New Tools for Quantitative Diagnosis of Parkinson's Disease Based on Scale Invariance of acceleration signals.
AbstractBackground: Movement disorders are neurologic syndromes that produce abnormal movements, being Parkinson’s disease (PD) the most frequent. Current diagnosis of Parkinson’s disease is clinical: it is based on expertise, which can result in diagnostic errors and treatment deficits. Quantitative diagnosis is reached searching objective parameters, a quantity that correlates with the phenomenon we want to quantify. Quantitative measures of movement are a novel approach to assist physicians in the diagnosis of Parkinson’s disease. Objective: To find a quantitative feature in accelerometry signals that distinguishes parkinsonian patients from age-matched control subjects. Method: We designed and developed an ergonomic wristband with proper electronics attached, including inertial sensors and control logic to manage the data signals. Additionaly, we developed a mobile app for logging and administration of data signals. We measured limb acceleration in two groups of people: a group of patients with Parkinson’s disease (n = 11) and an age-matched control group (n = 10). Acceleration signals were measured during the evaluation of finger tapping (from the motor subscore of the Unified Parkinson’s Disease Rating Scale - UPDRS part III). To characterize the signals we developed mathematical tools based on the analysis of temporal and frequency patterns. Power spectrum analysis of the signals exhibited a power law relationship between variables with different frequency ranges. Results: Interpreting the finger tapping task as an input to the motor system, variations in the scale invariance bandwidth can be related to the mechanic system's control response. Analyzing the power spectrum of the patient group signals, there is a statistical relevant reduction (p<0.005) in the frequency range that follows a power law (Linear Invariance Range, LIR) compared to the control group. LIR quantity calculated correlates with the gold standard clinical diagnosis MDS-UPDRS-III. Conclusion: The LIR quantity allows proposing a reduction of the scale. The quantity LIR found is proposed as a possible signal related biomarker of Parkinson’s disease. Keywords— Accelerometers, Biomedical engineering, Mobile Application, Movement Disorders, Parkinson’s Disease, Quantitative Diagnosis, Wearable Devices.