Deep learning for predictive simulation of gait and post-treatment functional benefit in neurological diseases.
- Authors
- Publication Date
- Oct 08, 2024
- Source
- HAL
- Keywords
- Language
- English
- License
- Unknown
- External links
Abstract
Neurological conditions often manifest as gait disorders, frequently linked to spasticity. Botulinum Toxin Type A (BTX-A) injectionscommonly treat spasticity-related gait issues. Achieving optimal treatment outcomes with a favourable benefit-risk ratio remains crucial. Kinematic improvements obtained by this treatment are sometimes very efficient, but at this moment they remain difficultly predictable. The aim of this thesis is to employ deep learning (DL) techniques to simulate the impact of BTX-A treatment on gait parameters. The simulator aims to display the most probable gait result, enhancing the process of decision-making in BTX-A treatment. The database consisted of 43 adults diagnosed with various diseases, including CP, MS, TBI, SCI, and stroke. Each participant had undergone at least one clinical gait analysis (CGA) both before and after receiving treatment.The UGCEAM laboratory obtained and processed kinematic gait data. Various regression techniques were employed, including LSTM, BiLSTM, attention mechanism, ensemble learning, and Multi-task learning (MTL). The evaluated methods and their efficacy were compared both amongst themselves and to alternative approaches documented in the literature. This study is the first to quantitatively simulate the impact of BTX-A treatment on the gait of adults with various diseases. It explores a wide range of treatment combinations and different gait patterns.
