The Fellowship of Puzzlemakers

The Fellowship of Puzzlemakers, by Samuel Burr.

This week I was invited by my friend Graeme Harper, Dean of Oakland University’s Honors College, to talk to his freshmen. The Honors College students are smart and interesting, but most are not studying science. So, I wanted to give them a taste of what a scientist’s life is like, but I didn’t want to give a very technical talk. Also, each fall the freshman class reads a book that becomes their theme for the year, and this year the book was The Fellowship of Puzzlemakers, by Samuel Burr. I tried to relate my talk to the book, at least loosely. I ended up discussing a paper I wrote thirty years ago: Roth BJ (1995) “A Mathematical Model of Make and Break Electrical Stimulation of Cardiac Tissue by a Unipolar Anode or Cathode,” IEEE Transactions on Biomedical Engineering, Volume 42, Pages 1174-1184. Russ Hobbie and I discussed make and break stimulation in Homework Problem 48 in Chapter 8 of Intermediate Physics for Medicine and Biology.

There are a lot of freshmen (over five hundred), so I talked to two sections of HC 1000, one on Wednesday afternoon and the other on Thursday morning. The enrollment is so large that we had to meet in the Oakland Center Founders Ballroom. It may be the largest group I’ve ever addressed. As I expected, the students were lively and engaged, and asked many good questions. Below is an approximate transcript of my talk. I had no Powerpoint to go with it, so the transcript gives you everything except some hand-waving.  

Introduction

How did you enjoy The Fellowship of Puzzlemakers? I thought it was good. Scientists solve puzzles, too. We’re a lot like the folks in the book. Today I’m gonna give you an example of solving a scientific puzzle. This is a story from my own personal experience. Solving puzzles for a living is great fun.

You know how the book would shift from past to present, back and forth, so it was hard to keep track of when something was happening? I’m gonna do that in this talk. Beware.

NIH

For seven years, from 1988 to 1995, I worked at the National Institutes of Health. NIH is like a big college campus with faculty but no students (what could be nicer?….no, just kidding, we love you students). It’s in Bethesda, Maryland which is just north of Washington, DC. As they say, it’s “inside the beltway.” If you know DC, it’s along the subway system’s red line. It’s so big it has its own stop.

NIH is the greatest biomedical research institution anywhere, ever. At least it used to be. Don’t get me started about what’s happened in the last two years. In any case, during the 1990s it was the best place to do biomedical research.

It’s a little surprising that I was at NIH, because I’m a physicist. I was a graduate student at Vanderbilt University, studying physics, and I especially liked electricity and magnetism. I’m a theoretical physicist, which means I use math and computer simulations to study problems. I can’t solve a puzzle definitively using these tools, but I can make predictions that experimentalists can then test. And, mathematical models can provide much insight into what is happening in a messy experiment. Finally, I also was interested in how physics applies to medicine and biology. Things like how nerves and muscle work. So, I ended up at NIH as a theoretical physicist interested in bioelectricity. One of the projects I was working on was electrical stimulation of the heart. This occurs in pacemakers, where a pulse of current is passed through an electrode to stimulate the heart.

Dekker

Sometimes at NIH I would get tired by the end of the day, and I would stop working on my computer calculations (all that sitting and typing can be exhausting) and would instead go browse in the NIH library. Not the National Library of Medicine, also on the NIH campus, because it has closed stacks. I preferred the more user-friendly NIH library in the basement of the Clinical Center. One day I stumbled upon this paper by an old Dutch doctor by the name of Egbart Dekker. Back in 1970 Dekker had done experiments where he found four different ways to stimulate the heart. He could make his electrode negative (a cathode) and the heart would be excited right after he turned the current on. He called this “cathode make” excitation (make because it happened when the current turned on). He could switch his polarity so the electrode was positive (an anode), and excite by a mechanism he called “anode make.” So both a cathode and anode could excite the heart, but the anode needed a bigger current.

He also saw break excitation. To understand this, I have to tell you a little about action potentials. At rest, the inside of a heart cell is negative (it has a net negative charge). When an action potential fires, sodium channels in the cell membrane open and positively-charged sodium ions rush in through these channels and the inside becomes positive. The inside then remains positive for a while, until potassium channels open, positively charged potassium ions rush out, and the inside returns to rest (negative, where it started). That’s the action potential. If you fire an action potential, and then try to fire a second one soon after, it doesn’t work. The cell is “refractory.” You have to wait a while until the cell returns to rest before you can excite it again.  

When the heart was refractory, Dekker applied a long cathodal stimulus. When the stimulus turned on nothing happened because the cells under the electrode were refractory. But when he turned the stimulus off, the heart tissue under the cathode fired an action potential. That’s surprising, because you usually don’t excite the heart by turning a stimulus off. He called that “cathode break” excitation. Break, because it didn’t happen until he turned the stimulus off. He could also do break excitation with an anode: “anode break” excitation.

As I read Dekker’s article, I thought “that’s odd.” Some people think science proceeds through great “Eureka!” moments when a scientist sees the answer in a flash of insight. But more often, science proceeds by someone saying “Hmmmm…. That’s funny. What’s going on?” 

Now I had my puzzle: why are there four mechanisms of excitation of the heart? What’s happening during each one? Dekker had been my puzzlemaker. He left the puzzle for me to solve. It’s almost as if he had put the puzzle in a bottle, corked it, and thrown it into the sea, and 20 years later I had stumbled upon it when walking on the beach.

Grad School

Now, to explain how I solved this puzzle, I have to go back to when I was in graduate school. I was studying along with my PhD advisor John Wikswo and his research associate Nestor Sepulveda. Nestor was a better programmer than me, and he knew how to use this fancy new technique called finite elements. He had been using this method to do simulations of the voltage created when you use a small electrode in a sheet of heart tissue. The tissue was anisotropic, meaning it had different electrical properties along the grain of the muscle fibers than across them. Just like a pork chop is easier to cut in one direction than the other, the electrical properties are different in the two directions. Nestor found that when you apply a cathodal stimulus, the voltage across the cell membrane right under the electrode is made positive by the stimulus: depolarization. (Remember, at rest the tissue is negative inside, so making it less negative means depolarization). He also predicted two weaker regions of negative voltage across the membrane (hyperpolarization, more negative than rest) on each side of the electrode along the fiber direction. So we were able to see where the tissue was depolarized and hyperpolarized. However, this model did not have ion channels in the membrane that could open and close, so it couldn’t predict action potentials.  

At NIH, I decided to use Nestor’s model, but I added ion channels so I could look at action potentials firing and refractory periods. I used the only ion channel model I knew: one for a nerve axon (some of you might have heard of it: the Hodgkin and Huxley model). I wrote the program and started applying stimuli. And guess what? I was able to simulate all four types of excitation: cathode make, anode make, cathode break, and anode break! I was very pleased, so I wrote up a paper and sent it in for publication.

Peer Review

When you submit a paper to a journal, it undergoes peer review. The journal editor sends copies of the paper to several scientists (you don’t know who) to read it, make comments on it, and recommend if it should be accepted, revised, or rejected. My reviews came back: rejection! The reviewers said that the idea was interesting, but the model of ion channels was wrong. I shouldn’t use a nerve model for the heart.

Peer review is important, but it’s not fun. Sometimes when I get back reviewers’ comments, I want to go play racquetball. But when I hit the ball, instead of thinking it’s a ball I pretend its the reviewer’s head. Once I calmed down, however, I said to myself “you know, Brad, why would anyone use a nerve model for the heart?” Those reviewers had a good point. I needed a heart model for the ion channels. So I found one, rewrote the computer program, and redid all the calculations. The results were as before—I got all four types of excitation—but now they looked even more like what Dekker saw.

Experiment  

I revised my paper and sent it to a different journal. I also sent a copy to John Wikswo, my former PhD advisor. Wikswo is an experimentalist, and his lab had just the technique to test this model, at least in an animal experiment. You apply a dye to the heart, and it gets absorbed into the cell membrane. When you shine green light on the heart, the dye absorbs the light and then emits red light. The amount of red light you get depends on the voltage across the cell membrane. In this way, you can use optical methods—cameras, lasers, and lenses—to measure voltage. It was exactly what we needed.

So Wikswo and his post doc Marc Lin decided to do the experiment. I waited a few months (these things take a while). Then Wikswo calls me on the phone and says “What would you say if I told you that my experiment looks exactly like your model for cathode make, anode make, and cathode break excitation, but is completely different for anode break?” I panicked, and started to list all the assumptions in my model that might explain this disaster. Wikswo just laughed and said “I didn’t say your model doesn’t match our experiment. I merely asked ‘what would you say if….’”. Ha, Ha. Very funny, Wikswo. The experiments did match the model, and I (really, we) had solved Dekker’s puzzle.

My theoretical paper was eventually published in the IEEE Transactions on Biomedical Engineering (it’s the paper I asked Dr. Harper to send to you all, so you can have a look if you want to see what a scientific paper looks like) and Wikswo’s experimental paper was published in the Biophysical Journal, and we all lived happily ever after.

Mechanisms

Before I end, let me tell you a little more about the mechanism for break excitation. When you apply a cathodal stimulus to refractory tissue, nothing happens. Right under the electrode, the tissue is strongly depolarized. Along the fiber direction, it is hyperpolarized in two places. The hyperpolarization pulls the refractory tissue back toward rest. In other words, it forces the tissue to end its refractory period early. It creates two islands of excitable tissue in a sea of refractory tissue. When the stimulus ends, the depolarization under the electrode can diffuse into these hyperpolarized and excitable regions, exciting them. This launches two action potentials that then eventually excite the whole heart. It was the hyperpolarization forcing the refractory tissue back to rest so it becomes excitable that was the key insight.

This insight turned out to be important in defibrillation. During fibrillation, the heart has many action potentials propagating chaotically, like turbulence. So the heart is quivering instead of beating in synchrony to pump blood. If your heart goes into fibrillation, it doesn’t pump blood you die in a few minutes unless you defibrillate. To do this, you apply a strong shock to the whole heart (like in the hospital TV shows, when they put the paddles on the chest and yell “clear!”). The shock can stop the action potentials, but when the shock ends it can cause break excitation. So, a defibrillation shock may fail if restarts fibrillation because of break excitation. There are some tricks you can do to prevent this from happening, which means you more effectively defibrillate. This idea is important. But applying these mechanisms to defibrillation wasn’t my idea, however, so I won’t talk much more about it.

Conclusion

So that’s my story. Imagine getting to spend you career solving puzzles like this one. I can understand why those folks in the book wanted to form a fellowship of puzzlemakers. That’s what science is. A Fellowship of Puzzlemakers and Puzzlesolvers. It’s a great life.

Writers Unplugged with Samuel Burr. 

https://www.youtube.com/watch?v=hpquvCAP9fU