Analysis and Classification of Tremor Characteristics of Hepatolenticular Degeneration

Abstract

Tremor is one of the typical early symptoms of hepatolenticular degeneration (HLD), also known as Wilson’s Disease (WD), and is usually diagnosed clinically by the experience of doctors. The system uses a nine - axis sensor to record the vibration acceleration signal. Through the time domain and frequency domain analysis of the collected data of 43 WD patients and 9 healthy subjects, significant features were extracted. An artificial neural network with 4 input, 2 hidden layer with 5 neuron which was fed into sigmoid function was built. And through the supervision of learning, realizes the classification effectively. Experimental result shows that the four features are correlated well with the judgments of a neurologist, and the classifier block and neural network had high accuracy. A tight correlation between the clinical grade and the peak power of acceleration values reveals 93% accuracy by neural network with two hidden layers. The system makes it possible for WD patients to be monitored at home, and provides an effective auxiliary detection method for early WD diagnosis and treatment.