A cross-sectional study of category fluency serial recall order in Italian individuals with amnestic mild cognitive impairment from the TREDEM registry: The role of word length

Abstract

Category Fluency (CF) “item-level” scoring methodologies identify neurocognitive processes that are distinct from those of standard quantitative scoring. One of these accounts for words’ serial recall order (SRO) and calculates the correlation between SRO and item-level complexity to operationalise the tendency to name increasingly complex words. We hypothesised this tendency would be altered in individuals with amnestic mild cognitive impairment (aMCI), to reflect lexical-semantic organisational decline resulting from potential functional alterations of perirhinal cortex. We expected aMCI individuals to show higher correlational indices than people with no objective cognitive decline (NOCD) due to these trends being more influenced by word count than overall complexity. Sixty aMCI and 54 NOCD individuals, recruited by the TREDEM initiative and representative of the North-East Italian population were administered a 1-min “animals” CF. Four complexity features were scored: frequency, age-of-acquisition, typicality and graphemic length. Individual z-converted correlation coefficients were calculated between SRO and complexity. The two groups were then compared via uncorrected t-tests, age- and education-corrected ANCOVAs, and further ANCOVAs also correcting for quantitative CF scores. aMCI individuals named fewer words, and words of lower and more variable complexity. They also showed a significantly stronger “SRO-graphemic length” correlation. This emerged from the uncorrected models and from the first set of ANCOVAs. Significant differences in lexical-semantic organisation exist in aMCI. Length is a diagnostically relevant feature of Italian words generated as part of CF. In line with cross-linguistic generalisability of findings, potential mechanisms differentiating phonological and orthographic length in languages with transparent/opaque orthographies are discussed.