The neural control of singing

Abstract

Singing provides a unique opportunity to examine music performance-the musical instrument is contained wholly within the body, thus eliminating the need for creating artificial instruments or tasks in neuroimaging experiments. Here, more than two decades of voice and singing research will be reviewed to give an overview of the sensory-motor control of the singing voice, starting from the vocal tract and leading up to the brain regions involved in singing. Additionally, to demonstrate how sensory feedback is integrated with vocal motor control, recent functional magnetic resonance imaging (fMRI) research on somatosensory and auditory feedback processing during singing will be presented. The relationship between the brain and singing behavior will be explored also by examining: (1) neuroplasticity as a function of various lengths and types of training, (2) vocal amusia due to a compromised singing network, and (3) singing performance in individuals with congenital amusia. Finally, the auditory-motor control network for singing will be considered alongside dual-stream models of auditory processing in music and speech to refine both these theoretical models and the singing network itself.

Keywords: audio-vocal integration; auditory processing; dual-stream model; non-musicians; singers; somatosensory; vocal pitch.

Figure 1

Neural networks of vocal motor control (central column), somatosensory (left) and auditory feedback processing (right), and hypothesized regions of sensory-motor control of voice [modified from a model proposed by Jürgens (2009)]. The vocal motor control hierarchy starts with the generation of complete vocal patterns from the reticular formation and phonatory motoneurons (white boxes), and then the next highest level of control (green boxes) stems from the anterior cingulate cortex (ACC) and periaqueductal gray (PAG), which can initiate and emotionally motivate vocal responses. The highest level of vocal control comes from the primary motor cortex (M1, blue box; its modulatory brain regions are not depicted), which is responsible for producing learned vocalizations (i.e., speech and song). Somatosensory feedback (dotted arrow) from various receptors distributed throughout the vocal tract is processed in the ascending somatosensory pathway (yellow boxes, left; black slanted lines indicate that only selected regions of this pathway are shown) and transmitted to the primary and secondary somatosensory cortex (S1, S2). Auditory feedback (dashed arrow) from the vocalization is processed by the ascending auditory pathway and auditory cortical regions (orange boxes, right). Potential neural regions that integrate sensory feedback processing with vocal motor control are indicated with red-outlined boxes, and their shared connections are represented by red arrows: (A) the PAG, (B) ACC, and (C) the insula (in purple, classified as a higher-order associative area).

Figure 2

Brain regions involved in auditory-motor control of singing, as observed in non-musicians and singers. (A) When voluntarily correcting for a 200-cent pitch shift in auditory feedback (“compensate 200c” task), non-musicians recruited more activity within the dorsal premotor cortex (dPMC) than singers. (B) Singers engaged the posterior superior temporal sulcus (pSTS), anterior cingulate cortex (ACC), and anterior insula (aINS) when performing the “compensate 200c” task. (C) Analyses of task-modulated functional connectivity revealed that relative to singing with normal auditory feedback, the 200-cent pitch shift specifically enhanced functional connectivity between right pSTS and intraparietal sulcus (IPS) during both the “ignore 200c” and “compensate 200c” tasks, as well as the postcentral gyrus (containing somatosensory cortex) during the “ignore 200c” task. Data from Zarate and colleagues (2008, 2010b).

Figure 3

A revised version of Berkowska and Dalla Bella's, Dalla Bella, and colleagues' (2009, 2011) vocal sensorimotor loop model for singing, updated with findings from Zarate and colleagues (2008, 2010b) fMRI studies. The covert pathway for pitch production (blue arrow) includes auditory cortex and inferior frontal gyrus (IFG), while the overt pathway for vocal pitch production (red arrows) is comprised of auditory cortex (STG/STS), intraparietal sulcus (IPS), anterior insula (aINS), anterior cingulate cortex (ACC), and dorsal premotor cortex (dPMC). Brain regions that are not visible normally from this lateral brain view are indicated in boxes outlined with dashes. Box colors are retained from Figure 1: light orange for auditory processing, green for vocal motor control, purple for multimodal processing.

Figure 4

Neural substrates for sensory-motor control of singing that are sensitive to the amount of vocal training [based on findings from Kleber et al. (2010, 2013), Zarate and Zatorre (2008), Zarate et al. (2010b)]. Brain regions that are not visible normally from this lateral brain view are indicated in boxes outlined with dashes, and box colors are retained from Figures 1 and 3. Activity within primary somatosensory cortex (S1) increases as a function of the amount of weekly vocal practice, suggesting a greater reliance on somatosensory feedback with more training and experience. After extensive vocal training and practice, the anterior insula (aINS) can serve a gating function for somatosensory feedback. Features within auditory feedback are processed and extracted by auditory cortex (STG/STS) and the intraparietal sulcus (IPS), and task-relevant auditory information is sent via the aINS to the dorsal premotor cortex (dPMC)—in people with little to no formal vocal training—or to the anterior cingulate cortex (ACC) in experienced singers to voluntarily adjust vocal output according to the singing task demands.