dario Farina of Imperial College of London has a similar approach.
At Farina’s keynote at EMBC in Berlin this summer he cited electrophysiology research performed by Emil du Bois-Reymond in 1849. “I found a very classical electrophysiological technique and what I think is the very first paper providing evidence that you can record electrical activity from muscles when you make a muscle contraction.” The author of the paper, who Farina supposes was also the subject of the experiment, placed his fingers into a liquid and used a potentiometer to detect very small changes in electrical signals. “This lecture started with very conventional techniques. I thought it was good to look backwards,” he said.
Farina later referred me to a much more contemporary Journal of Clinical Neuroscience paper outlining the history and founding fathers of EMG. I’m sure that all of you are familiar with this technological arc, but I, as a non-engineer, found the history to be fascinating. The authors dutifully laid out the pathway of development from the Leyden jar in 1745-1746, to Galvani and his dispute with Volta in the late 18th century (and Napoleon’s role in this whole saga), on to Bois-Reymond’s improvements of Oersted’s galvanometer, through the incremental improvements of the 19th century, and concluded with the more significant advances of the 20th century.
“There was a landmark paper in 1929 on the invention of a very specific electrode for muscle recording that we still use – basically unchanged, today.” All of these events and discoveries are the foundation for the EMG testing devices currently in use. Raise your hand if you’ve had the pleasure of being the subject of a nerve conduction study.
I asked Farina what he thought these early authors would think of the work that he is doing today. “I hope that some of the things we do will last. But the contributions that I mentioned in my talk are immense. We have done very small incremental steps,” he said humbly. “Of course, when you take so many incremental steps in so many generations of scientists, then the gap with respect to the early techniques becomes very big. Some examples of the individual work that I’ve discussed are far superior to any incremental steps that we have made in the past 19 years.” Baby steps.
The scope of Farina’s work emphatically contradicts his humility. Perhaps if I were able to communicate with him in his native Italian a little tiny bit of vanity would show through – but I doubt it. I believe that he would be happy to say little and let his 400+ (co)authored papers speak for themselves. His “incremental steps” are not without great leaps, too. “We managed to decode peripheral electrical activity from muscles and reconstruct the drive in the neuro cells and to see spikes from those neuro cells from non-invasive recordings,” he modestly explained. “That was a breakthrough moment. Of course, it didn’t come from evening to morning, but it was a breakthrough that came in a relatively short period of time. And then of course, we worked on refining it.”
As a part of my personal interest in better understanding the role of closed loop systems in neurotech research, I asked Farina about why the cardiac field has been able to close the loop rather quickly and why neuro is taking so much longer. Obviously, it’s complicated.
“If we intend to decode motor and encode sensory, which is just one aspect of closed loop, that – in my opinion - is very complicated. Encoding sensory is very, very challenging, much more challenging than just decoding motor,” Farina said. “One of the reasons for this is that in peripheral nerves there are far more sensory fibers than motor fibers in a ratio of 10:1. In order to encode the information to biomimetically reproduce that information and to transfer it to the central nervous system is an order of magnitude more difficult than decoding the motor part. The real bottleneck is the sensory feedback in these systems.”
There are several research projects underway seeking the optimal point within the nervous system to introduce the sensory feedback. Farina suggests that the simplest level would be at the nerve and perhaps the most difficult would be to stimulate the brain itself. Even if one were to undertake the “straightforward” route and stimulate at the nerve, producing natural sensation has been very challenging.
“Some particular aspects of sensory feedback, like proprioception, are basically impossible to reproduce naturally,” he elucidated. “It is extremely challenging to reproduce so much afferent information. This will remain a bottleneck for a long time. There are no real breakthrough pathways. Some people are doing a lot of work, but there is nobody who says they have a revolutionary idea to completely change the way in which we do things.”
I can fairly well assure you that the general public pays little attention to the importance of sensory feedback (until they get their feelings hurt, of course.) The prevailing thought is that if an amputee receives a robotic arm that the added benefit of sensation is cool but not necessary because they believe the robot to be inherently smart enough to accomplish tasks without the need for sensation. Amy Orsborn, Assistant Professor of Electrical & Computer Engineering and Bioengineering at the University of Washington recently posted the perfect demonstration of the need for sensory feedback in motor control. Orsborn showed a video of her dog’s first encounter with shoes.
“We know from clinical examples of patients who are completely without sensation, but who have complete motor control that they have big difficulties making movements,” said Farina. “If you have a perfect motor decoding, you have a very big difficulty performing even basic movements if you don’t have sensation.”
Beyond restoring sensation for SCI and neural prostheses, Farina sees great promise in using closed loop systems to improve plasticity. Being able to drive plasticity would move some current approaches from assistive technologies to fully realized therapies. “That’s a Holy Grail in neurotechnology. If we could drive the plasticity of the nervous system, the brain in particular, as we please, then we could approach stroke and many other debilitating diseases with much better and much faster results.”
With so many promising applications that can be advanced dramatically by having a closed loop platform, what is standing in the way? One answer is that it’s not the technology that’s standing in the way. When we stimulate the brain, we still don’t understand what is happening at an empirical level. Stimulation excites cortical cells but why and to what end?
“I think the basic neuroscience still has a lot to do in order to drive neuroprosthetics in general. In many cases, ignorance of the basic mechanisms is the bottleneck, more than technology,” lamented Farina. “Technologically of course, there are many bottlenecks depending on the application. If we think of recording inside the brain without having to open the skull, that has been a limitation that may be addressed by things like Neuralink. The possibility of mounting millions of electrodes – because that is the order of magnitude of the cells from which we could record – that is a technological bottleneck.”
In addition to the large populations of cells that are screaming for attention, there remains the challenge of transferring the signals outside of the body. In Farina’s view there is a circular need that presents one of the greatest challenges that plagues our field – we need a greater understanding of the systems so that we can better develop the technology, but we need better technology in order to improve our understanding of the systems.
“It’s probably a combination of mechanistic and basic science understanding that is still lacking,” Farina said. “Of course, if we could do everything technologically, if we could put hundreds of millions of electrodes in any part of the system, we could record from every single cell and we could solve any problem with technology.”
Again, y'all know that I’m not an engineer, so I rely upon those who are to add perspective. During a recent visit to CorTec’s HQ in Freiburg, Germany, I learned that they are working on 32 channels for recording and stimulation and that an increase in the number of channels of their closed-loop system is already on the agenda for the development of the next generation.
"Our Brain Interchange technology is specifically designed to explore the mechanisms of the brain more deeply," said Martin Schüttler, CTO and CEO of CorTec. “We are dedicated to providing researchers with new technological opportunities through our implant system in order to accelerate their discoveries and thus make new therapies possible faster.”
As is the case with so many neurotechnologies, the regulatory framework, which is essential for safety reasons, is unfortunately slowing development progress for devices that will be used in human studies. CorTec is confident that they have already created a good basis, which they are expanding step by step.
Some of my past (and upcoming) interviews delve into the importance of collaboration: Marom Bikson circling back to his Case Western roots, Robert Riener’s focus on collaborating with end users to inform design, Doug Weber and… everyone. As pure chance would have it, Farina and Weber are collaborating together.
Type…type… type… send. Weber responds to my inquiry in typical Weber-fast fashion.
“Here’s the story,” launched Weber. “Battelle developed a sleeve for doing FES in the forearm. When I started working with them, I suggested that we try using the sleeve for HDEMG. Battelle loaned me a sleeve and I assembled a recording system to measure EMG. I have used that system to measure HDEMG in healthy controls and people with stroke and SCI. The paper, presented by Jordyn Ting at EMBC, is our first report of HDEMG from muscles “paralyzed” by SCI. We’re working on the full paper now. I invited Dario to collaborate on this project because he has methods for identifying motor unit action potentials within HDEMG data.”
Farina expanded on the genesis of the project. “The collaboration with Battelle is fairly recent,” he noted. “It came about through a good friend of mine, Doug Weber. Doug visited me in London to give a seminar and he was showing data that they collected with this new device, which is very similar to a lab device, but they developed it for the clinic. In a follow up meeting in my office I asked him to send me the signals because I thought we could do very interesting things with the signals and, indeed, within a few months we did have something very exciting,” he continued. “Now we are writing a grant together with Battelle. These things, when they work well, they can develop very fast. You can get something very exciting going in a very short time.
“I’ve been very, very fortunate to have many collaborations with top people from whom we’ve learned a lot. I have been so fortunate to collaborate with excellent people. For sure there are more out there. My collaborators so far have been such good people and good scientists that I could not ask for more,” Farina concluded.
Farina, who spends about 80% of his time on research (and the correlate administration) is now focused on the next steps for his research: translation. “Research is always, for a big part, incremental,” said Farina as he deftly brings us back to the idea of Baby steps. “We have built on what we have built in the past. We have a lot of projects that are based on our technology. The technology is now mature enough to make the next step so that we have a full translation. We have many pathways in many different patient populations and many different technologies. We need to translate completely. We are in discussions with companies and we are working on the translational level. It’s less time for basic research, but it’s just as important because if you want to do right by patients, you have to take those translational steps.”
I always like to ask the people that I interview what advice they wish they had (or had listened to) as sort of a “Dear Abby of Neurotech” feature. I’ve gotten some interesting answers, but Farina’s seemed particularly compelling as the neurotech field grows in diversity and opportunity. “You have to have a broad vision,” he said. “I remember as a PhD student that I definitely had a very, very narrow vision that I can see now was very naïve. I would definitely go back and say to broaden the vision. If you do it as early possible in your career – then usually your career will do better – not just from the academic side, but from the research that you do. Broadening the vision was something that I missed early on.”
In the words of the inimitable Leo Marvin, “Baby steps. It means setting small, reasonable goals. One day at a time. One tiny step at a time.”
 Kazamel M, Province Warren P. History of electromyography and nerve conduction studies: A tribute to the founding fathers. J Clinical Neurosci 43 (2017) 54-60
 Adrian EDBD. The discharge of impulses in motor nerve fibres: Part II. The frequency of discharge in reflex and voluntary contractions. J Physiol 1929;67:119-51
 Ting J, Vecchio AD, Friedenberg D, Liu M, Schoenewald C, Sarma D, Collinger J, Sharma G, Farina D, Weber DJ. A wearable neural interface for detecting and decoding attempted hand movements in a person with tetraplegia. Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Conference, 2019