DOI
https://doi.org/10.25772/KQR4-5X73
Author ORCID Identifier
https://orcid.org/0000-0003-1924-0955
Defense Date
2023
Document Type
Dissertation
Degree Name
Doctor of Philosophy
Department
Biomedical Engineering
First Advisor
Dean Krusienski
Abstract
Speech is the first and foremost means of human communication. Millions of people worldwide suffer from severe speech disorders due to neurological diseases such as amyotrophic lateral sclerosis (ALS), brain stem stroke, and severe paralysis. A speech neuroprosthesis that decodes speech directly from neural signals could dramatically improve life for these individuals. Recent studies have demonstrated that it is possible to decode and synthesize various aspects of acoustic speech directly from intracranial measurements of electrophysiological brain activity. For those who have completely lost the ability to speak, the objective is to synthesize acoustic speech directly from brain activity during imagined speech. However, the lack of acoustic or behavioral output during imagined speech presents challenges in designing an effective decoding model. To cope with this limitation, it is common to utilize behavioral output from overt speech or mouthed speech (i.e., performing inaudible speaking articulations without vocalization) as a surrogate to study associated brain activity or to train decoding models for imagined speech applications. To improve the performance of these models, better understanding and modeling of imagined speech processes and a systematic comparison of neural processes across these speech modes in the context of speech decoding are required. This dissertation aims to provide a novel temporal, spatial, and spectral characterization of neural activity associated with the generation of overt, mouthed, and imagined speech. Stereotactic EEG (sEEG) data were collected from multiple subjects who performed overt, mouthed, and imagined speech trials while acoustic speech was simultaneously recorded. A systematic characterization and comparison of the neural activity during these different speech modes supports progress toward the development of a practical speech neuroprosthesis for the individuals with speech impairments. Specifically, this unique characterization provides important insights for the elusive goal of developing more effective imagined speech decoding models with respect to the better-established overt speech decoding counterparts. Furthermore, the results of this study provide a novel elucidation of the roles and of deeper brain structures, including white matter, and their potential contribution to a speech neuroprosthesis.
Rights
© The Author
Is Part Of
VCU University Archives
Is Part Of
VCU Theses and Dissertations
Date of Submission
3-15-2023