Abstract:
In neurodegenerative motor disorders, the impaired interaction among the
nervous system, muscles, the human skeleton, and the environment can impact
locomotion. Changes in walking patterns are directly associated with
a decreased quality of life and social engagement. The gait alterations serve
as indicators to determine the progression and severity of a disease, holding
significance for both, patients and healthcare providers. The selection of effective
therapy for maintaining or restoring gait requires the consideration of
the interdependencies between neuronal impairments and gait changes. Understanding
these dependencies is challenging given that the control of movements
results from this complex and dynamic interplay of the central and peripheral
nervous system, the musculoskeletal system, and the environment.
Neuro-musculoskeletal models offer a way to investigate these interactions in a
human-like simulated system. Digital gait recordings of healthy and impaired
participants can be used to verify the plausibility of the predicted gait patterns
by the neuro-musculoskeletal model.
The purpose of this work was to investigate and predict the gait in a neurodegenerative
motor disorder causing spasticity and muscle weakness, namely
hereditary spastic paraplegia. Of particular interest was the very early phase of
the disease, before participants and clinicians experienced any noticeable gait
alterations, as it is the most promising for therapeutic interventions. We call
this phase the prodromal stage. To achieve this goal, we conducted gait analysis
experiments with prodromal hereditary spastic paraplegia participants. We
used walking patterns as a functional outcome measure and linked it to physiological
alterations in the neuro-muscular control to predict the gait deviations
in a neuro-musculoskeletal model. This work focuses on understanding the natural
history of hereditary spastic paraplegia from the very beginning. Studying
the emergence of spastic gait patterns may help to maintain norm-like walking
patterns through early therapeutic interventions.
The first goal was to find specific and disease-related gait alterations already
in the prodromal stage that can be used as functional outcome measures to describe disease severity and progression. In an experimental study, we included
70 participants in the prodromal stage, manifest stage, or healthy controls.
All participants underwent an instrumented digital gait analysis in a movement
laboratory with high-precision camera measurements. The analysis was
carried out using two paths. First, gait features were analyzed considering a
priori knowledge about the disease with an in-depth analysis of spatiotemporal
gait trajectories. Second, a neural network was trained to learn gait alterations
that are decisive for the prodromal and manifest stages of hereditary spastic
paraplegia. We found specific gait changes that occurred already in the prodromal
stage and increased in more severe stages of the disease. Using convolutional
neural networks to learn decisive gait features, we found more specific
gait alterations in manifest patients. Therefore, we could describe the severity
in the prodromal and early manifest stages in an objective and physiologically
plausible way. The decisive features were further analyzed toward their predictive
capability to monitor longitudinal change in the prodromal phase. We
analyzed a two-year follow-up assessment and proved the severity-related gait
features as still valid after the two-year progression for the same participants.
For the same gait features, we found a significant longitudinal change in gait
characteristics toward the gait patterns of manifest patients. We conclude that
gait patterns can be used as functional outcome measures for future clinical
trials.
The second goal of this work was to link these altered walking patterns to
the impaired motor control processes in hereditary spastic paraplegia. We
used computer simulations using neuro-musculoskeletal models. By gradually
increasing disease-related symptoms (spasticity and muscle weakness), we
could predict kinematic and muscular changes comparable to the recorded gait
patterns of prodromal and manifest participants. With the gradual increase
of these parameters, we could reproduce the severity of the disease by mapping
altered sensory-motor control parameters to the severity-related kinematic
changes.
The overall results show that gait patterns can be used as decisive and functional
outcome measures for future therapeutic interventions already in the
prodromal stage of hereditary spastic paraplegia. Using neuro-musculoskeletal
models to predict and evaluate patient-relevant parameters, may be used to select
effective therapies or predict relevant outcome measures for clinical trials.
Neuronal Networks