Awardee Profile - Aaron Batista

BWF award helps researchers unravel neural patterns behind learning
Aaron Batista

When Aaron Batista graduated with philosophy and computer science degrees from the University of Pennsylvania in 1994, he serendipitously stumbled into the budding field of bioengineering.  An interest in neuroscience eventually led him to a postdoctoral fellowship at Stanford University, where he was surrounded by scientists trying to understand how populations of neurons communicate. In his own research, Batista wondered how neuronal communication can flex and change, allowing learning to take place.

Now an assistant professor of bioengineering at the University of Pittsburgh, Batista is putting his once only hypothetical ideas into practice. He and his colleagues have developed a new way to visualize patterns of neural activity. This technique has allowed them to measure the neural patterns behind learning and observe constraints on the adaptability of the brain. These findings will be published on August 28 as the cover story of the latest issue of Nature.

“With this project I can officially declare victory,” said Batista. “We have done what I proposed to do ten years ago.” The study was supported by a Burroughs Wellcome Fund Career Award in the Biomedical Sciences, which Batista received in 2003, as well as grants from the National Institutes of Health and the National Science Foundation.

In their study, Batista and his co-authors placed electrodes on the surface of the brains of two monkeys and hooked them up to a computer, creating a brain-computer interface. The electrodes, resting on a motor area of the monkeys’ brains, recorded the activity of about a hundred neurons. The firing patterns of these neurons could then be translated into the movement of a computer cursor. For example, one pattern of neural activity would move the cursor to the right and another would move the cursor to the left.

Over months of data collection, Batista’s graduate student, Patrick Sadtler, trained the monkeys to generate neural activity patterns that would move the cursor to different targets on a computer screen. When the cursor reached the target, the monkeys got a shot of juice as a reward.

The researchers then mathematically mapped the recorded activity of the monkey’s neurons onto a two-dimensional plane, like a piece of paper. This plane represented the network of neurons that communicate to move the cursor to a target. Translating the complex activity of a set of neurons into a geometrical surface allowed the researchers to visualize brain activity in a novel way.

“That was the key insight that enabled this really incisive experiment,” said Batista, who has been working on this idea for many years alongside Byron Yu, a co-author on the study and an assistant professor of electrical and computer engineering and biomedical engineering at Carnegie Mellon University. The two researchers are involved in the Center for the Neural Basis of Cognition, a joint program between Carnegie Mellon and Pitt.

With their new visualization technique in tow, the researchers could explore what happened if they changed the rules of the cursor game. After the monkeys learned to move the cursor with a particular pattern of neural activity, the researchers then altered the relationship between those patterns and where the cursor moved. Suddenly, patterns that moved the cursor left now moved it right. But over time, the monkeys relearned how to reach the targets with the cursor.

In other trials of the experiment, the researchers made a more substantial change. They reoriented the two-dimensional plane representing how the neurons talked to each other. Now the monkeys had to generate new patterns of neural communication to get the cursor to move to the target. This change was much harder to adapt to, and the monkeys’ performance on the cursor task was impaired.

These findings give insight into the common experience of learning skills related to previous knowledge more easily than entirely new skills. Imagine a professional ballet dancer learning a new routine. It’s much easier for her to learn one in the style of ballet, where she already knows the typical moves and technique, than a hip hop dance, which would require learning completely new ways to move her body.

Lee Miller, who studies how neurons communicate as part of networks, thinks it would be interesting to see how new neural activity patterns could be learned over the course of days or weeks, as opposed to the timescale of hours Batista and his colleagues used.

“This is somewhat of a theoretical study with potential ramifications for real world learning,” said Miller, a professor of physiology as well as physical medicine and rehabilitation at the Feinberg School of Medicine at Northwestern University.

This deeper understanding of the potential constraints on the adaptability of the brain may lead to improved interventions for patients who are paralyzed, have suffered a stroke, or have other conditions that are associated with abnormal neural activity patterns.

“Our study suggests a way in which we might be able to train these patients to show normal activity patterns again,” said Yu.

Some patients, such as those with paralysis, are already using brain-computer interfaces to learn to move robotic limbs. But Batista and his colleagues hope to explore how to predict which neural activity patterns would be harder for some patients to learn and coach them accordingly.

“Treatments for neurological conditions have been defined by drugs for decades,” said Batista. “We think direct electrical interventions need to go alongside drug therapeutic approaches.”

This study applies to a wide range of fields including computational neuroscience, engineering, and cellular biology, according to Yu, making it appropriate for the cover of the journal’s issue.

And this work serves as the next step forward in a long line of research, most of which has been made possible by the Burroughs Wellcome Fund. This independent private foundation is dedicated to advancing the biomedical sciences and has supported many researchers in Batista’s field who study how networks of neurons communicate, including his mentor at Stanford, Krishna Shenoy.

“It is safe to say this field was nurtured, and it’s exploded, due to the Burroughs Wellcome Fund in a way that it wouldn’t have otherwise,” said Batista.