Q&A: Physicist Christof Koch on leading a premier brain observatory

When Christof Koch’s PhD adviser moved from Germany to MIT in the early 1980s, he invited Koch to join him as a postdoc. The invitation was followed by a telegram in which, as Koch tells it, his adviser expressed concern: There was already someone in the US who modeled neurons. “He said, ‘You are going to be a father soon, and I am not sure I can guarantee you a job once you go out into the wider world.’” Undeterred, Koch went to MIT. He laughs about the telegram. “Now, of course, there are thousands of people who model the brain at all sorts of scales,” he says.

After a four-year stint at MIT, Koch joined the faculty at Caltech. Years later, in 2011, he left to take the reins of the Allen Institute for Brain Science in Seattle. His migration from physics to neuroscience was natural, he says. “I was always fascinated by computers, I was always fascinated by the brain, and using physics to study the brain just seemed like an obvious thing to do.”

At the Allen Institute, Koch directs a team of more than 330 people aiming to identify and list the different types of neurons in the brains of mice and humans and to build “brain observatories” to track the flickering activity of tens of thousands of neurons in mice as they carry out tasks. Another goal is to introduce common standards and reproducibility to the field of neuroscience. (See also the story on the International Brain Laboratory in Physics Today, May 2018, page 26 ; Koch is on the IBL’s scientific advisory board.) In addition, Koch is deeply interested in understanding consciousness from a physical point of view.

PT: Was the switch from physics to neuroscience difficult?

KOCH: I remember the first poster I gave at a neurobiology conference in Germany. I was a graduate student [at the University of Tübingen and the Max Planck Institute for Biological Cybernetics], and I was the only modeler—everybody else presented data from their experiments. My poster, on using cable theory to explain the features of dendritic spines, was relegated to the back of the room. Only two people came by. Two! One wanted to know where the bathroom was. He took pity on me and asked me to explain my computer simulations. That evening, I got quite drunk and wondered whether by modeling neurons I had made a career-killing mistake.

PT: Things clearly got better. How did you end up at Caltech?

KOCH: John Hopfield, a famous physicist who was one of the people who really boosted and gave credence to this field of computational science, spent a sabbatical at MIT. I shared an office with him. He said that Caltech was looking for someone who looks at brains from an engineering point of view and who looks at computers from a biology point of view. Caltech asked me out, and I got the job.

When I visited Caltech, the biologists said, “Christof is obviously very smart and all of that, but he is clearly an engineer.” The engineers said, “He is clearly a biologist.” Finally the provost had to break the logjam. He created a position that was in both divisions. That is common for people who work in border areas.

PT: At Caltech you pursued research on consciousness. What drew you to that topic?

KOCH: My most satisfying work has been on the question of consciousness. It’s an ancient question—Aristotle wrote about it. We find ourselves in a universe that is pretty well described by physics and allied sciences, but we are completely at a loss to understand why life should feel like anything.

Where does sadness come from? Where does happiness come from? I first thought about it with regard to pain. You have a toothache. We know the physical process: There is a receptor that is activated, the electrical activity goes along the nerves, goes inside the spinal cord, is switched up through a relay in the thalamus, and gives rise to electrical activity in your cortex. But so what? All that means is that there are some ions—sodium, potassium, chloride, manganese—sloshing around your brain. That’s just physics. Why should that hurt? Where does the feeling come from?

So back in 1989, when it was still very unpopular, I started the study of consciousness as a serious scientific enterprise with Francis Crick. We were the first of the modern wave of scientists who thought again about consciousness in a scientific way, in an empirical, measurable way. And we said let’s forget about the philosophy, let’s look for the footprints of consciousness in the brain. What are the specific mechanisms in the brain that need to be active in order for you to have pain or pleasure, or to see or smell? Is there special electrical activity associated with it, like oscillations? Are there specific genes associated with it? You can ask all those sorts of questions. Those are questions you can ask independent of your metaphysical position on whether consciousness is real. There are now hundreds of labs that look for the various footprints of consciousness in the brain. It’s big business.

PT: Has this research gotten anywhere?

KOCH: Yes. For example, there are thousands of patients for whom it’s difficult to know if anybody is home. They moan, they open and close their eyes, they move their head. Some will come back, some will die, and some will remain in a netherworld. Now there is a device being tested in clinical trials that probes the integrity of the cortex using a magnetic pulse. You measure the electrical response with a high-density network of electrodes, and then you can essentially look at the complexity of that response and infer whether the patient is conscious. It’s called transcranial magnetic stimulation.

PT: Are you still most interested in consciousness?

KOCH: Yes. This is one of a cluster of subjects relating to consciousness. Who else has it? If you are a cat or dog owner, there is no question to your mind that of course your dog [or cat] is conscious. So how low does it go? Is it only mammals? What about a squid or an octopus? They are very complicated creatures. What about a bee? A bee has only a million neurons; we have 81 billion. But a bee can recognize individuals. It has a complex waggle dance. Does it have consciousness?

We have to realize that the night is strange and full of wonders. Normal physics describes what the world looks like from the outside—magnetically, electrically, the strong, weak, and nuclear forces, et cetera. Maybe consciousness is how a particular type of physical system feels from the inside. It’s all the same—it’s just how it appears from the outside and the inside. I love thinking about consciousness from this point of view.

PT: What lured you away from Caltech?

KOCH: Partly it was that I like challenges. I was at Caltech for 27 years, and I knew I could be there literally until the day I died. Also, it was a time when I was becoming restless in my personal life, and I was looking for Christof 2.0.

PT: What is the challenge for you at the Allen Institute?

KOCH: One is that the biomedical sciences suffer from a replication crisis. Two out of three experiments, whether in brain imaging, cancer biology, or psychology, cannot be reproduced. That’s a scandal that goes right to the heart of the scientific enterprise.

There are a variety of reasons. One is that biomedical sciences are hypercompetitive. Another is that we are dealing with complex systems. Even the brain of a worm or a fly is vastly more complex and has vastly more degrees of freedom than a Higgs boson or a black hole, which are characterized by just a few numbers—charge, mass, angular momentum.

You need standardization. And you need to make all the data available to everybody on the planet. So I said to Paul Allen [the institute’s founder and funder], let’s build brain observatories. I want to bring the rigor of astronomy and the observatory model to neuroscience. He said yes. Our budget is roughly $100 million a year.

PT: How far does the comparison with astronomy hold?

KOCH: Well, one difference between astronomy and neuroscience is that there is only one sky. There is of course not just one animal. Even if you stick with one animal, like a mouse, every mouse is different. But we try to move toward a standard.

Also, unlike with the sky, with the brain I can turn individual neurons on and off with millisecond precision. So I am not only observing the system—watching the neurons in action as the mice are thinking—but also perturbing it, which is very powerful.

PT: What are some of the differences between running the Brain Institute and being a professor?

KOCH: It’s very different. I have a lot of latitude. I write columns and continue to interact with consciousness researchers. We are working on a project here at the Allen Institute, so I get to do that on the side. But I am an executive now.

We live and die by SMART goals: specific, measureable, actionable, realistic, timely. So I have to make go–no go decisions. It’s much closer to industry. There are never enough resources for everybody. People constantly complain. They are jealous. There are political issues and all of that. What’s difficult here is that I have to make decisions that directly affect people. It’s quite trying and often not very enjoyable. But it’s essential to achieving our larger vision.

PT: What are your plans for the future?

KOCH: I don’t know. There are so many things going on. The 10-year plan at the Brain Institute takes us to early 2022.

Biology is becoming mature, partly because of funding. NIH alone puts about $33 billion a year into biology. Overall it’s close to double that in the US, and worldwide it’s much more. Ultimately it behooves us to take insights and turn them into actionable things that help people. That’s difficult in a university. In physics, you have a rich ecosystem of small labs, national labs, international institutions. You need something similar in biology, and it’s beginning to happen.

It’s a very exciting time, and a very rich ecosystem of people. We really live in a golden age of brain science.