| Sumario: | Purpose: Although linguistic features in connected speech provide a valuable tool for identifying mild cognitive impairment (MCI), heterogeneous measures and tasks in MCI discourse studies have yielded widely varying outcomes, and the diversity of languages studied is highly limited. This study aims to identify effective measures and tasks to differentiate Chinese-speaking older adults with and without MCI, and to explore task-related differences in connected speech. Method: Speech samples from 30 Chinese-speaking older adults with MCI and 30 healthy controls (HCs) were collected using the single picture, picture sequence, and story retelling tasks. Lexical and semantic measures, including parts of speech, lexical diversity, informativeness, conciseness, and efficiency, were analyzed using univariate 2 × 3 mixed ANOVAs with Bonferroni corrections. Results: MCI participants produced significantly fewer information units and had lower success rates in generating information units across all three tasks than HCs. There were no significant differences in lexical measures between the MCI and HC groups in the picture sequence task. In the story retelling task, MCI participants showed significantly lower lexical efficiency than HCs. When comparing tasks, the story retelling task was associated with the lowest lexical diversity, information efficiency and density, and success rate of information unit produced. Speech samples elicited by the picture sequence task showed higher lexical efficiency but lower semantic richness than the single picture task, with no significant differences in the success rate of information units produced. Conclusions: Cognitive ability and task type both influence discourse performance. Regarding effective measures and tasks, fewer information units and a lower success rate of information units produced are reliable MCI indicators in each task, while lexical measures are less sensitive than semantic measures. The single picture task shows strong validity in identifying MCI, regardless of lexical or semantic measures employed. Regarding differences among tasks, the results suggest that the story retelling task is more demanding, and single picture and picture sequence tasks are better suited for analyzing lexical richness and semantic deficits of MCI patients. This study enhances the understanding of MCI in Chinese-speaking individuals and highlights the task-related differences in connected speech.
|
|---|