Author ORCID Identifier
Semester
Summer
Date of Graduation
2025
Document Type
Dissertation
Degree Type
PhD
College
School of Medicine
Department
Exercise Physiology
Committee Chair
Sergiy Yakovenko
Committee Member
Valeriya Gritsenko
Committee Member
Jennifer Collinger
Committee Member
Jessica Allen
Committee Member
Jean McCrory
Committee Member
Sergiy Yakovenko
Abstract
Movement is the primary means by which the nervous system converts intention into interaction with the environment. Voluntary limb motion enables standing, walking, reaching, and grasping—actions that underpin self‑care, communication, and economic participation. When diseases or injuries disrupt the pathways that couple neural commands to muscle forces—whether through cortical stroke, spinal cord injury, or peripheral nerve damage—mechanical outcomes such as weakness, loss of coordination, and ultimately reduced mobility emerge. These impairments cascade into secondary complications, increased dependence on caregivers, and diminished quality of life. Quantifying how the lesion alters the mechanics of movement is therefore essential for designing targeted interventions that restore function and prevent long‑term disability.
In this dissertation, I have developed biomechanical models and computational methods to improve the diagnosis of—and rehabilitation from—motor impairments that arise after injuries to the central nervous system (CNS) and the peripheral nervous system (PNS), detailed in respective chapters. Chapter 2 presents real‑time forward and inverse dynamic models of bipedal locomotion and details analysis of selected numerical solvers and stabilizing viscoelastic contributions that enable fast, stable, high‑fidelity biomechanical simulations. Chapter 3 details the use of these models for quantitative description of the biomechanical origins of knee flexion deficits observed after stroke; to that end, I leverage a torque‑decomposition approach—rarely used in biomechanics but common in mechanical engineering—to identify variables contributory to the deficit. Chapter 4 shifts the focus from central to peripheral lesions, proposing and testing a novel real‑time method capable of inferring missing muscle activity from sparse EMG recordings, which could be used to generate control signals for assistive and rehabilitative technologies. Chapter 5 presents a perspective on anatomically constrained neural networks describing new approaches to approximate musculoskeletal transformations, needed to accurize and personalize them. Chapter 6 recapitulates main results and outlines directions for clinical translation and algorithmic refinements / extensions of presented solutions.
Recommended Citation
Bahdasariants, Serhii, "Computational Approaches for Diagnosing Neurological Impairments and Restoring Functional Movement" (2025). Graduate Theses, Dissertations, and Problem Reports. 12965.
https://researchrepository.wvu.edu/etd/12965
Included in
Biomechanics Commons, Cardiovascular Diseases Commons, Computational Neuroscience Commons, Engineering Physics Commons, Exercise Physiology Commons, Motor Control Commons, Musculoskeletal System Commons, Nervous System Diseases Commons, Physical Therapy Commons, Systems Neuroscience Commons