Research Article: Heritability of semantic verbal fluency task using time-interval analysis

Date Published: June 11, 2019

Publisher: Public Library of Science

Author(s): T. P. Taporoski, N. E. Duarte, S. Pompéia, A. Sterr, L. M. Gómez, R. O. Alvim, A. R. V. R. Horimoto, J. E. Krieger, H. Vallada, A. C. Pereira, M. von Schantz, A. B. Negrão, Daniel E. Gustavson.


Individual variability in word generation is a product of genetic and environmental influences. The genetic effects on semantic verbal fluency were estimated in 1,735 participants from the Brazilian Baependi Heart Study. The numbers of exemplars produced in 60 s were broken down into time quartiles because of the involvement of different cognitive processes—predominantly automatic at the beginning, controlled/executive at the end. Heritability in the unadjusted model for the 60-s measure was 0.32. The best-fit model contained age, sex, years of schooling, and time of day as covariates, giving a heritability of 0.21. Schooling had the highest moderating effect. The highest heritability (0.17) was observed in the first quartile, decreasing to 0.09, 0.12, and 0.0003 in the following ones. Heritability for average production starting point (intercept) was 0.18, indicating genetic influences for automatic cognitive processes. Production decay (slope), indicative of controlled processes, was not significant. The genetic influence on different quartiles of the semantic verbal fluency test could potentially be exploited in clinical practice and genome-wide association studies.

Partial Text

The historical roots of the semantic verbal fluency (SVF) task lie in tests requiring the production of words from a certain semantic category [1]. This task involves long-term memory retrieval according to the meaning of words, or semantic memories [2]. The most common category used in SVF tests is “animals”, a semantic category with minimal differences across countries, generations, and feasible across different education levels [3]. Nonetheless, previous SVF studies have demonstrated that demographic characteristics such as sex [4], age [5], time of day [6], and education [7, 8] can influence performance. The task is broadly used in clinical and non-clinical research [9], mainly because its sensitivity to several neurobiological conditions associated with cognitive impairment, such as Alzheimer’s disease [10], ageing [11], mild cognitive impairment [12], schizophrenia [13], and Parkinson’s disease [14] among others.

Performance in the SVF task in the Baependi sample was sensitive to demographic and environmental factors, as previously described in the literature from other locations [4, 5, 8]. In this study, unadjusted heritability estimates for 60-s score increased with the addition of sex and age, and decreased when years of schooling was considered as a covariate. Although assessments at later times of day correlated positively with performance as found before [6], they did not change heritability estimates. The best fitting heritability model included age, sex, schooling and the time of day as covariates, but the biggest impact on performance was schooling.