?
Phonetic clustering characteristics in verbal fluency: A potential marker for differentiating subjective cognitive decline from mild cognitive impairment
Objective
Semantic and phonemic verbal fluency (VF) tasks are widely used to assess older adults’ cognition in clinical practice. Typical scoring only analyses the total number of correct words produced. We investigated whether differentiation between individuals with subjective cognitive decline (SCD) versus mild cognitive impairment (MCI), which is often challenging, could be enhanced by also assessing linguistic (clustering) characteristics of VF responses.
Method
In this cross-sectional study, we retrospectively analyzed 426 VF responses from 127 community-dwelling older adults with SCD or MCI who underwent cognitive assessment in a memory clinic setting. Using mixed-effect models, we tested whether including linguistic (clustering) characteristics in addition to the total number of correct words enhanced the prediction of the Montreal Cognitive Assessment (MoCA) score and clinical group (SCD or MCI), when adjusted for age, sex, education level, and repeated measures alongside with task type.
Results
A lower mean phonetic cluster size in both SVF and PVF was associated with a lower MoCA score. Adding clustering characteristics to the model of SVF (although not PVF) significantly improved classification into SCD versus MCI, compared to the model with the number of correct words alone.
Conclusions
Retrieving sequences of words based on their phonetic proximity while performing not only PVF but also SVF tasks pointed to more preserved cognitive functioning and appeared vulnerable to early cognitive changes. Linguistic analysis of VF performance can capture subtle cognitive reorganization missed by scoring of the total number of correct words only and may enhance early dementia-risk profiling.