The Avian Basal Ganglia Are a Source of Rapid Behavioral Variation That Enables Vocal Motor Exploration

Abstract

The basal ganglia (BG) participate in aspects of reinforcement learning that require evaluation and selection of motor programs associated with improved performance. However, whether the BG additionally contribute to behavioral variation ("motor exploration") that forms the substrate for such learning remains unclear. In songbirds, a tractable system for studying BG-dependent skill learning, a role for the BG in generating exploratory variability, has been challenged by the finding that lesions of Area X, the song-specific component of the BG, have no lasting effects on several forms of vocal variability that have been studied. Here we demonstrate that lesions of Area X in adult male zebra finches (Taeniopygia gutatta) permanently eliminate rapid within-syllable variation in fundamental frequency (FF), which can act as motor exploration to enable reinforcement-driven song learning. In addition, we found that this within-syllable variation is elevated in juveniles and in adults singing alone, conditions that have been linked to enhanced song plasticity and elevated neural variability in Area X. Consistent with a model that variability is relayed from Area X, via its cortical target, the lateral magnocellular nucleus of the anterior nidopallium (LMAN), to influence song motor circuitry, we found that lesions of LMAN also eliminate withinsyllable variability. Moreover, we found that electrical perturbation of LMAN can drive fluctuations in FF that mimic naturally occurring within-syllable variability. Together, these results demonstrate that the BG are a central source of rapid behavioral variation that can serve as motor exploration for vocal learning.

Cell-type specific outcome representation in primary motor cortex

“…In contrast, failure signals may promote network states that will generate behavioral variation on the next trial. If failure signals are generic error messages, the effects could be to put the PT network into a state that allows more explorations on future trials, analogous to a strategy thought to be used in songbird learning (Kojima et al, 2018). On the other hand, if the failure outcome signals provide information about the nature of the error, they might influence the network to produce a more specific adaptive correction.…”

Dopaminergic signals for reward, performance and social outcomes are dynamically gated during courtship

“…When a male zebra finch sings its courtship song to a female of interest, song is highly stereotyped and tonic levels of dopamine (DA) are increased in Area X, a vocal motor basal ganglia nucleus capable of regulating song variability 4–8 . Yet when males practice alone, song is highly variable and tonic DA levels are decreased in Area X 4, 9 .…”

“…Yet when males practice alone, song is highly variable and tonic DA levels are decreased in Area X 4, 9 . Blockade or disruption of striatal DA signaling eliminates the social context-dependent transition between ‘practice’ and female-directed ‘performance’ modes 10, 11 , suggesting that tonic DA levels actively regulate ongoing vocal variability 5, 7, 8, 12 .…”

“…To test how DA neurons may implement these dual functions, we recorded antidromically-identified VTAx neurons as we controlled both perceived error (with syllable-targeted distorted auditory feedback (DAF) 17, 22, 23 ) and behavioral state (with female present or absent) 5, 7, 8 . Surprisingly, tonic discharge patterns of VTAx neurons, including mean firing rate, median interspike interval (ISI), burstiness and firing regularity, did not significantly differ between undirected and directed song (Fig.…”

Spontaneous variability predicts adaptive motor response in vocal pitch control

“…The amount of this 73 compensation became larger when distributions of original and shifted FF variations are more overlapped 74 (Sober and Brainard, 2012;Kuebrich and Sober, 2015), linking the wider variability with the greater vocal 75 adaptations. It has also been shown that temporal patterns of FF fluctuation within a brief sound element guide 76 to keep and improve the song quality (Charlesworth et al, 2011;Kojima et al, 2018). Intriguingly, the vocal 77 variability in birdsongs is not simply due to the intrinsic noise in the peripheral motor system, but a certain 78 amount of them is 'actively' generated by a dedicated circuit that is required for song learning ( the active generation of variability in the motor processes is likely to suit to the adaptation-related motor 83 exploration (Dhawale et al, 2017).…”