Genetic architecture of the Delis-Kaplan Executive Function System Trail Making Test: evidence for distinct genetic influences on executive function

Abstract

Objective: To examine how genes and environments contribute to relationships among Trail Making Test (TMT) conditions and the extent to which these conditions have unique genetic and environmental influences.

Method: Participants included 1,237 middle-aged male twins from the Vietnam Era Twin Study of Aging. The Delis-Kaplan Executive Function System TMT included visual searching, number and letter sequencing, and set-shifting components.

Results: Phenotypic correlations among TMT conditions ranged from 0.29 to 0.60, and genes accounted for the majority (58-84%) of each correlation. Overall heritability ranged from 0.34 to 0.62 across conditions. Phenotypic factor analysis suggested a single factor. In contrast, genetic models revealed a single common genetic factor but also unique genetic influences separate from the common factor. Genetic variance (i.e., heritability) of number and letter sequencing was completely explained by the common genetic factor while unique genetic influences separate from the common factor accounted for 57% and 21% of the heritabilities of visual search and set shifting, respectively. After accounting for general cognitive ability, unique genetic influences accounted for 64% and 31% of those heritabilities.

Conclusion: A common genetic factor, most likely representing a combination of speed and sequencing, accounted for most of the correlation among TMT 1-4. Distinct genetic factors, however, accounted for a portion of variance in visual scanning and set shifting. Thus, although traditional phenotypic shared variance analysis techniques suggest only one general factor underlying different neuropsychological functions in nonpatient populations, examining the genetic underpinnings of cognitive processes with twin analysis can uncover more complex etiological processes.

Figure 1

Multivariate Genetic Models. Only one twin is represented in each figure. To note, for space reasons, only A (additive genetic) and E (non-shared environment) pathways are depicted. Figure 1A (Cholesky model), depicts the atheoretical, fully parameterized variance/covariance structure among Trails 1–4, including all source of variance: additive genetic influences (A₁–A₄) and non-shared environmental influences (E₁–E₄). Figures 1B (Independent Pathways Model), 1C (Common Pathways Model) and 1D (Measurement Model) depict theoretical models for the covariance structure among Trails 1–4. A_C=Additive genetic influences and E_C=Nonshared environmental influences that are common to all Trails conditions. A_S=Additive genetic influences and E_S=Nonshared environmental influences that are specific to each Trails condition.

Figure 2

Most parsimonious independent pathway model for the covariation among D-KEFS Trail Making Test conditions, for Trails values that were a) not adjusted for general cognitive ability and b) adjusted for general cognitive ability. A_C = common genetic factor; E_C = common non-shared environmental factor; A_S1 = specific genetic factor for visual searching condition; A_S2 = specific genetic factor for set shifting; E_S1–4 = specific non-shared environmental factors on each condition.