Analysis of the Pathological Tremor by the AR Model.

OKADA Kiyoshi, HANDO Shima, TERANISHI Miyoko, MATSUMOTO Yoshinobu, FUKUMOTO Ichiro

doi:10.11239/jsmbe1963.38.275

This paper discusses whether an application of the AR model to an acceleration data of a tremor is useful for a differential diagnosis between Parkinson's disease and other diseases with tremors, such as the essential tremor. The degree of the AR model was chosen as 7 according to Akaike's FPE criterion. The examinees were 19 Parkinson's disease patients, 21 essential tremor patients —a disease that mainly appears in the elderly and in Parkinson's disease patients— and 13 healthy elderly as the control. This study on the acceleration data showed that the first prediction coefficient, the same as the main tremor frequency, is the parameter that classifies the Parkinson's disease group and the essential tremor group. The 7th prediction coefficient is the parameter that classifies the pathological tremors observed in Parkinson's disease and essential tremor disease patients and the physiological tremors observed in healthy subjects. Although the effects of another prediction coefficients on the differential diagnosis between Parkinson's disease and other diseases with tremors are not clarified during this stage, adding the AR model parameters to diagnostics by the main tremor frequency can increase reliability of the diagnostics. This paper showed the usefulness of the AR model for the acceleration data of the tremor to classify Parkinson's disease and other diseases with tremors.

Analysis of the Pathological Tremor by the AR Model.

Bibliographic Information

Description

Journal

Citations (1)*help

Details 詳細情報について

Export

Report a problem