HomeOur ScientistsCarlos D. Brody

Our Scientists

Carlos D. Brody, PhD
Investigator / 2008–Present

Scientific Discipline

Computational Biology, Neuroscience

Host Institution

Princeton University

Current Position

Dr. Brody is also an associate professor in the Princeton Neuroscience Institute and the Department of Molecular Biology at Princeton University.

Current Research

Carlos Brody studies the neural mechanisms underlying cognition by fusing computational modeling with experimental neuroscience.
A decision-making task requiring accumulation of evidence over time...


Decision time. You've just sampled two flavors of ice cream, and you need to tell the person behind the counter whether you want rocky road or mint chip. What your brain needs to do is to consider the two options that are stored in your…

Decision time. You've just sampled two flavors of ice cream, and you need to tell the person behind the counter whether you want rocky road or mint chip. What your brain needs to do is to consider the two options that are stored in your short-term memory and decide between them.

Memory storage and decision making are both complex functions, and one might assume that neurons divide the labor among themselves. In reality, researchers have found that individual neurons in the prefrontal cortex, where these types of decisions are made, flip from memory storage to decision-making mode in the blink of an eye. Carlos Brody, a neuroscientist at Princeton University, wants to know: "What controls that switch?"

Brody began his career as a computational neuroscientist, creating computer models to explain how neurons in the prefrontal cortex make decisions under different circumstances. "We found that by changing the input, you could change [neurons' activities] from short-term memory to decision making," he says. "We're imagining that there might be either another part of the prefrontal cortex, or another cortical area, that exerts higher-level control, determining whether the prefrontal cortex is doing short-term memory or decision making."

Much of neural regulation depends on the balance of how neurons excite or inhibit the firing of other neurons, a mechanism Brody thinks is at work here. A small amount of inhibition, and the neurons stay in short-term memory mode; a larger amount, and they flip into decision-making mode. But, says Brody, "It's just a model. What we want to do now is figure out: Is it right or is it wrong?"

Answering that question means studying real animals. Most computational neuroscientists would have passed the task along to an experimental biologist. But Brody wanted to do it himself. He felt it was important to be directly involved in the experiments that would let him test and refine his models. "These experiments are difficult to do, and we're not really sure what the exact right test is. Developing the tests and the models side by side will make that process much, much faster," he says.

Brody's computer model was created by using data obtained from monkey studies carried out in the laboratory of Ranulfo Romo (an HHMI International Research Scholar), where Brody was a postdoctoral fellow. The monkeys had been trained to make decisions between choices stored in their short-term memory. Brody decided to try setting up a more efficient system, because his own lab was not well equipped to care for and train monkeys.

"We thought, if we could develop the same kind of behaviors in rats, maybe we—and other groups—could do tests much faster," he says. "Nobody else was developing the particular behavior [needed to test the model] in rats." Others thought examining such complex behaviors in rodents was impossible. That didn't stop Brody.

In collaboration with Zachary Mainen's and Anthony Zador's labs at Cold Spring Harbor Laboratory, Brody and his colleagues developed a system that trains large numbers of rats at once—by computer. Instead of sitting with one animal at a time, a single researcher can now oversee 20 animals while the computer offers each one a set of choices, determines when to reward the animal, when to increase repetitions, and when to move on to the next step. At day's end, the researcher gets a progress report and can adjust settings, revise computer code, or set up a whole new approach.

The automated system allows researchers to expand their studies and begin to investigate different types of decision making. The system has been so successful that researchers at other institutions, including groups at HHMI's Janelia Farm Research Campus, have begun using it to train their animals.

For Brody, the automated system is helping him test his model of how the brain prompts neurons to switch tasks. He's training his rats to respond to certain sounds. "The rats hear two sounds separated in time," he explains. "The first is a set of clicks. Then there's a gap, and then they hear another set of clicks. They have to decide which of the two had the higher frequency of clicks." The rats signal their decision by turning one way or the other, and they get a reward for a right answer. The behavior forces the rats to use their short-term memory, and then make a decision.

Once the animals learn their task, Brody's group attaches bundles of precisely spaced wires to their brains to measure the activity of individual neurons. The differences in those signals are then correlated with the rats' sensory function and motor activity. "Now we're going to look at their firing patterns as they engage in the behavior, and see if they conform to our model," Brody says. If they do, he plans to administer drugs to alter the level of inhibition in the prefrontal cortex, and see if this compels the rats to make decisions more quickly. Brody isn't stopping there. He plans to put the flexibility and speed of his rat-training methods to use in investigating other aspects of cognition.

Thinking about what he's accomplished with his crossover into experimentation, Brody calls it a gamble that paid off. "In retrospect, it was totally crazy," he laughs. "But it seems to be working out."

Show More


  • BA, physics, Oxford University
  • PhD, Computation and Neural Systems, California Institute of Technology


  • Alfred P. Sloan Foundation Research Fellow