Genetic and Environmental Influences on Verbal Fluency in Middle Age: A New Framework

Neuropsychological measures of verbal fluency are some of the most widely used tests for detecting age-related cognitive decline and dementia. Although substantial evidence suggests that phonemic and semantic fluency measures... [ view full abstract ]

Neuropsychological measures of verbal fluency are some of the most widely used tests for detecting age-related cognitive decline and dementia. Although substantial evidence suggests that phonemic and semantic fluency measures are differentially associated with neuropsychiatric conditions and other cognitive functions, few studies have examined the shared/unique variance between these aspects of fluency, especially using genetically-informative studies. We present a novel two-factor model of verbal fluency, describe its genetic/environmental architecture, and evaluate multiple hypotheses regarding the relevance of these fluency abilities to cognition and aging.

1464 middle-aged male twins from the Vietnam Era Twin Study of Aging (VETSA) completed letter (F-A-S) and category (animals, boys’ names, fruits/furniture) fluency tests at up to two time-points (mean age 56 at baseline, and 62 years at follow-up).

The best-fitting model at both waves comprised two factors. A General Fluency latent factor predicted variation in all letter and category subtests and a Semantic-Specific latent factor accounted for additional variation in category fluency not already captured by the common factor. Both factors were explained primarily by genetic influences at both waves (a² = .56 to .76) as well as nonshared environmental influences (e² = .24 to .44), but not at all by shared environmental influences (c² = .00). There was considerable stability of individual differences over six years (r = .90 for General Fluency, r = .81 for Semantic-Specific), especially for genetic influences (r_g = .99 and 1.0, respectively). Additionally, there was evidence for mean-level decline in General Fluency (d = -.22) but not Semantic-Specific Fluency (d = -.01). Further analyses of the first wave data revealed that the General Fluency factor was associated with vocabulary (β = .26), executive functions (β = .24), and working memory (β = .23), whereas the Semantic-Specific factor was strongly associated with episodic memory (β = .60). The Semantic-Specific factor at age 56 also predicted change of variance in episodic memory at age 62 and predicted conversion to amnestic mild cognitive impairment (OR = 2.00) controlling for episodic memory.

The results provide a new framework for viewing verbal fluency as a combination of general and semantic-specific processes, both of which have unique genetic underpinnings and differential relations with other cognitive variables. Semantic fluency is particularly impaired in Alzheimer’s disease. Therefore, the role of the heritable Semantic-Specific factor may be particularly important in future efforts to understand transitions to mild cognitive impairment and Alzheimer’s disease.