Classification of aphasic Chinese speakers: cluster and discriminant function analyses

Abstract

The Western Aphasia Battery (WAB) is a standardized aphasia test which classifies aphasias into eight syndromes based on the Wernicke-Lichtheim model, commonly known as the Bostonian model. Although the syndrome approach has its limitations it is a widely used procedure in clinical aphasiology as it allows users to classify patients into discrete syndromes and to make some inference about lesion sites (Kertesz 1983). It has been shown that the Wernicke-Lichtheim framework is also applicable in classifying aphasia in Chinese, although Chinese and English linguistic structures are very different (e.g. Naeser and Chan 1980, Packard 1986, Gao and Benson 1990, Yiu 1992). The use of a common framework to classify aphasia in different languages is a useful concept, as it allows aphasia to be compared across languages. Unfortunately, aphasia classification is not always a clear-cut procedure, and use of the WAB is often queried regarding the validity of the quantitative classification. Previous reports have shown that the agreement on classification using the criterion scores proposed by the WAB and that of statistical procedures varied from 30% to 74%. Such variability could either be attributed to the differences in the sample size and statistical methods employed, or the inherent problems with the classification criteria. This study re-examined the adequacy of the classification criterion scores proposed by Kertesz (1979, 1982) by comparing the results of the classification using the criterion scores (clinical classification) and two statistical procedures (statistical classification). The agreement between the clinical classification and the statistical classification varied between 60% and 88%. The results were interpreted to support the use of the criterion scores proposed by the WAB.