Development Of Wearable System For Freezing Of Gait Monitoring For Parkinson’s Disease Patients

Freezing of gait(FOG)is common gait impairment in Parkinson’s disease(PD).FOG is usually associated with fall risks and reduction of quality of patients’ life.Episodes of FOG are related to complex factors such as the state of the patient’s mode and the environments around the patient.Due to its episodic nature,FOG is difficult to predict and occurs under uncertain circumstances.It is usually the case that FOG doesn’t necessarily occur in clinical although the patient suffers FOG in daily time.Such cases make it difficult for clinicians to evaluate the stage of PD accurately.Therefore,using a daily FOG monitoring system to detect and assess the severity of FOG is very meaningful for clinical diagnosis and treatment of PD.In this study,we developed a wearable FOG monitoring system for PD patients.The system consists of a wearable node for motion data capture and a specialized client on the smartphone.MSP430 inside the wearable node accessed and controlled the inertial sensors through I2 C bus,for real-time motion data capture from the patient.Data was then sent to smartphone in term of data frames.A customized Android application was developed on the smartphone as a client for FOG monitoring and a display.The application received and decoded the data frames,and realized online FOG detection,gait symmetry and rhythmicity calculation,and the pedometer for the use of patients’ step counting during their daily walking training.The software of the application was all based on acceleration algorithms: a)FOG detection was realized by time-frequency combined analysis of acceleration.b)Gait symmetry and rhythmicity were derived from the unbiased autocorrelation coefficient of acceleration.And c)the pedometer counts steps according to the pseudo-periodicity of gait acceleration.All the monitoring items were displayed on the screen of the smartphone.Experiment results showed that the frequency of data capture reached150 Hz,which was much higher than the minimum frequency requirements50 Hz.Few data frames were missed during data transmission according to the timestamp sequence inside the data frames.Simulation on Matlab using cumulative eighty-minutes gait acceleration data,including 116 FOG episodes,captured from nine PD patients’ experiments showed the sensitivity and specificity of the FOG detection algorithm were respectively 90.8% and91.4%.The accuracy of pedometer was tested on five patients with altogether1157 steps,and got a result of 95.7%.The Android client used tables and histograms to display FOG episodes,gait symmetry and rhythmicity,and step counts in three different time periods,respectively daily,weekly and monthly.The system developed in this study was of good real-time performance and high reliability.With features such as high pertinence,low power supply and convenience in usage,the system was able to monitor FOG for long time in daily life,to provide a new assistant diagnosis method of PD.Furthermore,owing to real-time FOG detection,the system made context-aware FOG cuing possible in the future.This study made much sense in both clinical and daily life of PD patients.