UC Merced

Despite being a common activity, singing is a complex behavior. It involves processes such as: maintaining mental imagery of desired tones, activating a motor plan to make a vocal production, coordinating different parts of the vocal apparatus and respiratory system, and monitoring the auditory and kinesthetic output for accuracy. While singing can be achieved by a majority of people, some cannot do it accurately. In this dissertation, I investigate one of the posited causes of poor pitch matching and experimentally test it.The inverse-model deficit of poor pitch matching suggests that poor singing results from the brain’s inability to instantiate a precise motor plan with one’s vocal apparatus when one tries to vocally reproduce a pitch. This hypothesis is motivated by previous work in sensorimotor control suggesting that in order to perform an action in an environment, one relies on two distinct internal models. These internal models are thought to be instantiated by the cerebellum. In this dissertation, I embed this hypothesis within the larger context of cerebellar cognition and the Bayesian brain hypothesis.

First, I present multiple findings of the involvement of the cerebellum in cognitive tasks, including in musical production. This will provide some support for the role of the cerebellum in cognition, in addition to its role in motor control. This first part will also allow for an understanding of some of the neural underpinnings involved in implementing internal models.

Next, I present recent findings which could support the notion of internal models within the Bayesian brain framework. This will provide an understanding of these internal models at an algorithmic level (Marr, 1982).

Finally, I test one of the key predictions of the inverse-model deficit of poor pitch matching i.e. perturbing the cerebellum during singing task would lead to poorer pitch matching performance. Adult participants had to perform various singing assessment tasks before and after brain stimulations on different brain regions. I did not observe a decrease in pitch matching accuracy (relative to the individuals’ baseline) after stimulation of the cerebellum. While these results were inconclusive, further research using TMS to investigate pitch accuracy should be undertaken so as to better understand the role of internal models in vocal production such as singing.