Ambulatory analysis of gait and posture in Parkinson's disease: a novel method based on kinematics sensors

Introduction:Gait and posture in PD patients are generally analyzed in laboratory. However, laboratory techniques do not allow estimating long-term ON/OFF fluctuation in PD patients. Here, we propose an ambulatory system based on kinematic sensors that detects basic body postures and analyzes gait during long-term activity of free moving PD patients.Methods:Thirteen PD patients with subthalamic nucleus deep brain stimulation were monitored up to 5 hours with stimulation ON and OFF. A group of 10 healthy subjects performing typical daily activity was considered as control. Three wireless kinematic sensors were attached on the trunk and shanks to record basic postures (sitting, standing, lying) and gait periods. Anterio-posteror trunk’s velocity (TV) during sitting-standing transitions as well as transition duration (TD) was estimated. Also for each walking episode, gait parameters such as double support (DS), stride length (SL) and speed (SP) were extracted. Outcomes were compared to the patients’ UPDRS score.Results:Gait detection had an accuracy of 97%. A moderate significant correlation (r ≈ 0.68) between UPDRS sub-score and SL, SP and DS was found. Comparing to the normal, patients had significantly longer TD (3.3±1.0 vs. 2.7±0.5 sec) and slower TV (9.2±3.4 vs. 13.8±4.1 deg/sec). There were also significant differences in motor function between ON and OFF states.Discussion and ConclusionOur system assessed motor performance during long-term monitoring in free moving PD patients and showed significant correlation to the clinical scores. Based on gait and trunk movement it provides a new way to estimated ON/OFF fluctuation during daily activity of PD patients.