Low-cost, 3-dimension, office-based inertial sensors for automated tremor assessment: technical development and experimental verification.

BACKGROUND

Tremors are common clinical complaints among the elderly and non-specialist physicians frequently are challenged by the need to provide an accurate diagnosis of various tremor syndromes, particularly Parkinson's disease and essential tremor in their busy practices.

OBJECTIVE

We sought to develop an easy-to-use, mobile robust, accurate, and cost-effective instrument that can objectively quantify tremors.

METHOD: The low-cost, 3-dimension, inertial sensors were developed for automated tremor assessment. The main sensor unit consists of a 3-axis accelerometer and a 3-axis gyroscope for the purpose of measuring the tilting angle relative to the gravity, linear acceleration, and angular velocity of the body segments affected by tremors. The transmitter consists of five main modules, including a microcontroller, power management module, sensor module, external memory interface module, and Bluetooth™ communication interface module, which connects to the sensors by a thin wire. The signal processing utilized fast Fourier transform analysis to include RMS angular rate, RMS angle, RMS rate, RMS velocity, peak frequency, peak frequency magnitude, and dispersion of frequency as variables.

RESULT: The prototype was tested with a tremor simulator at programmable angular rates of 2-, 4-, and 8-Hz confirming its accuracy. Twenty subjects (10 PD and 10 age-matched ET patients) participated as part of the experimental verification to perform three tremor tasks, including rest, postural, and kinetic tremor according to the teaching videotape of the motor section of the UPDRS. The mean peak frequency was significantly lower in PD than ET patients at rest on the x- (p < 0.01) and z-axis (p < 0.01). In PD patients, the RMS angular rate, RMS angle, RMS rate, RMS velocity, and peak magnitude were all significantly higher than those values in ET patient at rest while the data was not significantly difference during postural and kinetic actions. ET patients had significantly higher peak frequency during postural action in the y-axis than PD patients (p < 0.05).

CONCLUSION

The study provides the technical development of an accurate, inexpensive, and simple method to measure the kinematics of tremor in humans. Further studies are warranted to confirm the validity of each parameter and the diagnostic accuracy in each tremor syndrome.