Abstract
•
Purpose:
This study aims to establish objective verb clustering criteria in action verbal fluency (VF) using hierarchical clustering analysis (HCA) based on the Lancaster sensorimotor norms. We explored age-related differences in switching and cluster diversity (word retrieval strategies) between younger and older adults and correlations among age, education, working memory (WM) capacity, action VF performance, and word retrieval strategies for each group.
•
Method:
Sixty-two native Korean speakers participated in the study, with 31 young adults (Mage = 27.39) and 31 older adults (Mage = 70.45). Participants completed a 1-min action VF task, and generated verbs were classified into 15 clusters based on the Lancaster sensorimotor ratings using HCA. We analyzed switching (shifts between clusters) and cluster diversity (number of unique verb clusters) to assess word retrieval strategies. WM capacity was measured through word-forward and word-backward (WB) tasks.
•
Results:
Older adults demonstrated significantly fewer switchings and lower cluster diversity compared to younger adults, indicating restricted word retrieval strategies. WB task scores in older adults positively correlated with word retrieval strategies and action VF performance. Older individuals with younger age and higher education employed more word retrieval strategies. Both groups exhibited positive correlations between word retrieval strategies and action VF performance.
•
Conclusions:
The novel HCA approach, based on the Lancaster sensorimotor norms, successfully classified verb clusters that revealed aging-related differences in word retrieval strategies and the relationship between WM and these strategies. These results highlight the potential of using the HCA method for verb clustering analyses, particularly in providing qualitative insights into action VF tasks.
Keywords
•
Action Verbal Fluency
•
Hierarchical Clustering Analysis
•
Older Adults
Major Figures
•
Figure 1. Colored dendrogram of verb clusters derived from hierarchical clustering analysis with cluster’s ID (right) * Excluded verbs based on the subjective survey results
•
Figure 2. Example of calculating the number of switchings and verb clusters from a participant’s performance
•
Figure 3. The number of switchings (a) and cluster diversity (b) between the groups. *p < .05 , **p < .001
•
Figure 4. The visualization of nonlinear multidimensional scaling among variables for each group
