In April, when Barack Obama announced the launch of the BRAIN Initiative, a well-funded drive to better understand the brain, some neuroscientists raised a sceptical eyebrow. The project’s fuzzy goals included mapping the activity of vast numbers of neurons in the brain as they fired—effectively, watching thoughts in real-time. Some people talked about mapping hundreds of thousands of neurons. Others spoke about recording activity from every single one—all 86 billion of them.
It seemed far-fetched. Today’s technology makes it difficult to record the activity of a handful of neurons, let alone thousands or billions. For example, you could use a technique called patch-clamping, which involves impaling neurons with microscopic electrodes. It’s arcane, cumbersome, very difficult to pull off in the brain of a living animal, and virtually impossible to do for more than a couple of neurons at the same time.
That’s a problem, when what we really want to do is to study entire neural circuits. Neurons work together to produce memories, make decisions, and compute information. If you can only study them one at a time, it’s like trying to understand a movie by only watching one pixel on a screen. “Understanding how a neural circuit is performing a computation requires knowing the electrical activity of all of the neurons in that circuit,” explains Michael Nitabach from Yale University.
His team, led by former member Guan Cao, has now taken us closer to that goal, with a new tool called Arclight. It allows them to record the activity of many, specifically-chosen neurons, in lots of detail, all without using any electrodes. Instead, it relies on light.
As neurons fire, waves of changing voltage spread over their surface. Arclight uses glowing proteins that can detect and respond to these fluctuations; the higher the voltage, the stronger the glow. By loading these proteins into the brain of a fly and watching the insect under a microscope, Cao could see its neurons firing as it responded to smells or fell asleep.
“This was absolutely transformative for us,” says Nitabach. “We’ve been doing electrode-based physiology for decades. Being able to look into the brain of a behaving animal and, for the first time, see membrane events occurring in real-time is just awesome.”
“These efforts will be critical to the success of BRAIN,” says Bill Newsome, a neuroscientist from Stanford University School of Medicine, who is co-leader of the initiative. He also points out that Arclight is the latest in a long line of increasingly powerful tools that can visualise neural activity without needing to stick anything into the brain.
For example, some proteins can flash in response to the changing levels of calcium ions that happen when neurons fire. A few months ago, a team at the Janelia Farm Research Campus used these proteins to film extraordinary movies of fish brains in action. It’s an amazing technique but also an indirect one—the calcium sensors are measuring a consequence of electrical activity, rather than the activity itself.
“They don’t reveal the moment-to-moment, nitty-gritty integration of inputs from a neuron’s many partners,” explains Newsome. Every neurons is, in itself, a living computer that receives and processes signals from its neighbours, through a multitude of branches. If the combined weight of these signals passes a certain threshold, the neuron fires. Calcium sensors tell you about that last bit, but not any of the computation that went before.
But Arclight can. It’s sensitive to even the small “sub-threshold” changes in voltage that build up in a neuron’s branches—the ones that determine whether it fires. Cao’s team could use it to draw a continuous map of electrical activity across a whole neuron, from its chunky body (soma) to its far-reaching branches (dendrites). “That’s been doable in cultured neurons, but we can do this for the first time in [a living brain],” says Nitabach. Put it this way: Calcium indicators can tell you which party won an election, but Arclight tells you how everyone voted.
Arclight is a gift from the sea. The protein used in the technique is a fusion of two others—a voltage-sensitive one from a sea squirt, and a flourescent one from a luminous jellyfish. Vincent Pieribone, a co-author on the new study, engineered the protein, while Nitabach’s team developed it for use in a living animal.
The gene that makes the hybrid protein can be delivered into a fly, so that it homes in on specific types of neurons. For example, the team loaded it into neurons that control the body clock of a fly, or those involved in processing smells, by linking it to genes used only in those cells. This “genetic targeting” is another attractive feature—it allows scientists to use the technique to study neural circuits that are involved in very specific roles, behaviours or parts of the brain.
So far, the team have successfully used Arclight to visualise handfuls of neurons, falling well short of the tens of thousands in earlier calcium-sensor studies. But it’s still early days. “This is a proof-of-principle,” says Nitabach. “The promise of the technique is to be able to do this on dozens or hundreds simultaneously.”
His group are now working on better techniques for detecting the light given off by the Arclight proteins. They need better resolution in both time and space, so they can work out what’s going on at specific parts of individual neurons within thousandths of a second.
But they’re already starting to use Arclight in its current unrefined form to study what happens in a fly’s brain as they fall asleep and wake up—something Nitabach has been working on for a decade. “We have a catalogue of hypotheses that we’ve been eager to test, but that we couldn’t until now,” he says.
For obvious reasons, Arclight currently works best in animals with transparent-ish brains, such as flies, zebrafish or nematode worms. The team’s colleagues are already using it to study mouse neurons, but they’re restricted to those that are near the surface. That still leaves us a long way from the goals of the BRAIN Initiative, but Nitabach says that’s okay. “Trying to achieve the BRAIN Initiative’s goal for a human brain, or even a non-human primate or rodent one, is grossly unfeasible right now,” he says. “Attempting to simultaneously record from every neuron in the worm or fly brain is a stepping stone to doing the same thing in a human.”
Reference: Cao, Platisa, Pieribone, Raccuglia, Kunst & Nitabach. 2013. Genetically Targeted Optical Electrophysiology in Intact Neural Circuits. Cell http://dx.doi.org/10.1016/j.cell.2013.07.027