The Dopamine System and Automatization of Movement Sequences: A Review With Relevance for Speech and Stuttering

Abstract

The last decades of research have gradually elucidated the complex functions of the dopamine system in the vertebrate brain. The multiple roles of dopamine in motor function, learning, attention, motivation, and the emotions have been difficult to reconcile. A broad and detailed understanding of the physiology of cerebral dopamine is of importance in understanding a range of human disorders. One of the core functions of dopamine involves the basal ganglia and the learning and execution of automatized sequences of movements. Speech is one of the most complex and highly automatized sequential motor behaviors, though the exact roles that the basal ganglia and dopamine play in speech have been difficult to determine. Stuttering is a speech disorder that has been hypothesized to be related to the functions of the basal ganglia and dopamine. The aim of this review was to provide an overview of the current understanding of the cerebral dopamine system, in particular the mechanisms related to motor learning and the execution of movement sequences. The primary aim was not to review research on speech and stuttering, but to provide a platform of neurophysiological mechanisms, which may be utilized for further research and theoretical development on speech, speech disorders, and other behavioral disorders. Stuttering and speech are discussed here only briefly. The review indicates that a primary mechanism for the automatization of movement sequences is the merging of isolated movements into chunks that can be executed as units. In turn, chunks can be utilized hierarchically, as building blocks of longer chunks. It is likely that these mechanisms apply also to speech, so that frequent syllables and words are produced as motor chunks. It is further indicated that the main learning principle for sequence learning is reinforcement learning, with the phasic release of dopamine as the primary teaching signal indicating successful sequences. It is proposed that the dynamics of the dopamine system constitute the main neural basis underlying the situational variability of stuttering.

Keywords: Parkinson’s disease; automatization; basal ganglia; chunking; dopamine; movement sequences; speech; stuttering.

FIGURE 1

Schematic outline showing the midbrain dopaminergic sources and main pathways to the striatum and the cerebral cortex. In particular the cortical pathways are incompletely characterized, and motoric cortical regions may be innervated by both the VTA and the SNc (Gaspar et al., 1989; Björklund and Dunnett, 2007; Hosp et al., 2019). VTA, ventral tegmental area; SNc, substantia nigra pars compacta; NAcc, nucleus accumbens (The background sagittal brain view is from Patrick J. Lynch, medical illustrator, CC BY 2.5 <https://creativecommons.org/licenses/by/2.5>, via Wikimedia Commons, https://commons.wikimedia.org/wiki/File:Brain_human_sagittal_section.svg. Illustration of the basal ganglia and the dopamine pathways by Per A. Alm).

FIGURE 2

Schematic showing the hierarchical organization model of action sequences, with sub-sequences as sub-chunks. For speech the lowest levels (a) may be the movements of individual speech muscles, the next level (C) may be phonemes, and the higher level (S) may be syllables. Syllables may, in turn, be chunked into automatized multisyllable words. Reprinted from Jin and Costa (2015) with permission.

FIGURE 3

Schematic illustration showing the three different types of learning: According to this model, the basal ganglia are specialized for reinforcement learning, based on rewards in the form of dopamine signaling. The cerebellum is specialized for “supervised learning,” which adjusts the output based on movement error, independent of reward. Finally, the cerebral cortex is specialized for unsupervised “Hebbian” learning, which, might be modulated by inputs from the basal ganglia. Reprinted from Doya (2000) with permission.

FIGURE 4

Different patterns of signaling were observed in neurons in the basal ganglia during the execution of a learned action sequence with eight units. Based on recordings from the SNc, striatal projection neurons, and from basal ganglia output neurons, in mice. The red dots at the top represents a timeline of eight actions in an action sequence. The black surfaces represent the variations in firing in different populations of basal ganglia neurons. For example, one population signals for every action, while other neurons only signal at the start and the end of the sequence. Reprinted from Jin and Costa (2015) with permission.