Kinetic tremor analysis using wearable sensors and fuzzy inference systems in Parkinson's disease

Background: Computer systems for evaluating Parkinson's disease (PD) have recently increased. Many existing methods allow the quantification of tremors and extraction of some characteristics of the acquired signals by analysing manoeuvres established in the MDS-UPDRS. Some of these current methods interpret the finger-to-nose test, which includes kinetic tremors of the hands; however, an evaluation strictly based on the guidelines of the MDS Unified Parkinson's Disease Rating Scale (MDS-UPDRS) is not performed, in addition to not using additional biomechanical indicators that make more robust and accurate monitoring of the patient's evolution. New method: The proposed method consists of a fuzzy logic system that evaluates PD patients in a range based on the MDS-UPDRS. The system evaluates by taking biomechanical features extracted from signals recorded with inertial measurement units (IMUs), which were previously processed for obtaining meaningful characteristics according to the MDS-UPDRS and other additional ones. Comparison with existing methods: In addition to the characteristics established by the MDS-UPDRS for the classification, this method uses other procedures that were considered necessary for the achievement of an accurate evaluation, such as the amplitude of the tremors in the different stages of the finger-to-nose manoeuvre, the tremors frequency and the voluntary movement frequency. Conclusions: kinetic tremors were measured based on a sensor network formed by IMUs. A Fuzzy Logic system obtains an accurate and repeatable biomechanical assessment of PD patients. This system will permit physicians to follow up on each patient with objective assessments improving medical treatments.