ABSTRACT
Parkinson’s diseases are the second leading cause of disability globally. Freezing of gait (FOG) is an episodic gait disturbance affecting patients’ locomotion, which closely relate to the risk of fall and is the most serious symptom of Parkinson’s disease. Prompt detection of freezing of gait (FoG) is crucial to fall prevention and effective intervention. This paper studies the accurate detection of FoG based on pure accelerations and multimodal data. An experiment was carefully designed to trigger FoG and multimodal data, including electroencephalogram, electromyogram, skin conductance, and acceleration, were recorded. A total number of 12 PD patients completed the experiments and produced a total length of 3 hours and 42 minutes of valid data. Several methods have been proposed to improve the accuracy of the FoG detection and promote the wearability of the sensoring system. Based on the acceleration data, a new time-frequency spectrum estimation method has first been proposed to improve the temporal and frequency resolution and then improve the performance of the Freeze index (FI), which is defined as the ratio of the areas under power spectra in ‘freeze’ band and in ‘locomotion’ band. Results showed that FOGs can be predicted in advance of its occurrence in most cases using the new calculated FI.
