Mechanisms of sensorimotor adaptation in a hierarchical state feedback control model of speech

Abstract

Upon perceiving sensory errors during movements, the human sensorimotor system updates future movements to compensate for the errors, a phenomenon called sensorimotor adaptation. One component of this adaptation is thought to be driven by sensory prediction errors–discrepancies between predicted and actual sensory feedback. However, the mechanisms by which prediction errors drive adaptation remain unclear. Here, auditory prediction error-based mechanisms involved in speech auditory-motor adaptation were examined via the feedback aware control of tasks in speech (FACTS) model. Consistent with theoretical perspectives in both non-speech and speech motor control, the hierarchical architecture of FACTS relies on both the higher-level task (vocal tract constrictions) as well as lower-level articulatory state representations. Importantly, FACTS also computes sensory prediction errors as a part of its state feedback control mechanism, a well-established framework in the field of motor control. We explored potential adaptation mechanisms and found that adaptive behavior was present only when prediction errors updated the articulatory-to-task state transformation. In contrast, designs in which prediction errors updated forward sensory prediction models alone did not generate adaptation. Thus, FACTS demonstrated that 1) prediction errors can drive adaptation through task-level updates, and 2) adaptation is likely driven by updates to task-level control rather than (only) to forward predictive models. Additionally, simulating adaptation with FACTS generated a number of important hypotheses regarding previously reported phenomena such as identifying the source(s) of incomplete adaptation and driving factor(s) for changes in the second formant frequency during adaptation to the first formant perturbation. The proposed model design paves the way for a hierarchical state feedback control framework to be examined in the context of sensorimotor adaptation in both speech and non-speech effector systems.

Neurophysiological evidence of sensory prediction errors driving speech sensorimotor adaptation

“…In addition, speech adaptation also seems to be sensitive to auditory feedback delays (i.e., 100 ms delay can eliminate adaptation), which highlights the importance of prediction errors that require temporally precise comparison of prediction and the actual feedback (Max & Maffett, 2015; Shiller et al, 2020). More recently, a computational model, Feedback-Aware Control of Tasks in Speech (FACTS, Parrell et al, 2019) also generated simulations of adaptation driven by auditory prediction errors (K. S. Kim et al, 2023).…”

“…Recently, another potential explanation, which is that adaptation is halted by prediction errors which quickly decrease throughout adaptation because of both the motor output changes and sensory prediction updates, has been put forth by a computational model, FACTS (K. S. Kim et al, 2023). In the simulation, the adaptive motor output produced lower F1 in response to F1 upshift perturbation, resulting in perturbed sensory feedback to become more like the baseline sensory feedback (i.e., lower perturbed feedback in F1).…”

“…Given that the right hemisphere SIS is known to be highly variable across tasks and individuals (K. X. Kim et al, 2023), the discrepancy may have been due to the sampling issue.…”

“…Another potential explanation is that adaptation is halted by prediction errors which quickly decrease throughout adaptation because of both the motor output changes and sensory prediction updates, an idea put forth by a computational model, FACTS (K. S. Kim et al, 2023).…”

Bayesian inference of state feedback control parameters forfoperturbation responses in cerebellar ataxia

“…A mechanistic computational model can thus be a powerful tool in evaluating these hypotheses by simulating the effects of specific model changes on the observable output. Tuning the parameters of a computational model to fit observed behavioral data has the potential to distinguish how different model components contribute to the observed effect [11][12][13][14][15][16].…”