Episode 25: Cynthia Chestek, PhD

The following interview was conducted in-class, during the Spring 2024 session of Hidden Figures: Brain Science through Diversity, taught by Dr. Adema Ribic at the University of Virginia. What follows is an edited transcript of the interview, transcribed by Brooke Alexandra Negron, Aidan McQuarrie Fleming, Nicole Grace Garibaldi, Annie Rui Lucie Curran, Faith Taylor Crosley, Lucas C Nelson, Lily Elizabeth Kauffman, Neha Amarnath, Sam Marie Patrick, Neena C Peterson, Adrian Rodriguez, Sharzeeda Mandiya Attah Diallo, Olivia Sophie Pemberton and Victoria Lu, who also drafted Dr. Chestek’s biography. The final editing was by Dr. Adema Ribic.

Dr. Cynthia Chestek is an Associate Chair of Research for Biomedical Engineering in the College of Engineering at University of Michigan where she ​​runs their Cortical Neural Prosthetics Lab, which focuses on brain and nerve control of finger movements, as well as high-density carbon fiber electrode arrays. She is also an associate professor for robotics, biomedical engineering, electrical engineering and computer science. Dr. Chestek received her bachelor's and master's degrees in electrical engineering from Case Western Reserve University, and her doctorate in electrical engineering from Stanford University. Her current research is based on implantation devices within the premotor and motor cortices, working towards allowing paralyzed or amputee patients to move prosthetic devices, as well as their own limbs, using functional electrical stimulation and assistive exoskeletons.

What does your lab currently work on?

My lab does basically 3 things. [We are working on] a nerve-controlled prosthetic hand, where a person has very small electrodes that are implanted in their brain such that they're picking up nerve signals and controlling prosthetic hands. We also work on brain-machine interfaces trying to do able-bodied control of individual fingers. The last thing my lab does is the carbon fiber electrodes which are electrodes that are smaller than neurons. They're so small they can sink into the cortex, doing very little damage and they provoke very little immune response.

How did you get into biomedical engineering?

I started this path freshman year, but in high school, I was a theater kid and was really into music. However, I was good at math, and by the time I was a freshman in college, I realized I should probably major in something that involves math. I actually went into physics first because I thought that was really interesting, and I was reading lots of books about physics. At some point, I realized I wasn't gonna discover the mysteries of the universe, so I switched to electrical engineering, which was just sort of very practical. I'd always been interested in electricity and programming.

I got into the neuroscience because I was looking for a funded summer research position and I could not find very many so I joined a neuroscience lab, where they love electrical engineers. I started recording from neurons and that was amazing. Then I settled on brain machine interfaces as kind of the average of electrical engineering and neuroscience.

I went to Stanford after that and I kind of picked it because I could stay an electrical engineer but do neuroscience. We were interpreting the activity of neurons in real time and back then we were controlling computer cursors on a screen. That was going really, really well: this is a field where we have had large jumps in performance at about these 5 year intervals. At some point, you're faced with the choice: do I stay in academia or go to industry, where industry is the more boring version of whatever I'm doing. In my case, the more boring version of this are wearables like my Fit-bit which is processing electrical signals in real time. However, I could not leave because it was going way too fast, and I really want to be a part of it. I really wanted to keep working on it, so I became a professor in 2012.

Was there a key moment in your life that helped you understand your passion for neuroscience?

There were a couple of key moments, key conversations, that changed the whole course of my career. One of them was meeting with my curriculum advisor. He looked at my grades and he was like, “Okay, well, what are you gonna do?”. And I said, “I don’t know. I'm gonna go work at a company or something.” Then he says, “ Your grades are really good. The bigger things are open to you. You should go look into all the different kinds of research that are going on around the department, talk to the professors, and just see if any of them are interesting to you.” That started a long path that didn't happen quickly, but it was after that conversation that I kind of realized things that I didn't think about before were open to me.

I also had an amazing undergraduate research mentor. This was a neuroscientist and, to this day, this is somebody that always has a lot of undergraduates in the lab and treats them like grad students. He talks about science for long, long hours with all of his students and that was a lot of fun. I really loved being in that lab environment.

I also had an amazing Ph.D. mentor, Krishna Shenoy, who sadly passed away about a year and a half ago. It was really hard to lose him. We had an amazing and important relationship. We had numerous conversations when I was trying to think about whether I wanted to become a professor or not. I can remember sitting in his office going through his calendar trying to figure out what life would be like if I was a professor.

What drew you to originally study math/physics?

As a kid, up till the age of 14, I had been involved in math competitions. Strangely, I met my husband at a math competition when I was 14. I was always competitive and I enjoyed that, so I think I had a good appreciation for how math was different from my other skill sets. Math is very deep. Whereas a lot of the other skill sets are broad, math is something that very quickly seems like magic if you see something that's even 1 or 2 semesters ahead of you. It's just very, very valuable, and more math makes everything better.

Going through college did you see yourself pursuing research?

Certainly not at first. It originally just didn't even occur to me. I was from a smaller town, Erie, Pennsylvania. I was not thinking at all in terms of doing research or anything like that, until I had that one conversation with my curriculum advisor.

What motivates you to do your job?

This job feels very meaningful. In the sense that your job is to move the whole world, but you don't get to move it very much. The unit of one academic paper generally doesn't cause revolutions, but it is the only way that human knowledge advances. And if you look back, you know hundreds or thousands of years, it's the most important thing that's been happening. I really do feel a part of that: when we do something cool and we put it out in the world, we watch the world change right through it. Other groups that do similar things start adopting the technique. Sometimes we get to see industry adopt a technique. It feels very much like academic engineering research is raising the floor under everything humans can do. Then companies try to jump off that floor and grab onto new things first, but raising the floor feels really meaningful. In that way, it would be very, very hard to go and just work for one company instead of working for the whole world.

How difficult is it to engineer systems that mimic the brain and what mechanisms go into creating an algorithm for controlling prosthetics?

It’s interesting because it ends up being much more machine learning than neuroscience. We're using the best possible algorithms. There's some really cool research on using natural plasticity to control brain machine interfaces. My lab actually doesn't do very much of that, we're just reading machine learning literature. We are trying to deploy the best algorithms and get the highest performance. That said, recently, virtually all of the high performance approaches involve neural networks. It's such a privilege in my lab; every day we have a population of neurons, and the students are hooking up a population of artificial neurons and interpreting it and putting those two things together. I really hope that within the next 10 years we start to be able to draw more parallels between the best of artificial neural networks and what's actually going on in the brain.

You have highlighted the superior performance of long short term memory algorithms compared to other decoding architectures. Could you elaborate on the advantages of long short term memory algorithms and tell us how they contribute to achieving more accurate neural networks?

For that paper, we were looking at a lot of different neural network architectures that had been tried without trying them in real time (actually controlling things and sort of putting them all together). It absolutely was the highest performing approach, but it was a little bit higher than the other neural network approaches. All of those are still in use (LSTMs) and we are also using a fair number of fully random recurrent neural networks that are working pretty well. The feed forward networks are good for a lot of things. They're very simple. They're very easy to put on a microcontroller or things like that. I would say the biggest benefit of the recurrent neural network architectures is that they can memorize dynamics, meaning they can memorize our task right.

Have you ever had trouble securing any grants and how do you face this obstacle?

That’s hilarious, because even if you're really really good at getting grants, you're never gonna get more than one in three. At first the job starts with writing ten and getting one of them. I would say that your grit is critical. It's the non optional part of research: research is trying things that nine out of ten times don't work, but when one works it's like a step forward for the entire human species. You get a lot every time you have success and that's true both in the lab and in grants. It teaches perseverance and it teaches emotional fortitude. I have to say, I just got a big grant this morning. I was very excited. I was stopping myself from opening the reviews because I put my heart and soul and many hours of work into it. I know that two out of three times I open those reviews, it's gonna be like being punched in the face.

What kind of innovations within neuroscience and biomedical engineering do you hope to witness in the following years? Could you reflect on any notable developments you've observed in your time in the field?

In 2002, brain machine interfaces didn't even have any real tools. It was people assembling wires to try to make electrode arrays. So by the time I got involved there were a couple of really good devices, work horses that people can use, and multi electrode arrays. These (array) electrodes can't really live in harmony with neurons forever, so you could do it for the length of an experiment, maybe for a couple of years. A lot of electrodes now don't cause that kind of scarring. It’s gonna take a while for those to make it out to a medical device. The current fiber electrodes that we use, it's only one of a few electrodes that are similar to a neuron in size, like capillary size. Besides the size, we have to figure out other problems like how to get them into the brain. We have to make them not get attacked by the body. There are lots of engineering challenges. I have been in this field for 15 years and you see these dominoes fall, and there's not very many dominoes. What's interesting to me is that things are actually much better off today in my field than I thought they were going to be as a PhD student. It's because there's good luck, too, and it's hard to predict the good luck, if you have bad luck all the time. Occasionally something happens which is better than you think. You can't really predict that. Things go quickly in where the field is at right now. They're decoding 70 words per minute from the brain activity of people who can't talk. I just heard the first system went home with somebody-it’s a research system, but they can use it to speak to their family.

Given your substantial research on implantation devices and knowing how much they can impact a person's life, are there any specific milestones that you'd hope to reach in your field, perhaps within the next five to ten years?

We still have not had any of these systems put in as a clinical treatment. There's very cool videos of people controlling prosthetic hands, but all of those were done as part of a research study. It's gonna be a major milestone when somebody goes home with something that's really supposed to help them. That's also a really important thing to note about our research participants. We are not helping them yet. They are helping the field, and they are helping people that come after them. There's no guarantee that they're gonna benefit from these studies, so it's a real privilege to work with them. At some point in the future, I can't wait until we see somebody take home a system that they can actually use.

We already talked about difficulties with getting grants, but were there any other setbacks towards any goals in your career? Specifically, did you experience sexism in the field of engineering which is a men-dominated field.

The dynamics are interesting. I would say that I can give you a couple of instances of sexism, but I think that's mostly not the reason why there's fewer women. Overt sexism is a rare thing, but the dynamic can be weird. Electrical engineering has a culture. When I was in undergrad, the physics department was issuing invitations to join them, but the electrical engineering department was not and that's because with the electrical engineering department, everybody has amazing jobs with super high salaries. It doesn't need to be advertised. Anybody who's in the “in group” knows why you need these skills and you want this degree. There's no welcoming committee and there's also a heavily competitive culture. If you have the smallest amount of cultural difference in who feels comfortable bragging, that has a major outsized effect. In our culture, if men feel 5% more comfortable telling you how great they are, and women tend to put themselves down a little bit. This gets amplified like crazy when you're actually taking these classes. I'm just glad that I stuck with it for long enough to realize that the guys were all talk. Those people that were talking about a big game were not getting A's. I was getting A’s, and I never once felt good at it. I just did everything; I did all the work, did all the studying for all my tests, I got good grades. It took at least a year to realize what the disconnect was there. After that point, it’s easy. You start judging yourself by like real numbers, real successes as opposed to how you feel you are doing. I think that was the reason I saw a lot of people dropping out of undergrad.

Grad school is different. Grad school is when you stop learning from classes and you start learning from other people. If you feel socially intimidated out of doing that, you lose all that opportunity for learning. I never collaborated with another woman until I became a professor. My colleagues in grad school were all men. They were super nice.

Why do you think so few women are attracted to math, physics, engineering, and science?

I think there's a couple of reasons for that. One is that it's insider information that these careers are so valuable. It's not obvious at all: you need somebody to tell you that this is the major that you have to have in order to do these things. Second, I think it is very fair to look at a group that is very homogeneous and conclude that there must be something dangerous, something that's scary happening there. I like that this has not been my experience–engineers are just about the least intimidating men in the world. But I can understand how somebody can think that from the outside, and I also understand how a small number of jerks can make things seem worse than they are.

At UVA, there’s a constant competition to be the best and be the top of your class. So, therefore, in your undergrad, how did you enjoy what you were doing and continue on this path, if you didn't think that you were good enough or good at all?

Enjoy is kind of a strong word for engineering. I love what I can do. I've found it hard to explain to students that it's what you can do later. That is what you're doing this for. There's just a lot more delayed gratification in engineering. I actually feel like engineering is like if you want to get a major in cooking rather than eating. Eating is the skill of enjoyment. But if you learn to cook, you're gonna eat. You really do have to know why you want it and what you want to do afterwards. I got a great deal of enjoyment out of interacting with other people. There's something that makes humans happy about doing hard things in groups. I feel that still to this day, if my group, if my students are working in a team of three or four people. It's much easier to get really hard things done than when you're working alone.

Would your advice to undergraduate students be to find those people?

Yes, you have to find people. This is important. How to join a group, whether that's a lab or a study group. I always tell people to join a lab, become helpful first. You're really just trying to get to a point where you feel you can ask other people for help. You can always just go get help. You're feeling self-conscious about asking for help all the time, and you're gonna need to ask for help all the time, you can be a helpful person. There's always ways of being helpful, like when somebody new joins a lab. There's always something that somebody needs help with. If you find those things first, then you know you're part of the economy of helpfulness and you start learning from other people.

How can a background in physics and electrical engineering contribute to the study of biomedical engineering or other disciplines?

That's the beauty of both physics and electrical engineering, and I would put computer science and mechanical engineering in this bucket. It's easy to go from those disciplines to other disciplines. If you're not sure what you want to do with your life, I'd recommend one of those things because you can branch into anything. I really do believe biomedical engineering is an amazing major if you know exactly what you want to do. If I want to do prosthetics or biomechanics or something like that, you can get more specialized by being in a biomedical engineering department.

What is one major challenge you’ve encountered in your career and how did you overcome it?

The obstacle of switching into a teaching role is a big one. When you finish your PhD, the best scientists in the world are post docs. Those are the people that are in the lab all the time. They know how to do everything, and they can do things so quickly and so well. Then, becoming a professor and becoming a manager like in any job [means] you're no longer allowed to achieve things with your own hands.

It’s extremely difficult. That switch was important to me. I got past it by doing things with people until I understood why they couldn’t do them right. It’s really easy to forget what you learned. I had somebody graduate from the lab and start a lab. They're really struggling with this; they just want to do it themselves. My husband was recently pushed into a management role at his company and he was texting me, “Okay, can I just do it myself?” I said, “No, that is not your job anymore.” You have to find ways to meet people where they're at. You learn that if people aren't doing something, it's usually because they don't know how, and they don't know how to tell you that. The only way to do that is often to walk with them to where they are attempting to do the thing, make them attempt it, and then you can help them. That's the moment you can say, “It is okay you don't know how to do this, now I can help you.”

What has been your most meaningful collaboration or work experience?

One of my best memories is when we had this insane DARPA deadline [in graduate school]. DARPA is the most unforgiving funding agency, and we had multiple program officers visiting the next day. We were trying to meet this milestone that the whole project relied on and we were working all night in the lab. I remember going out for pizza at 2 in the morning with my lab mates and just talking about circuits, our experiments, our neural signals. It was hard. It took a lot of hours. But it was a lot of fun. Of course, it worked–we got it all to work. We kept our grant and it was a very rewarding experience at the end.

What is Neuralink and what is your relation to the company?

Neuralink is a company that has consulting arrangements with various places. I actually have royalties from Neuralink, so I do have a conflict of interest. They license all of the brain machine interface patents out of Stanford, so I get a few 100 bucks a year from Neuralink. There's a number of new things that are all being done at the same time in the Neuralink device. No one, to my knowledge, outside of that company really knows exactly how those signals are looking.

They just did their first human trial: their first human implant. What we know from that is that they have gone through many rounds of back and forth with the FDA. They're only working right now with people that have severe paralysis.

How have you seen your field change and where does Neuralink fit in?

My field didn't exist at all 20 years ago. The first real papers with monkeys controlling computer cursors were from 2002. And that was the first time that it was put together. Fast forward 5 years, and it was becoming an engineering discipline rather than neuroscience. Many more people getting involved are focusing more on the machine, and learning the signal processing and control. Probably 10 years after the field was founded, we started to look actually good.

Computer cursor control was working at that point, and that was when I think translation really began. The first human results were very early, but better human experiments started happening, and then 5 years ago there started to be a real commercial push. We haven’t really seen the companies yet, but there are competitors to Neuralink: there’s Paradromics, Blackrock, and others.

Are there any unexpected skills you’ve gained through neuroscience and engineering?

I think of myself as speaking multiple languages because I interact with neuroscience. I interact with electrical engineering, and with biomedical engineering which is in between the two. Neuroscientists are very nice, don't over brag, don't put people down. Then I'll find myself in a room full of electrical engineers and we'll start quantifying people. I'll just feel like, “Oh, yeah, I speak this language, too.” But I know not to get offended. I also call it speaking DARPA. If you go to DARPA meetings everybody just walks around talking about how great they are, and at some point I had to pick up that skill set. It's very unnatural for me, but I learned how to speak it. A part of me thought that if my husband ever heard me at a DARPA meeting speaking about how great I am that I would be embarrassed because that’s not who I am.

Is there currently a specific project that you're working on, if you're willing to share, and are there any setbacks? What do you think the results of the project will be?

We're working on our neural dust project. It's been way slower going than I'd hoped. In this project, we take our 8 micron carbon fiber, and we have 2 chips that we put on top of them. The sad truth is that lots of research projects stall out on some dumb thing. We have all of our chips, but we can't get the 2 chips to go together in a perfect alignment. We're now at the point where we're trying it 3 different ways, and we've been trying it for a couple of years. Things are just not working, and we're discovering things that have to go. But everything almost works until it works. We're hoping at some point, we'll be able to write a big paper on carbon fibers with basically 200 micron “ice cubes” attached that can do a wireless neural recording. That's important because only this capillary-sized electrode is entering the brain.

How often do you get to work with real patients? Do you work with MDs often?

With patients, I meet our study participants. I try to come to the initial meetings, but it's my students that really get to know our study participants. The vast majority of the world's scientists, particularly in engineering where there's fewer post docs, are Ph.D. students. They see everything first. Of course, I get to hear about it at the weekly meeting, but they're the ones that actually do it. Most breakthroughs in engineering that I've been a part of or been aware of have come from Ph.D. students. They get to interact with people a lot. At this point, I have a bigger lab and I don't get to go to experiments very often. It's a little sad for me, but there's something very enabling about enabling other people. I do interact with our participants, but not as much as my students.

I interact with MDs all the time. This is the kind of research that would not happen at all without very interested MDs. These are, I would say, rare people. It takes a lot of courage to innovate when you're a medical doctor. They have to be very involved and very engaged; they have to make it safe. I'm an electrical engineer. I can't do surgery. I can't judge the safety of anything that we do. That’s why I have a number of different collaborators. Pretty much everything I work on, maybe some more than others, like the rodent research, involves an MD pretty heavily.

You fulfill a lot of roles at the University of Michigan: professor, head researcher, associate chair. Which would you say is your favorite role to take on?

Mentoring PhD students. That's easy. They're great. That would be the part of this job that I could not possibly walk away from. Teaching undergrads, it's rewarding in a different way. But teaching Ph.D. students is amazing.

What is it like getting a PhD? What is the biggest challenge for PhD students, since it is so long and intense?

A PhD is having the skill set to do a peer reviewed publication. That's true across all the disciplines. I once sat down and calculated that it takes my students an average of 4 years to produce their first paper. Then, I checked that for my advisor, because I wondered if it was high for my lab, and his was 4.5 years. Then I sat down with all my students. We made a list of 30 reasons they take that long, and it's mostly not a lack of technical skill. Yes, you have to learn some things, but it is mostly learning how to be focused and rigorous and direct everything you're doing towards answering a question, rather than doing a cool thing. There's a realization that you can do an infinite amount of experiments that teach you nothing. Ultimately, you get into this very rigorous mindset: there is a thing I need to know next, and this is my precise question, and I need data that addresses this precise question for this precise figure.

How important are grades in a PhD application? Are there specific traits or characteristics that stand out to you that your PhD students possess? Do you find that grit and passion for research is the number one thing that makes a good PhD student or are there other factors?

That's a hard question to answer. First of all, grades are important. I will sort the applications by grades. We admit students that more or less succeeded at undergrad. But grades certainly aren’t the most important thing, and it doesn't predict any performance after that. I would say grit is important. Having spent some time in industry is great, because they realize what the deal is and the trade they are making to transition to academia (lower salaries, but you get to work on cutting edge research).

I do think that not giving up is really important, and not being intimidated by not knowing how to do things. If that makes sense it almost doesn't matter what your starting skill is. If you care about looking smart, you're not going to learn anything. If you are open to asking lots of questions, and not concluding that you're just bad at something, that sort of growth mindset is pretty critical. Especially when you're supposed to learn in a really unstructured way for 5 years. I do think all the scaffolding of the classes comes away, and you have to bring about that learning through interacting with others, asking questions, trying and failing, trying again.

What are your hobbies and interests outside of work? Do you think these hobbies relate to your work at all?

I actually have a lot of hobbies and I've had a lot of hobbies over the years. When my kids were little, I had no hobbies. All the hobbies just went away for a bit. Honestly looking back, it was not that much of my life. My hobbies have always involved making things. I used to do a lot of medieval reenactments. That was great, because you just get to make everything. I spent 2 years getting obsessed with knitting, and then lately I've been obsessed with gardening. I play a lot of board games. A lot of gaming in general. We took our whole family to Gencon this past year, which was fun.

If you've had an annoying day at work, what is one thing that you like to do that helps you blow off steam or relax?

I think Covid made me an outdoors person, because the only place I had to go was my backyard around the block, and that gave me really good habits for taking a walk, getting under sunshine, all of those things. I tend to take a walk, even if I'm indoors in the middle of winter. I'll walk around my complex or if I can, I will get outside and just be outside doing things. I also talk to my husband. My husband is extremely supportive; he's on the receiving end of learning everything that isn't going well at work. It's very important to have a really supportive partner.

Going off what you just said about the very technical and detailed nature of your work. Do you find that it's easy to share with your family and friends and your husband, or do you find you have to explain yourself a lot?

I think I can talk to people about what I do at all ranges. I'm a teacher, so I'm good at explaining what I do to a 12 year old. I do talk a lot with my family about what I do. I will say, I think I do have a bit of a cultural disconnect with my hometown where it is hard to relate to

them. It can be hard to talk about what I do, and it's unclear why I left my hometown to a lot of people. Otherwise, I think I definitely can. I also love university environments and I've enjoyed being on campuses. That's part of why I became a professor. A lot of people buy into progress as a life goal.

What advice would you give to a student interested in pursuing engineering and an undergraduate student in general?

To an undergraduate student, I would say–and just in general for engineering–get your feet under you first. I remember the transition to freshman year [of college], which was the hardest thing I have ever done. All the things that came after giving birth are not nearly as hard as freshman year of college, because you make that transition to self motivated work.

What guidance would you give someone who’s apprehensive about switching fields like you did?

I would say that a different couple of classes takes you a really long way. You can decide, “Hey, I want to get a little deeper in the math sequence” or “I want to take an extra programming course,” or something like that. Those things help a lot. My husband is a self taught programmer; he was actually a psychology major, but he had taken a programming course. You can end up with one programming course having more economic value than the rest of a psychology degree, so he ended up doing that and he got into tech support.

I remember I had to start college at calculus one, which felt like the end of the world, because I was surrounded by people from great prep schools. They all had AP calculus and I felt like I was way behind. Now I see the absurdity of an 18 year old thinking that they're way behind on something because they're 4 months behind on something. You have all the time in the world. You're so young. You can make these kinds of changes. A year is not the end of the world.