The singer and the song: The neuromechanics of avian sound production

Highlights

• The source-filter theory for human voice production is applicable to songbirds.

• Biomechanical effects of muscles remain only partially understood.

• Multiple mechanisms of sensorimotor feedback have been detected in songbirds.

• Simple models explain complex dynamics of respiration and song production.

Song is crucial to songbirds for establishing territories and signaling genetic quality and an important driver in speciation. Songbirds also have become a widely used experimental model system to study the neural basis of vocal learning, a form of imitation learning with strong parallels to human speech learning. While there is a strong focus on central processing of song production, we still have limited insights into the functional output of the motor neural circuits. This review focuses on recent developments in motor control, biomechanics and feedback mechanisms of sound production in songbirds.

Introduction

Neuromechanicists seek to understand how muscles, sense organs, motor pattern generators, and brain interact to produce coordinated movement [1], including locomotion, eye–limb coordination and also sound production. Many vertebrates have evolved the ability to produce sounds with highly specialized organs [2], driven by complex motor patterns [3, 4] and executed by exceptional muscles [5, 6].

Songbirds produce among the most sophisticated communication signals in nature and song is of critical importance to them. It helps birds establish and maintain territories and to signal reproductive quality to mates [7, 8]. The physical and neural mechanisms by which songs are perceived, produced, and learned are substrates chosen for and against by sexual and natural selection, and song is an important aspect driving speciation [9, 10]. Songbirds and their songs have also proven an invaluable model system to answer fundamental questions in neuroscience [3]. As a result songbirds have become a widely used experimental model system [11], especially for studies aimed at understanding the neural basis of vocal production learning, a complex form of imitation learning with strong parallels to human speech learning [12, 13, 14]. Although we have an increasing understanding of the physics [15] and sensorimotor control [16] of speech and song production in humans, there we lack the experimental manipulations to understand its learned control. In contrast, through studying songbirds, we have the unique potential to quantify the entire neuromechanical system, including neural circuitry interactions, peripheral biomechanics, and feedback mechanisms.

Vocal learning depends on integrated action of neural systems for auditory perception, song learning, and song production. The songbirds, that with over 4000 species comprise almost half of all living bird species, have a set of brain nuclei whose sole function is song learning and production and is referred to as the song system [11]. This circuitry is molecularly distinct from its immediately surrounding tissue [17] and comprises a song motor pathway, which spans from the telencephalon (nucleus HVC) to a brainstem vocal-respiratory network, and an anterior forebrain pathway traversing the telencephalon, striatum, and thalamus [18]. The vocal-respiratory network controls the three major motor systems essential to sound production: the respiratory system, vocal organ, and upper vocal tract. Our understanding of the central processing of song is advancing rapidly [11], but the output and function of these motor neural circuits is still not well understood. Here I review recent developments with a focus on the peripheral motor control and feedback mechanisms to the central control of sound production, and highlight gaps in our knowledge on how motor commands are translated into sound.