Parkinson's disease Assessment using Fuzzy Expert System and Nonlinear Dynamics

Parkinson's disease represents a chronic movement disorder, which is generally proportionally to age. The status of Parkinson's disease is traditionally classified through ordinal scale strategies, such as the Unified Parkinson's Disease Rating Scale. However, the application of the ordinal scale strategy inherently requires highly specialized and limited medical resources for interpretation. An alternative strategy involves the implementation of an iPhone application that enables the device to serve as a functional wireless accelerometer system. The Parkinson's disease tremor attributes may be recorded in either an effectively autonomous public or private setting, for which the resultant accelerometer signal of the tremor can be conveyed wireless and through email to a remote location for data post-processing. The initial testing and evaluation of the iPhone wireless accelerometer application for quantifying Parkinson's disease tremor successfully demonstrates the capacity to acquire tremor characteristics in an effectively autonomous environment, while potentially alleviating strain on limited and highly specialized medical resources.

The aim of the study was to develop a low-cost and compact system for analysis of tremor using a three-axis accelerometer (the Wii Remote (Nintendo)). To analyze tremor, we hypothesized that the influence of gravitational acceleration should be separated from that of movement. This hypothesis was tested experimentally and we also attempted to record and analyze tremor using our system in a clinical ward. A system for tremor measurement and analysis was developed using the three-axis accelerometer built into the Wii Remote. The frequency and amplitude of mechanical oscillation were calculated using methods for frequency analysis of the axis of largest variance and an estimation of tremor amplitude. The system consists of a program for measurement and analysis of Wii Remote acceleration (Tremor Analyzer), a Wii Remote, a Bluetooth USB adapter and a Web camera. The Tremor Analyzer has a GUI (graphical user interface) that is divided into five seg- ments. The sampling period of the analyzer is 30 msec. To confirm the hypothesis, mechanical oscillations were fed to the Wii Remote. The peak frequency of the power spectrum and the frequency of the oscillation generator were in good agreement, except at 1 Hz (0.01 G) and 2 Hz (0.02 G). With a change in the sum of squares of the three axes from 1.0 to 1.8 (G), the estimated and generated amplitude (0.3 cm) were in close agreement. This system using a Wii Remote is capable of analyzing frequency and estimated amplitude of tremor between 3 Hz and 15 Hz.