Duke University School of Medicine scientists have discovered a hidden brain circuit that helps explain how we learn from experience.
The finding published March 18 in Nature solves a long-standing mystery about the cerebellum, a small region at the back of the brain essential for learning and refining movements from walking and reaching to playing a musical instrument.
Neuroscientists from Duke and Harvard Medical School collaborated on the study that could also offer new clues to what goes awry in neurological conditions that impair learning and movement.
How the brain corrects mistakes
The cerebellum relies on powerful error signals carried by climbing fibers, which are unique neural structures that fire when movement doesn’t go as planned. When a movement is off — a missed step or a mistimed throw — climbing fibers send a signal that something needs fixing.
Those signals activate Purkinje cells, the cerebellum’s main output cells, triggering bursts of calcium that help rewire brain connections. These changes, known as plasticity, are how the brain learns.
But, according to co-senior study author Court Hull, PhD, there’s been a lingering scientific paradox. Hull, an associate professor of neurobiology at Duke, led the study alongside Harvard neuroscientist Wade G. Regehr, PhD.
“Climbing fibers also activate inhibitory cells that should prevent those calcium signals,” Hull said. “So, the question has been: how can climbing fibers promote learning and suppress it at the same time?”
The new research shows the brain resolves that conflict by briefly shutting inhibition off.
A hidden circuit explains it all
Using high-resolution electron microscopy, brain slice experiments, and recordings in living mice, first study author Fernando Santos Valencia, a Duke graduate student, found that climbing fibers don’t activate all inhibitory cells equally. Instead, they preferentially target a specific group known as ML12 cells.
These cells don’t inhibit Purkinje cells directly. Instead, they shut down another group of inhibitory neurons — ML11 cells — whose normal job is to suppress learning.
“Until Fernando’s finding it was not known that this type of disinhibitory circuitry existed in the cerebellum,” Hull said. “Or how climbing fiber inputs could engage this circuitry to promote learning.”
The effect is strongest when multiple climbing fibers fire at once. That kind of synchronized activity often happens during sensory experiences, like tripping on a hidden object, hearing a loud sound, or seeing a sudden movement.
When those signals arrive together, the brain briefly releases its internal brakes. Inhibitory activity drops, allowing Purkinje cells to generate strong calcium signals that reshape brain connections.
Because learning depends on how the brain takes in and organizes sensory information, this process links experiences to long-term learning. The finding helps explain why synchronized climbing fiber activity is especially effective in triggering cerebellar learning.
Just as important, researchers say the study highlights why inhibition matters in the first place.
“The key is having 'brakes' that can control neural plasticity,” said Santos Valencia. “Rather than constantly increasing error messages to produce plasticity and learning, a braking mechanism allows the brain to open a window for learning when needed and closing it when it’s not.”
The discovery could eventually help scientists better understand brain disorders.
“An imbalance of excitation and inhibition in the cerebellum could also lead to motor dysfunction or impaired motor learning,” Hull said.
“The hope is that by understanding how the circuit works to allow learning in normal conditions, we can pinpoint what elements are not working normally in cerebellar diseases such as ataxias, or other diseases thought to involve the cerebellum such as autism spectrum disorders."
Support for the study was provided by the National Institute of Neurological Disorders and Stroke, Edward R. and Anne G. Lefler Center, Nancy Lurie Marks Foundation, Alice and Joseph Brooks Fund, Ruth K. Broad Biomedical Research Foundation, Bertarelli Program in Translational Neuroscience and Neuroengineering and Stanley and Theodora Feldberg Fund.