Amanda Rubin

This week, the New York Times reported on a new Obama initiative that, in comparison to gun control or the economy, might seem a little frivolous. It’s called the “Brain Activity Map.”

Three Billion Dollars

The name of the project says it all: The goal is to map the connections in the brain in the same way the Human Genome Project mapped out the genes in human DNA. It’s expected to cost about $3 billion dollars over ten years.

If that seems like a pretty heavy price tag for the American people to take on, especially now, just to let scientist go poke around in the brain, you are right. But you are also wrong. Because the pay off could be tremendous.

Let’s use the Human Genome Project as our comparison. That project cost around the same amount as the Brain Map, coming in just under $4 billion when it was completed ahead of schedule in 2003. But the payoff from that project has been an incredible 150 fold, or $800 billion in 10 years. For any investment, that is a handsome return.

But can we replicate that kind of return? Especially in neuroscience, where scientists would be looking at a single organ, instead of the entire human genome? Absolutely. Alzheimer’s Disease, one of the poster children of neurological dysfunction, is expected to cost Americans $1.1 trillion dollars by 2050. That would create a 500 percent increase in Medicare and Medicaid spending. So the question is not really “is this project worth it?,” but more “Why haven’t we done this already?!”

A Difficult Path to Tread

 

Independently, scientists have been trying. But mapping the human brain is nothing like mapping the genome. DNA is a single straight line, a clear path. Conversely, following a signal in the brain, from neuron to neuron, is anything but. It is like walking a path that at every step splits in 10,000 directions. Or goes backwards. Or just stops.  As a person learns or forgets, all of a sudden the trail shifts. If that person get Alzheimer’s, whole sections of the path fall into an abyss.

To avoid this overwhelming complexity, scientists have looked at tiny pieces of the map, trying to understand the little part of the network as best they can. In theory, this approach could work, and it has worked for smaller problems. It has cured diseases where only a small piece of the network is being affected. But to understand whole-brain diseases, this approach is too uncoordinated and leaves enormous swaths of “unknown” separating small and growing “knowns.”

Putting Our Heads Together

So we know what we want to do, and we know what we are up against, but is it even possible at this point? What tools do we have? What other tools will we need?

For the latter questions, it depends who you ask.  Gary Marcus, a psychology professor at NYU, writing in the New Yorker blog, has suggested the initiative break itself into five separate projects. Some scientists would decipher the basic elements of neural computation.  Others would try to understand how neurons develop and then change connections over time. Still others would look the connection between genes, circuits, and behavior. And so on.

The trouble with this approach is twofold. First, it misunderstands what neuroscience has achieved. All of the above describe the “known” bits of how the brain works. True, these elements are not completely known, but they are bright lights of knowledge compared to the big black sea of enigma surrounding them.

Second, the Marcus approach misunderstands the way research is conducted and the potential gains from a paradigm shift. Small collaborations of scientists are already at work on the smaller issues mentioned. But for the bigger problems, we need all hands on deck. These are undertakings so data-intensive that Google and Microsoft came together to check that we have the computing power to see them through.  These high-level experiments are going to require fancy new tools, armies of scientists, and the money to keep them all going.

So how to approach this mammoth undertaking? The Europeans have taken one route, building a super-computer simulated brain based on the “knowns” we already have. Their device is trying to simulate everything we know about the brain, down to the neuron. The approach is a bit like modeling in economics, reducing a complicated web to something more manageable. But in biology, where an entire disease can hinge on a single molecular mutation, cutting out the details seems ill advised.

Brawn for the Brain

Assuming Obama’s monolithic “live brain in action” approach is the right one, which I believe it is, the only question left is what are the right tools for implementation. fMRIs, a research tool with a lot of popular press, can’t help us here. It can only look at more macro-level activity, but can’t see individual neurons or smaller parts of the path where all the action is happening.

So, scientists have been developing new tools. Just to see the physical map itself, to distinguish individual neurons from the grey mass, scientists at Harvard developed a technique to individually color neurons. This allows others to create their own wiring maps. This tool, called “Brainbow,” creates some truly stunning photographs, but they are just static images.

A similar technique uses an altered virus to jump from neuron to neuron, which then leaves a trail of where it’s been.  This provides a great way to discover individual paths and networks. It is now being used in a limited way to look at activity as well, the same viruses are being used to turn on and off whole pathways.

But where existing techniques fall short in tracking individual neurons’ activity, scientists are proposing new alternatives. Engineers are imagining wireless microcircuits to report activity to a computer.  Synthetic biologists are hoping to co-opt the neuron’s own machinery. This approach would basically create an in-cell ticker-tape of activity, that would be written out in real time on new strands of DNA.  Each newly created strand would hold 7 days worth of activity, which could later be read by researchers.

Whatever the future holds, it’s going to be complicated, and it’s going to be expensive.  It might unearth issues for ethics and regulation. But, if the Brain Activity Map is able to give Americans even a fraction of the research, health, and economic benefits it promises, it really is a no-brainer.

Amanda Rubin is a first year law student at Stanford, where she is also pursuing a Ph.D. in neuroscience. She is co-president of Stanford’s Interdisciplinary Group in Neuroscience and the Law (SIGNAL) and president of the BioLaw and Health Policy Society.