Verbal fluency in Parkinson's disease with mild cognitive impairment: quantitative, qualitative, and semantic network analyses.

BACKGROUND

Verbal fluency (VF) tasks are widely used in the cognitive assessment of Parkinson's disease (PD). However, the total number of correct responses does not capture all aspects of performance, and the underlying cognitive mechanisms remain unclear.

OBJECTIVE

To compare semantic and phonemic VF performance in PD patients with mild cognitive impairment (PD-MCI) and those without cognitive impairment (PD-nCI).

METHODS

Eighty participants with PD (50 % PD-MCI), no differences by sex, age, or education, completed a comprehensive neuropsychological battery and two verbal fluency tasks. The number of correct responses (CR), clusters, switches, mean cluster size, intrusions, and perseverations were compared between groups. Linear regression analyses were conducted to identify neuropsychological predictors of each VF measure. In addition, two semantic network maps were created for each group and compared both quantitatively and qualitatively.

RESULTS

The PD-nCI group outperformed the PD-MCI group in CR on both tasks, as well as in the number of clusters and switches in semantic fluency, with large effect sizes. Episodic verbal memory, naming, and attention predicted semantic fluency performance, while attention, working memory, and executive functions predicted phonemic fluency. Semantic network analysis revealed significant differences in average shortest path length and clustering coefficient, with higher values in PD-nCI, whereas modularity coefficient did not differ. The PD-nCI network showed more numerous and clearly defined communities.

CONCLUSION

Participants with PD-MCI showed greater impairment in semantic fluency, with fewer clusters and switches. Semantic network analysis revealed alterations in the network structure of PD-MCI, indicating reduced local semantic cohesion.