Objective: To evaluate the visual-perceptive assessment to detect upper extremity bradykinesia by using marker-less full body motion capture with Microsoft Kinect™ V2.

Background: Instrumental motor assessment for symptoms of Parkinson’s disease (PD) has been tested using different technologies, mostly using sensors or markers worn on the patient’s body. We explore marker-less motion capture as an easily applicable alternative. This may pave the way for low-cost and feasible home-based multi-point assessment of PD motor symptoms, e.g. to aid treatment decisions.

Methods: We used a protocol (PASS-PD), that includes four upper extremity motor tasks including two items of UPDRS (finger tapping, hand grip, arm holding and finger-nose test) which are performed in front of a Microsoft Kinect™ V2 sensor. The protocol was performed by 15 patients with PD (age: 44-81, MDS-UPDRS III: 7-52) in 2-3 different medication states. We used custom-written software (Motognosis Labs) to extract kinematic parameters for the description of bradykinesia and asymmetry. Main movement frequencies (besides arm holding) and ranges of motion were extracted for each item and each bodyside respectively. For repetitive movement tasks, similarity measures and arrhythmicity were calculated as well.

Results: There was a tendency for intra-individual performance differences between OFF and ON medication states, though non-significant in this still small number with only 13 patients with respective data pairs available. Patients showed lower average finger tapping movement frequency in OFF state (2.24Hz) than in ON state (2.49Hz) with similar results in hand gripping (OFF: 2.25Hz ON: 2.51Hz). No difference in movement frequencies was found in finger-nose-test (OFF: 0.35Hz ON: 0.36Hz). The complete set of resulting kinematic parameters also showed asymmetry. To assess asymmetry, significant differences in finger tapping and hand gripping test between better and worst body side could be found (p < .001).

Conclusions: The test series demonstrated good applicability and protocol compliance and we confirm and extend previous findings of the applicability and potential of visual-perceptive computing by an easy-to-use means to quantify bradykinesia in PD. We were able to see individual differences in different medication states and possible relations to disease severity as well as bradykinesia.

« Back to 2018 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/instrumental-assessment-of-upper-extremity-bradykinesia-in-updrs-testing-based-on-depth-video-data/