Proposing a quick, objective, comprehensive analysis of impaired spoken discourse in post-stroke aphasia

Abstract

Purpose: Aphasia is an acquired language disorder, most commonly caused by a stroke, that adversely affects one's ability to understand, speak, read, and write. Diagnosis of aphasia is typically done through administration of standardized aphasia batteries, many of which lack a detailed and adequate evaluation of spoken discourse. The aim of this paper was to propose a quick, objective, and comprehensive analytic system that addressed the micro- and macro-linguistic aspects of oral narratives in persons with aphasia (PWA). Methods: Using a subgroup of unimpaired native speakers of English from the public corpus of AphasiaBank, task-specific normative data of three narrative tasks were constructed. Scoring criteria on the production of events, use of corresponding lexical items in these events, and the order of sequence of presenting the events were also developed. Twelve PWA were recruited and their spoken discourse was quantified using this newly proposed analytic system.Results: Significant correlations were found between PWA's performance on formal aphasia batteries and selected indices on the number of events produced as well as informative words used. Fluent and non-fluent PWA performed significantly differently in terms of their use of informative words and event sequence. The analytic system also demonstrated excellent intra- and inter-rater reliabilities. Conclusions: The clinical feasibility and diagnostic values of this proposed approach to quantify PWA's spoken output are confirmed. The present investigation also offered more evidence in supporting the needed psychometric properties of spoken discourse analysis that can supplement traditional formal aphasia batteries.