Investigating Subcortical Contributions to Speech Sequencing in Deep Brain Stimulator Recipients

Boston University (Charles River Campus), Boston MA

Investigators

Linked publications, trials & patents

Abstract

PROJECT SUMMARY The overall aim of this project is to leverage the capabilities of modern deep brain stimulation (DBS) devices to investigate the roles of 2 key subcortical circuits, the basal ganglia (BG) circuit and the cerebellar (CB) circuit, in speech production. No studies to date have investigated DBS effects on learning and execution of speech sequences, functions central to fluent speech and well-known to involve BG and CB circuits. We aim to fill this knowledge gap by investigating the roles of the BG and CB circuits in speech motor sequence learning and execution in DBS implant recipients whose devices can be turned on and off in an experimental context. According to the GODIVA neurocomputational model of speech sequencing, the BG and CB play different roles in sequence learning: the BG contribute to learning the order of items in a sequence, whereas the CB contribute to optimizing the muscle activation patterns (motor programs) for producing the sequence. These hypotheses will be tested for 2 types of speech sequences: phoneme sequences within a syllable (Aim 1) and syllable sequences in multisyllabic pseudowords (Aim 2). Each study will involve 20 participants with DBS in the BG (specifically the subthalamic nucleus, or STN-DBS) and 20 participants with DBS in the CB circuit (specifically, the ventral intermediate nucleus of the thalamus, or VIM-DBS). DBS state will be manipulated separately during training and testing phases to distinguish roles in learning from roles in execution. Local field potentials will be measured from the DBS electrodes to test hypotheses concerning neural activity changes with learning. Based on the GODIVA model, we predict that sequence learning will be impaired if DBS is off during training in participants with STN-DBS. However, since the CB is not hypothesized to play a role in learning sequence order, we predict that sequence learning will not be affected for VIM-DBS participants trained with DBS off. Instead, we predict worse motor program learning with VIM-DBS off, evidenced by longer utterance durations. For both STN- DBS and VIM-DBS, we predict DBS state during the test phase will not affect sequence accuracy but will affect movement duration. We further predict that LFP beta-band power during perceptual processing will be decreased for learned sequences, indicating earlier onset of motor preparation as perceptual processing demands decrease with learning. In Aim 3, we will modify the GODIVA model software to include both BG and CB circuits to allow us to simulate the Aim 1 and 2 experiments and directly compare model performance to the experimental results. A neurotypical version of the model will be compared to BG-impaired and CB-impaired versions to test its account of sequence learning and execution in both DBS ON and DBS OFF conditions for both participant groups. Regardless of whether our initial hypotheses prove to be correct, this project will provide crucial knowledge concerning the effects of DBS, and more generally the roles of the BG and CB circuits, in speech learning and execution, thereby enabling improved treatments for those with speech disorders due to BG or CB impairments.