Total output and switching in category fluency successfully discriminates Alzheimer's disease from Mild Cognitive Impairment, but not from frontotemporal dementia

Verbal fluency tasks require generation of words beginning with a letter (phonemic fluency; PF) or from a category (category fluency; CF) within a limited time period. Generally, total output on CF has been used to discriminate Mild Cognitive Impairment (MCI) from Alzheimer's disease (AD), while poor PF has been used as a marker for behavioral-variant frontotemporal dementia (bvFTD). However, in the absence of this disparate performance, further characterization of the task becomes necessary. OBJECTIVE: We examined whether fluency, as well as its components, clustering (successively generated words belonging to a category) and switching (shifting between categories) carried diagnostic utility in discriminating AD from MCI and bvFTD. METHODS: PF (letter 'P') and CF ('animals') tasks were administered in English to patients with MCI (n=25), AD (n=37), and bvFTD (n=17). Clustering and switching scores were calculated using established criteria. RESULTS: Our findings suggested that up to 85% of AD and MCI could be successfully discriminated based on total number of responses and switching in CF alone. PF-CF disparity was not noted in AD or bvFTD. Performance on clustering or switching also proved insufficient to discriminate AD from bvFTD. CONCLUSION: Switching was found to be useful when differentiating AD from MCI. In AD and bvFTD, the course of progression of the disease may lead to attenuation of total number of responses produced on both tasks to an extent where clustering and switching may not be useful measures to discriminate these dementias from each other.