A baby’s babbling may sound like nonsense, but it’s actually an extended act of trial-and-error learning. As babies produce different sounds, their brains note which attempts succeed and which ones fail. Over time, that feedback leads to improvements and, eventually, fluid speech.

Learning Through Internal Feedback

HHMI Freeman Hrabowski Scholar Vikram Gadagkar wanted to understand the neural mechanisms behind this process: how the brain improves motor skills through practice.

It is well established that dopamine plays a major role in reward prediction error — the difference between the actual reward received and the expected reward. But this framework applies to animals (including humans) that perform specific actions in anticipation of tangible, external rewards, like food. Would the same mechanisms hold true when the goal is improving performance for entirely internal motives — for example, learning for mastery rather than juice?

Enter the humble male zebra finch. Many birds have dedicated song systems in their brains, making vocal learning quantifiable and easier to measure than in the more complex mammalian brain, where many circuits contribute to speech. Unlike birds with large repertoires — nightingales, for example, can mix and match hundreds of songs — male zebra finches learn just one simple song. That simplicity reduces confounding variables and makes them ideal for controlled experiments.

Tracking Vocal Mistakes Through Dopamine

The initial goal of Gadagkar and his colleagues was simple: identify the long-sought evaluation signal for song learning in zebra finches. To do so, they used auditory feedback to distort the bird’s song, tricking it into thinking it had made a mistake on the fourth note. When that syllable was distorted, the bird’s dopamine signals dropped. When the fourth note was restored, however, the dopamine levels rose above normal. Because the birds had grown accustomed to hearing this note incorrectly, hearing it sung properly became a rare and especially satisfying treat.

These results supported Gadagkar’s theory that even intrinsic, subtle rewards employ dopamine. With the evaluation signalexternal link, opens in a new tab identified, new questions emerged. Would the brain handle motor errors differently during practice and performance? After all, the goals are different. During practice, the bird is just trying to learn. During performance, the stakes are much higher, because the zebra finch’s song functions as a courtship ritual.

Zebra finches mate for lengthy periods — sometimes for life — so the male’s performance has major consequences for its long-term prospects.

How the Brain Shifts from Practice to Performance

Researchers had long observed that birds sang far more consistently during performances than in practice. In their next experiments, Gadagkar and colleagues found that the dopaminergic evaluation signal, so strong during practice, dropped sharply during female-directed performancesexternal link, opens in a new tab.

In other words, zebra finches were no longer focused on learning while performing. What, then, would trigger a brain signal reward for a performance well done? Although female zebra finches don’t sing, they do produce short vocal calls. It turned out that when the females called back during the male’s song, the dopamine surges returned. Like any musical performer, male birds seek positive reinforcement from their audience.

“The females also have song system in their brain, even though they don’t sing,” Gadagkar explains. “We think that it’s for evaluating and showing preference for songs that they like. We’re now recording activity in these brain regions as females listen to song, to see how this works.”

Dopamine Fluctuations Facilitate Adolescent Learning

Video demonstrating vocal learning by juvenile zebra finches. Credit: Spectrographs and audio clips from The Gadagkar Lab; Illustrations: by Matteo Farinella, courtesy of the Gadagkar Lab

Since zebra finches are among the very few animals to show vocal learning, they offer unique opportunities to study not only practice and performance, but the process of language acquisition itself. Having confirmed findings in adult zebra finches, Gadagkar and his team at Columbia Universityexternal link, opens in a new tab turned their attention to juveniles, who were learning to produce sound for the first time.

A juvenile’s birdsong is much more variable than a mature bird’s, so the results were even more pronounced. The more accurate the young birds’ songs were, the higher their dopamine levelsexternal link, opens in a new tab. Furthermore, Gadagkar found that spikes in dopamine accurately predicted how their song changed. Dopamine signaling accounted for not only the bird’s most recent rendition, but a whole sequence of past efforts, helping the young bird hone his song over time.

Linking Practice to Broader Brain Function

Because dopamine plays such a central role in human behavior — and disease — these studies could someday inform medical research. For example, people with Parkinson’s disease have impaired dopamine neurons; they also experience deficits in motor skills and performance. Understanding how dopamine contributes to motor skill acquisition and maintenanceexternal link, opens in a new tab may help guide future treatment strategies.

Similarly, the principles of how female zebra finches use social evaluation skills could potentially apply to humans, too. This line of research may offer insights into conditions in which social evaluation is disrupted, such as autism.

As Gadagkar puts it, “These birds only have a few syllables; they aren’t speaking Shakespeare. But the brain principles could still apply.”