Development of a Machine Learning Model to Detect Freezing of Gait in Parkinson Patients

Freezing of gait (FoG) is a highly incapacitating motor symptom of Parkinson’s Disease (PD), affecting on average 50% of early and 80% of advanced PD patients. It is characterised by a brief episodic absence or marked reduction of forward progression of the feet despite the intention to walk. Episode manifestation and frequency depend heavily on situation, environment, and patient. Ambulatory cueing could serve as symptomatic treatment of FoG, but detection techniques are needed to facilitate this. Convolutional neural network (CNN) based architectures using inertial measurement unit (IMU) data have shown promise in solving this problem. Therefore, this research sought to develop such a classification model to detect FoG episodes using IMU data of the ankle.