Page 95 - Like me, or else... - Michelle Achterberg

Heritability of aggression following social evaluation
93
Phenotypic similarities among twin pairs can be divided into similarities due to shared genetic factors (A) and shared environmental factors (C), while dissimilarities are ascribed to unique environmental influences and measurement error (E). We used behavioral genetic modeling with the OpenMX package (Neale et al., 2016) in R (R Core Team, 2015) to get an estimate of these A, C, and E components. Comparisons of the ACE model with more parsimonious models (AE model; CE model; or E model) are described in the Supplementary Materials. When ACE models show the best fit, both heritability, shared and unique environment are important contributors to explain the variance in the outcome variable. AE models indicate that genetic and unique environmental factors play a role; whilst CE models indicate influences of the shared environment and unique environment. If the E model has no worse fit than AE or CE models, variance in the outcome variable is accounted for by unique environmental factors and measurement error.
Statistical Analysis
In order to detect outliers in the data, we transformed the raw data to z-values. Based on the Z-distribution, 99.9% of z-scores lie between -3.29 and +3.29. Z- values outside this range (<-3.29 or >3.29) were defined as outliers. Outlying scores were winsorized (Tabachnick and Fidell, 2013). To assess effects of condition (positive, neutral, negative) on noise blast duration (in ms) we used a linear mixed-effect model approach using the lme4 package in R (Bates et al., 2015) in R (R Core Team, 2015). Data was fitted on the average response times after positive, neutral and negative trials. Random intercepts per participants and per family allows to account for the nesting of condition within participant (ChildID) and the nesting of twin-pairs within families (FamilyID). Additionally, a random slope of condition was included per participant. Fixed effects included condition (factor with 3 levels), as well as participant’s age and IQ as covariates, which were grand mean centered. All main effects and two-way interactions between age х condition and age x IQ were included. P-values were determined using Kenward-Rogers approximation as implemented in the mixed function in the afex package (Singmann, 2013). The fitted mixed-effect model is specified in
R as: !"#$%&'($) ~ ,"!-#)#"! ∗ (/%_1%(!,%!)%2%- + ,"!-#)#"! ∗ 45_1%(!,%!)%2%- + (,"!-#)#"!|,h#'-49) + (1|<(1#'=49).
To derive a measure of individual differences in aggression we calculated the differences in noise blast duration between conditions (negative-positive; negative-neutral; neutral-positive). Brain-behavior associations were investigated by least square regressions with ROI activation predicting noise blast difference scores. Due to the nested nature of twin data, the data violates the assumption
4