These researchers are testing a Particle Filter (Sequential Monte Carlo Model) on finger tapping data sampled at 10k Hz in Parkinson's Disease (PD) patients and control patients. The Particle Filter tracks seven or nine parameters simultaneously, and attempts to solve differential equations in real time to trace the tapping data collected by the accelerometer. PD patients exhibits a phenomenon called "hastening," and the researchers aim to quantify hastening in the finger tapping data, and classify PD patients from control patients using the Particle Filter.
As this project progresses, the researchers are looking for parameters/features that are consistent across patients in order to perform classifications. This requires continuously modifying and testing the code.