Gait monitoring system for patients with Parkinson’s disease

• Autor(a) principal: Gonçalves, Helena Raquel Gouveia Silva
• Data de Publicação: 2021
• Outros Autores: Rodrigues, Ana Margarida Fernandes Marques, Santos, Cristina
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: https://hdl.handle.net/1822/83195
• topic: Wearable sensors; Adaptive computational method; Real-time application; Gait monitoring system; Parkinson’s disease; Engenharia e Tecnologia::Engenharia Médica; Science & Technology; Saúde de qualidade

Background: Wearable monitoring devices based on inertial sensors have the potential to be used as a quantitative method in clinical practice for continuous assessment of gait disabilities in Parkinson’s disease (PD). Methods: This manuscript introduces a new gait monitoring system adapted to patients with PD, based on a wearable monitoring device. To eliminate inter- and intra-subject variability, the computational method was based on heuristic rules with adaptive thresholds and ranges and a motion compensation strategy. The experimental trials involved repeated measurements of walking trials from two cross-sectional studies: the first study was performed in order to validate the effectiveness of the system against a robust 3D motion analysis with 10 healthy subjects; and the second-one aimed to validate our approach against a well-studied wearable IMU-based system on a hospital environment with 20 patients with PD. Results: The proposed system proved to be efficient (Experiment I: sensitivity = 95,09% and accuracy = 93,64%; Experiment II: sensitivity = 99,53% and accuracy = 97,42%), time-effective (Experiment I: earlies = 13,71 ms and delays = 12,91 ms; Experiment II: earlies = 12,94 ms and delays = 12,71 ms), user and user-motion adaptable and a low computational-load strategy for real-time gait events detection. Further, it was measured the percentage of absolute error classified with a good acceptability (Experiment I: 3,02 ≤ ε%≤12,94; Experiment II: 2,81 ≤ ε%≤13,45). Lastly, regarding the measured gait parameters, it was observed a reflection of the typical levels of motor impairment for the different disease stages. Conclusion: The achieved outcomes enabled to verify that the proposed system can be suitable for gait analysis in the assistance and rehabilitation fields.