Discrimination of Stabilograms using Machalanobis-Taguchi Method

Abstract

Stabilograms of healthy subjects and clinical cases with vertigo and equilibrium disturbances were analyzed for the purpose of discriminating normal from abnormal stabilograms of patients with labyrinthine and non-labyrinthine disturbances using the Machalanobis-Taguchi (MT) Method.
Method: Stabilograms were recorded with the eyes of the subjects open and closed for 60 seconds. The area, length/time, length/area, X- and Y- deviation, power spectrum, position and velocity vectors, and standard deviation, skewness and kurtosis of the amplitude histogram were measured. The total number of items measured was 69. The Machlanobis distance (MD) was calculated using these items.
Result: (1) On analysis using healthy subjects as the unit space and clinical cases as the signal data, 64% of the signal cases showed MD values exceeding the discrimination value and could be diagnosed as abnormal. The MD and abnormal ratio in the cases with brain diseases were larger than those in the cases with labyrinthine diseases. (2) On analysis using cases with labyrinthine diseases as the unit space and cases with non-labyrinthine diseases as the signal data, 68% of the signal cases could be discriminated as non-labyrinthine disorders. (3) On examination of the percent contribution of the items, the factorial effect graph showed that almost all of the 69 items contributed to differentiation of the abnormal from normal stabilograms.
Conclusion: The results suggest that the MT method is useful for the discrimination of stabilograms of healthy subjects from those of cases with labyrinthine and central nervous disorders.