James Cavanagh has been at The University of New Mexico for 13 years studying cognitive neuroscience. He has been well-supported by the National Institutes of Health in his quest to use imaging tools to understand psychiatric and neurological disorders.
His most recent publication is a theoretical review of a specific brain signal called Reward Positivity. This unique electrical signal is only sensitive to rewards: it emerges as a burst of positive voltage from about 200 to 500 milliseconds after someone gets a reward. "It hasn't been a long time since it was discovered," said Cavanagh, a professor in the department of psychology. "We're still trying to figure out what it means and why it's so specific to this one type of outcome."
This brain signal is evoked whenever a test subject is given a task and then given a dollar, or told "that's correct," or even just given a 'thumbs up' icon. This signal gets larger, the more surprising the reward is.
"This sensitivity to surprise aligns with a foundational theory underlying all learning," Cavanagh said. "Reinforcement learning is a fundamental algorithm that is used to understand learning in humans, vertebrates, computers, etc. It's all based on how surprising rewards are."
That means Cavanagh, alongside Clay B. Holroyd from Ghent University, has found a neural signal, tightly aligned with theory, that can be measured in humans noninvasively. However further work on this new theory has suggested that researchers may have been using the wrong terminology: they have made a fundamental mistake confusing "reward" and "goal."
Rewards and goals usually align – but not always. "A goal can be something you work towards. If you're at Blake's Lotaburger and you're getting a drink, you might decide to get water instead of soda. The reward is the sugar, the soda. The goal is to be healthy," Cavanagh said. "It shows us something above and beyond the ideas of simple reinforcement learning—we're seeing how people set what they've decided is important to learn. It's like the brain checking off items on a list."
Cavanagh said this distinction is critical. While "rewards" imply dopaminergic systems, the word "goal" implies this system is being used for higher-order cognitive control. This is the system that links high level control and low-level reward processes together.
He hopes to bridge this theoretical work with a practical, clinical application. This work was funded by a grant looking into anhedonia—the decreased ability to experience pleasure as it relates to clinical depression. Cavanagh said they have a series of papers demonstrating that the Reward Positivity signal is smaller in people with depression, and researchers don't know why.
"Now we have new motivation to probe this signal, what it means, why it's smaller, and even leverage it for clinical benefit. For example, other groups have shown this signal can predict antidepressant response," Cavanagh said.
Many clinical trials fail because they're aiming for a very difficult target. Cavanagh believes testing the Reward Positivity signal could be used earlier in the process. Because it's very easy to assess this signal, Cavanagh hopes this can be used as a phase two or three clinical trial outcome measure, meaning its effectiveness is tested on a large scale.
"Instead of just measuring somebody's depressive symptoms, you could have a suite of small-scale tasks to see if brain markers that predict neural health are getting larger," Cavanagh said. "I think this also will help us on the theoretical end to define ideas that we use clinically, like anhedonia, to be more specific and more brain-aligned," Cavanagh said.
