Time-frequency representations from inertial sensors to characterize the gait in Parkinson’s disease

Abstract

Parkinson’s Disease (PD) is a neurodegenerative disorder of the central nervous system whose main symptoms include rigidity, bradykinesia, and loss of postural reflexes. PD diagnosis is based on an analysis of the medical record and physical examinations of the patient. Besides, the neurological state of patients is monitored with subjective evaluations by neurologists. Gait analysis using inertial sensors was introduced as a simple and useful tool that supports the diagnosis and monitoring of PD patients. This work used the eGaIT system to capture the signals of the accelerometer and the gyroscope of the gait in order to evaluate the motor skills of patients. Fourier and wavelet transform were used to extract measurements based on energy and entropy in the time-frequency domain. The extracted characteristics were used to recognize differences between PD patients and healthy individuals. The results enabled to classify said groups with an accuracy of up to 94%.

Keywords

• Parkinson’s Disease
• inertial sensors
• time-frequency representation
• wavelet transform
• gait analysis
• supervised classification