Patterns of Word Class Production between Picture Description and Narrative Tasks in Aphasia

Lee, Seung Jin; Kim, HyangHee; Seo, Sangkyu; Kim, Mikyong

ORIGINAL ARTICLE

Commun Sci Disord. 2009;14(4): 470-483.

Seung Jin Lee^` , HyangHee Kim^` , Sangkyu Seo^` , and Mikyong Kim^`

이승진(Seung Jin Lee)| 김향희(HyangHee Kim)| 서상규(Sangkyu Seo)| 김미경(Mikyong Kim)

ABSTRACT

Background & Objectives
Impairment in word production during spontaneous speech is considered to be one of the most distinguishable characteristics of aphasic speech production and has been shown to depend on the unique characteristics of each patient’s language and tasks. This study investigates whether aphasic patients with a wide range in fluency levels are divergent inthe Korean word class production proportions during spontaneous speech.
Methods
Spontaneous speech samples using the picture description task and narrative task of the Korean traditional fairy tale, ‘Heungbu and Nolbu,’ were elicited and analyzed using nine Korean word classes, three sentence components, and nineerror types. The proportions of frequencies and errors of each word class were compared between and within the two subject groups: 55 Korean patients with aphasia and 25 normal Korean adult controls. Patients were divided into three subgroups based on their fluency levels measured by syllables per utterance during the picture description task.
Results
The most frequently produced 3-word classes were noun, verb, and postposition in both tasks. However, error patterns were divergent among the groups in that the non-fluent aphasics produced significantly higher error proportions in nouns, verbs and prepositions, as compared to the control subjects. The common error types in both tasks were phonemic paraphasias and substitutions.
Discussion & Conclusion
Narrative tasks may be used in clinical settings as a supplementation of the picture description task in order to assess aphasic patients’ spontaneous speech. In further studies, error patterns in various types of narrative tasks or other quantitative indices, such as mean length of utterance, should be examined.