Ambulatory gait analysis in Parkinson's Disease: Application of a novel method based on kinematics sensors

Summary: Using our ambulatory monitoring system, ASUR, 13 Parkinsonian patients were monitored for periods of up to 5 hours while they were free to perform activities at will. Every one hour patients did a UPDRS test as the clinical assessment. During the period of the monitoring, subjects had their Stimulation ON and OFF. Analyzing the data, periods of gait have been detected and gait parameters were calculated. We’ve found significant correlation between the results of this objective method and those of the clinical score.Conclusions: With minimal number of sensor sites, we could detect and analyze the gait parameters of PD patients as they were performing normal, unrestricted activities; while the results show a correlation to clinical scores. Using this method an objective evaluation method to assess gait in PD patients for extended periods of time is possible. INTRODUCTION : Parkinson’s disease (PD) influences gait with aggravation of quantitative parameters like gait cycle time (GCT), period of swing (SW), double support (DS), stride length (SL), speed (SP), range of rotation of shank (RS) and peak angular velocity of shank (PV) that have been shown to be improved by bilateral STN [1]. We have designed a new ambulatory system (ASUR) to monitor PD patients while they are performing they daily activity. We have used our analysis algorithm to quantify gait parameters in PD subjects while they were performing normal, unrestricted activities.PATIENTS/MATERIALS and METHODS : 13 PD patients with the age of 64±8 participated in this study. Subjects were treated by Sub-Thalamic Nucleolus Deep Brain Stimulation (STN-DBS). Each measurement took up to 5 hours during which subjects were free to move around in the hospital area. Two observers took note of the subject’s activities using portable computers. To measure gait parameters, ASUR devices, with an integrated uni-axial gyroscope in sagittal plane, where attached to each shank of the subjects. System had a sampling rate of 200Hz and angular velocities of shanks were recorded in a range of ±600º/sec. At the beginning of the measurement and then after each one hour, a UPDRS test was performed. Subjects started with STN ON, subsequently it was turned OFF for 3 hours and ON again for the last hour. By running our analysis program after the measurements, gait periods were automatically detected and respective gait parameters for each gait cycle were calculated. To compare to the clinical score, for each hour the average of the values of the gait parameters during that period was calculated. The first period was used as the baseline and the correlation between changes in UPDRS sub-scores (27;28;29;30) and changes in gait parameters from the baseline were calculated using Pearson method.RESULTS: Walking periods could be detected with a specificity and sensitivity of more than 97%. For each walking trial, gait parameters for each gait cycle were calculated. We could find a significant (p < 0.0001) correlation between the UPDRS sub-score and RS (r=-0.68), SL (r=-0.68), SP (r=-0.65), PS (r=-0.65), SW (r=-0.65) and DS (r=+0.65). GCT however, had no significant correlation to the clinical score.DISCUSSION: Gait in PD has already been assessed in gait labs with standard methods (like using force-plate or camera). Also in our own previous studies, good correlation between clinical scores and gait parameters has been reported but in our knowledge so far there has been no available method to analyze gait in PD, follow the fluctuations and have a correlation to the clinical scores while subjects were performing activities on will for several hours.