When does vocal learning happen?
Most vocal-learning species have an early sensitive period when the brain is especially receptive to acquiring new sounds. In songbirds, young males typically learn their songs by hearing adult songs early in life, and the timing of that exposure matters [3]. Parrots also begin a plastic phase of vocal development early, but their learning is not simply a copy of songbird development: parrots evolved different brain circuits for vocal repertoire acquisition, and their early learning appears to favor the growth and survival functions of corticosteroids rather than the sex steroids that shape songbird song [1].
Some species do not close the door on vocal learning after this early phase. Budgerigars, small parrots that live in flocks, continue to adjust their call types as adults when group membership changes. In a study comparing young adult and older adult budgerigars, both age groups showed equivalent vocal plasticity and converged on shared call types over a 20-day learning period [5]. Humans are another species that retains vocal plasticity across life, relying on auditory feedback to calibrate speech throughout adulthood [4].
Why early learning may depend on developmental hormones
Early vocal learning appears to be organized by hormones that shape the developing brain. In wild green-rumped parrotlets, giving corticosterone—a stress hormone—during the first week of vocal babbling led to larger vocal repertoires before fledging. When treatment began a week earlier in development, the effect was even stronger: early corticosterone treatment produced significantly larger repertoires than later treatment [1]. This is consistent with the idea that corticosteroids have more formative effects at earlier developmental stages, when they may help construct the neural circuits that support vocal repertoire acquisition [1].
The parrotlet study also found that corticosterone affected different stages of babbling in different ways. Both treatment groups showed weak negative effects on the early, reduplicated stage of babbling, but strong positive effects on the later, variegated stage [1]. That pattern suggests the hormone is not simply turning up or down a single learning switch; its influence depends on where the young bird is in the sequence of vocal development.
How later learning works: feedback and flexible circuits
Once an animal already has a vocal repertoire, later learning looks more like adjusting existing sounds than building a repertoire from scratch. Adult songbirds can learn to change the pitch of specific syllables when a sound or a mild skin stimulation is used as feedback, and the brain circuit involved—the anterior forebrain pathway—is required for this kind of adaptive change [4]. In experiments with Bengalese finches, adult birds shifted the pitch of a target syllable in response to cutaneous stimulation, and lesions of LMAN, the output nucleus of that pathway, blocked the learning. The same pathway and its dopamine inputs were also necessary for auditory-feedback-driven learning [4].
This suggests that later vocal learning may reuse a general feedback-learning circuit rather than relying on the same early hormonal organizing process. The songbird work shows that the circuit can use both auditory and non-auditory signals to guide vocal changes, which the authors propose may reflect a general principle for vocal plasticity across species [4]. In budgerigars, older adults showed persistent vocal plasticity and equivalent vocal convergence to young adults, but individual differences in learning were not explained by levels of the neural protein FoxP2, suggesting other mechanisms also contribute [5].
What is still unclear about early versus late mechanisms
The evidence supports a broad division of labor—early hormone-influenced organization versus later feedback-driven adjustment—but it does not show that these are entirely separate systems. The parrotlet study measured vocal repertoire size and babbling stages, not the underlying brain circuitry directly, so the proposed organizational role of the HPA axis remains a hypothesis consistent with the data rather than a demonstrated mechanism [1]. The budgerigar study found persistent plasticity in older adults but no link between FoxP2 expression and individual learning performance, leaving open which neurogenetic mechanisms regulate late-life vocal learning [5].
There is also a gap in how early social experience fits in. Human infants appear to learn the social efficacy of their vocalizations between 2 and 5 months of age: only 5-month-olds showed a vocal extinction burst when an adult stopped responding, indicating they had learned to expect responses to their sounds [2]. That is a different kind of learning from acquiring a species-typical song or call, and it is not yet clear how this early social learning connects to the hormonal and feedback mechanisms identified in other species.
Sources used in this answer
Early corticosterone increases vocal complexity in a wild parrot: An organizational role of the hypothalamic–pituitary–adrenal axis in vocal learning?
In wild green-rumped parrotlets, early corticosterone treatment produced larger vocal repertoires than later treatment, with weak negative effects on early reduplicated babbling and strong positive effects on later variegated babbling, consistent with more formative effects of corticosteroids at earlier developmental stages.
Learning how to learn from social feedback: The origins of early vocal development
Human infants learn the social efficacy of their vocalizations between 2 and 5 months of age, as shown by a vocal extinction burst in 5-month-olds but not 2-month-olds during still-face interactions.
Species differences in auditory responsiveness in early vocal learning
The paper examines species differences in auditory responsiveness during early vocal learning in song sparrows, with hints that early and late exposure to songs may affect juvenile behavior.
Author response: Shared mechanisms of auditory and non-auditory vocal learning in the songbird brain
Adult songbirds can adaptively change syllable pitch in response to non-auditory somatosensory feedback, and this learning requires the anterior forebrain pathway and its dopaminergic inputs, the same circuit needed for auditory-feedback-driven vocal learning.
Persistent vocal learning in an aging open-ended learner reflected in neural FoxP2 expression
Young adult and older adult budgerigars showed equivalent vocal plasticity and vocal convergence over a 20-day learning period, with persistent downregulation of FoxP2 in the MMSt but no relationship between FoxP2 expression and individual vocal learning performance.