|Abstract (english)|| |
INTRODUCTION: Traditionally, verbal fluency research has differentiated between semantic (SF) and letter fluency (LF). However, although linguistic research has strongly suggested that there are pronounced differences in the processing of different types of semantic categories, most researchers in this domain have uncritically assumed that there are no category-specific effects in verbal fluency. Limited research on SF has thus far revealed that disproportionate performances across different categories are in certain contexts associated with category size, while clinical populations might show normative deficits in one category, but not the other. Nevertheless, this issue remains poorly investigated, particularly in LF.
METHODS: We recruited 16 Croatian-speaking students from the University of Zagreb. For verbal fluency, we administered the categories animals and trees (SF), and the letters K and M (LF) (60 s). Participants further performed on a neuropsychological battery encompassing the Trail Making Test (TMT), Wisconsin Card Sorting Test (WCST), digit span, and lexical decision (LD). Paired-sample t-tests, and Spearman correlation coefficients and regression analyses were used for statistical analyses.
RESULTS: On the animal compared to the tree task participants produced more words, had lower intrusion rates, produced larger clusters, and were faster in inter-cluster transitions. Furthermore, while the total production of animals was positively predicted by the number of clusters, in trees it was positively predicted by cluster size. A negative relationship between performance on TMT and speed of inter-cluster transitions was found in trees, and a positive relationship between performance on WCST and the same fluency variable in animals. Higher digit span was associated with lower clustering rates in animals, but with higher clustering rates in trees, while accuracy on abstract LD trials was positively associated with total production on trees. Analyses of LF revealed that participants produced more words and were faster in inter-cluster transitions on the K compared to the M task. Total production in K was predicted by cluster size, but in M it was predicted by switching. Accuracy on TMT negatively predicted total production and the number of clusters for M. Also, performance on WCST was positively associated with clustering rates and inter-cluster transition speed on M.
CONCLUSIONS: Our results indicate that there are important differences in the phenomena and processes underlying performance on different SF and LF tasks. While the differences in the two SF tasks might be partly explained by presumed differences in category size, higher rates of intrusions in the tree task suggest that the boundaries of the category trees are less fixed compared to the category animals. Additionally, disproportionate relationships with neuropsychological variables suggest that clustering in trees might rely more on working memory and abstract semantic processes. Regarding LF, we found that performance on K relies more on clustering compared to M, while the neuropsychological data suggested that performance on the M task was inversely associated with visual attention, possibly suggesting that visuospatial imagery strategies might hinder optimal performance. Also, data suggest that clustering on the M, but not the K task is an executively relatively demanding process.