Grants and Contracts Details

Description

The goal of the proposed work is to broaden our understanding of three processes of speech sensorimotor control: 1) articulatory kinematics , 2) generalization of learning, and 3) articulatory coordination (mechanical equivalence). We will study these processes in talkers with non-progressive dysarthria and age-matched controls by analyzing the effects of spectral (formant frequency) modifications to auditory feedback on articulatory movements. University of Kentucky project component Dr. Johnson's team at the University of Kentucky will be responsible for the engineering and signal processing aspects of the study, including contributing to the software design and acoustic feedback software components as well as implementation of both kinematic and acoustic data analysis. One key aspect of this is the modeling of relationships between relevant task variables, in this specific case representing the control parameters of our articulatory synthesizer system, and the associated articulator variables which represent specific EMA measurement data in the articulatory space. Approaches to this modeling will include statistical methods such as locally-weighted regression as well as direct mapping methods such as neural network approaches. Project Specific Aims Non-progressive dysarthria is a speech sensorimotor disorder that afflicts 33% of permanently disabled survivors of traumatic brain injury (TBI) and up to 51% of stroke survivors. These two populations contribute approximately 480,000 new cases annually in the US. Dysarthria occurs on a continuum of severity. More severe impairment is associated with unintelligible speech that severely diminishes quality of life. Treatments often focus on teaching compensatory strategies for communication deficits, rather than rehabilitating speech to a premorbid form, particularly when dysarthria is severe. We have a limited understanding of how various processes of speech sensorimotor control are affected by non-progressive brain injury and how sensorimotor control deficits vary as a function of dysarthria severity, type, and individual differences. Consequently, dominant treatment approaches that focus on modifying speech loudness, clarity, and speaking rate, while functional and efficacious for some patients, are not substantively informed by an understanding of individual sensorimotor control deficits. We aim to broaden our understanding of three processes of speech sensorimotor control: 1) articulatory kinematics, 2) generalization of learning, and 3) articulatory coordination (mechanical equivalence). We will study these processes in talkers with non-progressive dysarthria and age-matched controls by analyzing the effects of spectral (formant frequency) modifications to auditory feedback on articulatory movements. It is well established that spectral modifications to auditory feedback can be used experimentally to elicit changes in speech movements that illuminate processes of sensorimotor control, but no research has used such an experimental paradigm to study dysarthria secondary to TBI or stroke. Three published research studies have used spectral modifications to auditory feedback to study sensorimotor learning in progressive dysarthria. Those studies all examined talkers with sub-clinically mild dysarthria secondary to neuro-degenerative disease processes (Parkinson's and Cerebellar Degeneration) using an acoustic resynthesis method to perturb auditory feedback. That method is only viable for talkers with very mild impairments. Our approach uses articulatory resynthesis, allowing for participants across the severity continuum. Our method converts a participant's articulatory movements (via electromagnetic articulography) into synthesized, "virtual speech" in real time to control auditory feedback. Virtual speech training elicits adaptive changes in articulatory movements that generalize to real speech. We will examine these adaptive changes in articulation to bolster our understanding of aspects of sensorimotor control in non-progressive dysarthria. Aim 1: Characterize baseline articulatory degrees of freedom and determine how they are affected by the size and location of the acoustic "task" space. Hypothesis: auditory feedback that mimics the acoustic space of a talker's real speech will maximize the specificity of articulatory learning, but will elicit smaller magnitude changes in articulation compared to learning within a completely novel (unfamiliar) acoustic space. Our method allows complete specification of both the acoustic space used for feedback and the mapping between articulator movements and synthesis controls. A talker must learn new articulatory movements to control the device even if the acoustic space mimics his real speech. This aim will help us further understand how manipulating the task space can affect articulation. Aim 2: Evaluate how the requirements of a virtual speech learning task affect generalization to real speech. Hypothesis: articulatory changes elicited in virtual speech will generalize optimally to real speech when learning involves connected speech tasks (rather than isolated sounds) and talkers with dysarthria will exhibit different magnitudes of generalization than controls. Preliminary data indicate that participants with dysarthria not only generalize virtual speech learning to real speech more effectively than neurologically normal talkers, but may do so more efficiently than typical talkers. This aim will allow us to evaluate how long generalization effects persist and if they bolster future learning. Aim 3: Characterize kinematic redundancy (motor equivalence) in the articulatory system and how it is coordinated during adaptive learning. Hypothesis: coordination within and between articulators will be altered in participants with dysarthria compared to controls. Dysarthria has been hypothesized to involve differences in articulatory coordination. Perturbations of auditory feedback can be used to examine "motor equivalent" adjustments to articulation that reflect adaptive changes in coordination. We will evaluate perturbation-elicited changes in articulation to examine how the brain coordinates kinematic redundancy during adaptive learning. This aim will help us understand if and how articulatory coordination may be affected by in non-progressive dysarthria.
StatusFinished
Effective start/end date8/14/187/30/22

Funding

  • Marquette University: $167,556.00

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.