Stephanie Palmer: A (real) theoretical neuroscientist


Stephanie Palmer, PhD, assistant professor of organismal biology and anatomy, shows middle school students a real brain during her neuroscience workshop.

As a chemical physics major in college and Rhodes Scholar in theoretical physics at Oxford, Stephanie Palmer, PhD, assistant professor of organismal biology and anatomy, didn’t exactly seem destined for a career in neuroscience. But though an encounter with some friends, a unique fellowship in theoretical neurobiology and a lot of hard work, Palmer now finds herself working to shape our understanding of how brains recognize and predict things about the world around us.

If equations, experiments and developing new theories about how the brain works weren’t enough, Palmer and her team also run workshops that introduce middle school students to the not-so-scary world of neuroscience (and yes, that is a dissected cockroach leg dancing to the sweet beats of Mark Ronson and Bruno Mars).

ScienceLife spoke with Palmer about life as a (real) theoretical neuroscientist:

What is theoretical neuroscience?

Stephanie Palmer: The task of theoretical neuroscience is to make contact between the really big questions about how our brains work and the kinds of data we can collect in the lab. Our job is try to make theories of things like neural computation. How does a stimulus, like something you see, go through multiple stages and layers in the brain, through millions and millions of neurons and synapses, to become some output behavior? How do you store memories in networks of neurons, how do you retrieve them, how do you erase them? All that kind of stuff is the domain of theoretical neuroscience.

While we call ourselves a theory group, we collaborate really closely with experimental groups. We also do a little bit of experiment here because I guess I’m a little crazy and my interests are broad. We even have a butterfly vision project going.

So what’s the big question about the brain you’re trying to answer?

SP: We want to unravel how the brain does computation. We’re focusing on one important computation that we think is really important in lots of brain areas on lots of different information processing levels, and that’s prediction. To do so, we’re looking at how our visual system represents what will happen next in the visual world in the retina.

Why the retina?

The neural computations needed for prediction begin in the retina.

The neural computations needed for prediction begin in the retina.

SP: The simple picture of the retina you might have is that photo receptors are like pixels on a CCD camera that relay frames of images of the world to your central brain. But the retina is actually part of your central brain that happens to be sitting out in your eye – and it does a lot of computation.

The hypothesis is that if we understood that computation better, we could do a better job at building retinal prosthetics. This requires that we understand the computational transforms that are happening. We have to essentially speak the language of the central brain. Prediction is a very, very important component of the computation that’s going on in the retina, so that’s what we’re trying to understand.

Wait, the retina computes? And predicts?

SP: Yes! Part of the computation for prediction takes place in the retina. How do you take in raw information from the outside world – all the different light levels that are hitting your photoreceptors – and from this input derive where a moving object will be next? We’ve shown that the retina represents information about the future of the visual inputs and extrapolates forward optimally. What seems to occur is that groups of cells in the retina have some hard-coded way, something about their tuning in a dynamic world, which represents predictive information preferentially over information that isn’t predictive.

To give a concrete example: You’re sitting in a park and you see a ball coming at your face. What is your brain going to represent? The ball trajectory, position and velocity, or is it going to represent the exact texture on a fluttering leaf that’s in the background to the side? How does it decide? All this information could be important. What we found is that the retina throws away stuff that’s irrelevant for prediction, throws away things that are noisy, throws away things that are static background, and keeps the information that’ll help the brain figure out where things will be next.

Is this how baseball players are able to hit a 100mph fastball?

Prince Fielder Home Run Swing - ImgurSP: It’s part of the answer. There’s a pretty long delay between the eye and the brain, and then there’s motor delays too. So the loop from when you see something to when you grab it could be something from 50-200 milliseconds. Professional baseball and tennis and cricket players, if you look at where their eyes are when they’re about to hit a ball, it’s not on the ball – if it were on the ball then they’d already be overcoming the delay since they’ve put their eyes on where the ball is now. They’re doing even better than that, they’re putting their eyes where the ball will be.

Even something not moving as fast as that, say like 60mph, is going to have moved several feet in that 50ms time window. We couldn’t solve that problem if we didn’t overcome this processing delay. It’d be like trying to drive a car if the front windshield were blacked out and you could only look through the rearview mirror.

Are you interested in studying athletes?

SP: We’re actually more interested in regular human performance. What we’re most excited about are things we can’t predict. For example, I have a student who has this nebulous goal of making a Turing test for natural motion. That sounds a little crazy so let me explain.

Reduced to a few points of light, we can easily tell a the difference between a man and a woman walking. Palmer wants to know what's needed to tell natural motion apart from background motion.

Reduced to a few points of light, we can easily tell a the difference between a man and a woman walking. Palmer wants to know what’s needed to tell natural motion apart from background motion.

If I showed you a video of butterflies flying around, then replaced the butterflies with dots, and said, ‘Now I’m going to show you lots of different sequences of moving dots – tell me which ones were taken from markers on natural objects and which ones are not.’ This would reveal what naturalistic motions look like when reduced to a few markers.

Coming from my physics background, we’re trying to understand the equations that govern this kind of natural motion as projected onto our retina. As opposed to the kind of stochastic flow you get from turbulence driven motion – leaves rustling, grasses swaying, water lapping, etc. We want to understand how the eye tells textural background apart from something important, like ball coming at you or a bird as soon as it moves in a tree. That’s something that machine vision can’t do yet. We actually just completed our first major effort in the lab, a two terabyte database of natural motion from in and around the Chicago area, and we’re going to publish a paper on it as a methods paper and a follow up analysis of the motion.

You were a Rhodes Scholar in theoretical physics at Oxford right?

SP: I started off when I was a young wanting to be a brain surgeon but I become totally engrossed by my physics and math classes and gradually moved away from medicine as an undergrad at Michigan State. My grandma (Mary Jane Palmer, née Morrison) actually went to UChicago back in the day. She did physics here and graduated in 1941. She and my dad are both physicists, so I always thought it was an interesting field. When I went off to grad school I started in condensed matter theory. And I was looking at ordering transitions in, what you could call, somewhat exotic magnets.

What kind of theoretical physics did you do?

[ScienceLife editor’s note: rather than attempt to transcribe a conversation about theoretical physics, we thought it’d be easier to just have Dr. Palmer explain it on video!]

Wow. How did you go from that to neuroscience?

SP: It just seemed like fun [laughs]

The theory that I was working on was well developed and we figured out this neat thing, but I didn’t see this huge, long term career path that I was as happy about in terms of the breadth of new open questions. I had some friends at Oxford who were working in a neuro lab – they were working on memory in the hippocampus. I would talk to them about math-y things and they would show me their experiments. I just got fascinated by it and was sold.

When looked for postdocs, I applied for this fantastic Sloan-Schwartz fellowship at UCSF where they took mathematicians and theoretical physicists and brought you into neuroscience with funding to kind of just learn neuroscience. I fell in love with these questions about what happens to our brain as we age that limits our ability to learn new things, critical periods for language learning and things like that. These questions remain largely unanswered, and they’re huge fundamental questions about how our brain works.

What was the transition like?

SP: It was kind of like starting your PhD all over again, but you’re a postdoc so there’s this tension where you feel like you don’t know anything but you’re supposed to. It was a bit of a shock to the system, but I got my PhD right when I turned 25, so I felt like I had time. I gave myself some years to flail around a little, just take in lots of information, before I wrote my first paper in neuroscience.


You run a neuroscience workshop for middle school students called Brains! How did it get started?

SP: I’ve always taught. I taught in chemistry labs when I was an undergrad, I taught electromagnetism in grad school and I did outreach teaching with local middle schools in San Francisco. I heard about these guys from Backyard Brains who had these wonderful little handheld amplifiers. You can cheaply solder and make these spike amplifiers yourself, so that students can see and hear neural spikes from, say, cockroach legs or their own arms. I thought that was a really cool way of demystifying science. I wanted to do something with these Backyard Brains kits so kids could have a hands on experience.

I happened to sit next to Sara Stoelinga, director of the Urban Education Institute, at a committee meeting one day and I told her about this idea. She hooked me up with LaVerne Wright, who was then at Woodlawn Middle School. I talked to LaVerne, she said great, let’s do it now. I talked to her in the winter and we had our first workshop that next spring. I got some graduate student volunteers together and we spent a few nights soldering up kits and ordering cockroaches and figuring out how to keep them in the lab. Since then, I’ve just followed LaVerne, so whatever school she’s at, we’ve gone and coordinated. This is the third year we’ve run the class.

What do students get out of the workshop?

SP: How do you look at the brain and see how it works? It’s not like the heart, which is obviously a pump. The brain is just this mush, this lump of tissue. How do you get kids to have a tangible experience with neurons? The workshop lets kids do the experiments themselves and see and hear neural spikes for themselves in real time, and not through some super expensive piece of equipment.

The brain is normally really abstract since its workings are electrical and chemical. But with the kits and an iPhone, they can see the spikes as a cockroach leg twitches, or as their own muscles move. They can also stimulate the cockroach legs. You can actually play a video from YouTube, feed it into the amplifier and make the cockroach legs dance.

[ScienceLife editor’s note: This is super cool!!!]

We also do this great experiment where they all hold hands and record how long it takes for squeeze to make its way around the circle. They can then measure and add up the lengths of all their arms to estimate the speed of a neural impulse.

The kids get to put on gloves and hold a brain from the anatomy lab. That’s something I didn’t get to do until my 20’s. It’s pretty profound. It’s also very personal, since you’re trying to understand your own body. I think the connection is pretty easy to convince students to get excited about. I think it’s really important for kids to have a firsthand experience with science, and especially neuroscience, which you don’t usually interact with until maybe college or even after.

This slideshow requires JavaScript.

Any plans to expand?

SP: We’ve done this every spring now, we’ve had 40-70 kids come through every time in groups of about 20. But we can’t do it every week – between the setup and take down, we have retool the whole lab so the kids can come through.

What I’d really like to do in the future is to run teacher training sessions. The first couple of times I funded it with my own startup funds. This year we were lucky enough to get funding from the Grossman Institute and from my department, Organismal Biology and Anatomy. I’m writing it into some grants, and the idea is if we got funded, we could have teachers come in. They could solder up their own set of kits for their science labs back at their schools, and we could give them worksheets on how to use the equipment, how to keep the cockroaches alive, etc.

That way, you have something embedded in the curriculum of local schools that would be more long lasting, and we could come in for guest lectures with kids who already have experience in neuroscience. I’m hoping that the word spreads. It’s clear that grad students get a lot of out of it, the students and teachers love it, and I’m hoping we can make it so that classrooms around the Southside can participate.

What’s the most important thing you want students to know about the brain?

SP: There are tangible factual things, like neurons communicate via electrical signaling. You know your brain controls everything, and it does so with these electrical signals that you can see and hear and they’re real. If that’s what they take away, great.

But I guess more abstractly, it’s just the feeling that things that sound intimidating or complex can be really easy to understand if they’re presented in this other context. It’s demystifying, it’s less scary. Neuroscience is a big word. It seems like a scary word, but I’d love it for kids to think ‘well it’s not actually that hard.’ I want the kids to come away from it feeling like science is interesting and fun and accessible and not intimidating and just white lab coats and stuff.

What interests you most about the brain?

SP: I feel really frustrated that your brain changes as you age. If you’re past the age of 12 you can’t learn a foreign language without an accent without severe training or innate talent for this kind of stuff that most people don’t have. That’s really the big question I find most fascinating about the brain. There’s this window of time where you can take in all this information, and then it kind of closes. I’d really love to understand how you can crack that and open us all up to learning violin, French, whatever, whenever we want.

About Kevin Jiang (147 Articles)
Kevin Jiang is a Science Writer and Media Relations Specialist at the University of Chicago Medicine. He focuses on neuroscience and neurosurgery, orthopedics, psychology, genetics, biology, evolution, biomedical and basic science research.
%d bloggers like this: