Making the tools that solve biology’s biggest problems: an interview with Michael Lin

By Kristin Muench

Michael Lin, an associate professor of neuroscience and bioengineering and a member of the Stanford Neurosciences Institute, has earned acclaim for his development of new tools for studying biology, among them designer proteins that emit light when neurons in the brain communicate via electrical signals called action potentials. Graduate student Kristin Muench sat down with Lin to ask him about his research and the promise of technology in neuroscience research.

Why do you choose to work on biological tools instead of investigating a single question in biology?

It always seemed to me to be that science was limited by tools. We didn’t know that cells existed until microscopes were invented. We didn’t know viruses existed until electron microscopes were invented. We couldn’t do neuroscience until electrodes were invented. It always seemed like if you really wanted to have a big effect on a field, you would try to identify its limiting technology and make that technology. The more generally useful the tool is that you develop, the more impact you can have.

However, there is a downside to developing biological tools, which is that you may not get to work on any particular problem for very long. Problems are still, I think, the most interesting thing to many people, especially the public at large. People are very curious about specific topics, like why we have emotions, or why we dream, or why we get neurodegenerative diseases, and how can we cure them. If you want to make a big contribution to those topics, it pays to focus. Unfortunately, you can only live one life, so you have to pick your poison.

Neuroscientists need special tools to study the action potentials that neurons use to communicate. Your lab designs indicators that sense when action potentials trigger a flood of calcium ions into a neuron. It takes a lot of precision and control to make those indicators, but biology is very noisy. How do bioengineers manage to work at that interface?

That’s actually what makes working on tools nice. Tools provide specific measurements, and since we have the specifications we want to achieve, we can know definitively whether we’ve achieved them or not. There is a lot of trial and error, but that variability is actually motivating in terms of trying to make sensors better.

You can look at the example of imaging neural activity. Early on, people said that imaging calcium was difficult because the calcium signal from a single action potential was barely above noise. However, through persistence, calcium indicators were improved and this problem was solved. Now, in terms of just having a qualitative look at whether a particular neuron type is more or less active, calcium works great. Engineering calcium sensors has taken fifteen years of work, but it’s really revolutionized neuroscience.

To what extent do you think novel discoveries in biology rely on new tools?

Sidney Brenner is known to have said that “progress in science depends on new techniques, new discoveries and new ideas, probably in that order”. I generally agree with this.

If your timeframe is more than ten or twenty years, I would say that novel discoveries rely completely on new tools. That is the time it takes a field to use a new technology until all the possible discoveries that could be made with it are made. I think our global research community is large enough, and people are hungry enough, that low hanging fruits are not left unpicked very long.

For example, optogenetics was introduced in 2005, and thirteen years later, it is still an essential technique in systems neuroscience. However, in ten years or fifteen years, every circuit that’s known will have been probed by optogenetics. We will then be looking for new technology to gain further insights into the brain.

What developments are you most excited about neuroscience in the near future?

It’s really hard to pick one. We’re making a lot of progress in understanding the microanatomy of the brain. The brain is in some sense is a pre-anatomical structure. Unlike the kidney or the liver, where we know how the anatomy underlies function, we know very little about the microanatomy of the brain underlies its function. We know the macroanatomy of the brain, but that doesn’t really help you because neurons respond to individual neurons, not to regions. We really need to learn about what makes up the brain at the cellular level.

The most interesting aspect of neuroscience is what has become possible in such a short period of time. If you watched science fiction movies ten or twenty years ago, like the Matrix, or Eternal Sunshine of the Spotless Mind, you would think “This is really fun because it will never be true.” Now, we’re getting to the point where some aspects of these movies are actually possible.

In the Matrix, the characters can use neural activity to get an idea of what someone is thinking about. We are already doing that with mice to a low resolution degree. Specifically, we know what a mouse is about to do based on its pattern of firing recorded by calcium indicators. With voltage indicators, we’ll get an even more precise idea of what the mouse is thinking, in a sense. We just hope that these abilities can be used only for solving true problems.

As opposed to fictional problems?

As opposed to giving some people unfair advantage, or as opposed to using it for more nefarious purposes.

With great technologies comes greater responsibility.

Yeah. But you know, at least for now, these things are really hard to do. They’re not something that a garage inventor or Dr. Evil can do. It really requires an academic institution, and even then, we’re still far away from actually reading minds. Hopefully that will keep things reasonable for the forseeable future.