From Prediction to Action: Neural Correlates of Musical Rhythm Production
Abstract
Music is one of the most pleasurable experiences for humans, being ubiquitous throughout history and across cultures. One of its backbones is the temporal organization of events – the rhythm dimension. Although it is clear in the perceptual domain that rhythm comprehension is a predictive processing, the bridge between auditory processing and motor control is lacking. This research gap led to this thesis, which attempts to study how prediction modulates rhythm production and its neurophysiological correlates.
Two studies were carried out in which non-musicians learned to produce several rhythms that varied in complexity. Specifically, measures of predictability were computed according to a model that captures enculturation. Additionally, visual representations capturing the rhythm’s temporal structure were provided. Together, both studies show that predictability allows for better performance overall and flattens the learning curves in synchronization and reproduction, suggesting differential engagement of motor control.
EEG recordings of the first study allowed the identification of performance monitoring mechanisms indexed by the Error Negativity (Ne) and error awareness indexed by the Error Positivity (Pe). In this experiment, asynchrony as a measure of error was treated continuously and related to changes in amplitude in these components as a function of predictability. The findings suggested a reinterpretation of these components in rhythm learning. The Ne was understood as a motor control prediction error in which the unpredictability of the sounds dampened this component due to imprecise encoding of the expected response. On the other hand, the Pe reflected a salient mechanism that tracked the complexity of the response.
As temporal information is critical for motor control in rhythm learning, the encoding of the rhythm template was examined. Previously, this temporal encoding was differentiated from motor control processes, as the IFG and IPL encoded the temporal structure. The fMRI findings of the second study suggest that the rhythm template is distributed across different parts of the motor control circuitry. In particular, visual information integrated in the basal ganglia allowed for better control over the timing of motor commands.
Together, the findings of this thesis highlight the integrated nature of timing in motor control and foster an interpretation of this function within a predictive framework. Specifically, rhythm production is a process in which auditory and motor predictions are connected through the encoding of temporal information to be performed.