Discordant cerebral lateralisation for verbal fluency is not an artefact of attention: evidence from MzHd twins

Abstract

The current study investigated how the brain structure–function relationships between covert and overt verbal fluency in monozygotic handedness discordant (MzHd) twins relates to broader attentional measures. Evidence presented here shows that the structure–function correlation demonstrated between functional cerebral lateralisation and corpus callosum widths 22–39, contiguous with Broca’s area and the middle frontal gyri (Gurd et al. in Brain Struct Funct 218:491–509, 2013), is not an artefact of attention. Twenty-five pairs of female MzHd twins performed a verbal switching task titrated for number of switches. The paradigm permits calculation of switch rates and costs. The switch rate and cost varied as a function of number of switches (1, 2, 3). There were no differences in any measure in relation to right–left handedness in twin pairs. This was supported by large and significant within-twin pair correlations. Atypical functional lateralisation of inferior and middle frontal lobes does not appear to be associated with better or worse performance on verbal task-switching. Discordant lateralisation for verbal fluency does not predict behavioural performance profiles in MzHd twins. This evidence is compatible with a view that attentional components of verbal fluency task performance may have significant heritable components. It does not indicate that neural correlates of frontal cerebral laterality for verbal fluency in MzHd twins are significantly accountable for by heritable components.

Appendices

Appendix 1

Differences between RH and LH twins were computed for 2-, 3-, and 4-category switch rate and switch cost measures. Summary statistics are listed in Tables 3 and 4. Figures plotting the mean within-twin pair differences ± 95 % confidence intervals are plotted for switch rate (Fig. 3) and switch cost (Fig. 4). The figures show mean differences near zero, with confidence intervals spanning the zero line for all measures.

Table 3 Summary statistics for differences in switch rate for RH–LH twins

Table 4 Summary statistics for differences in switch cost for RH–LH twins

Fig. 3

Mean differences between RH–LH twin pairs in switch rate for 2-, 3-, and 4-category switch (±95 % CIs) (CS2 2-category, CS3 3-category, CS4 4-category; LH left handed; RH right handed)

Fig. 4

Mean differences between RH–LH twin pairs in switch cost for 2-, 3-, and 4-category switch (±95 % CIs). (CS2 2-category, CS3 3-category, CS4 4-category; LH left handed; RH right handed)

Appendix 2

Errors: Means (SD) per group and per condition were few in numbers and comparable in proportion, per number of categories between handedness groups:

$$\begin{gathered} {\text{RH: 2-category}} = 0.00 \, \left( {0.00} \right), \, {\text{3-category}} = 0.23 \, \left( {0.65} \right),{\text{ 4-category}} = 0.64 \, \left( {1.09} \right) \hfill \\ {\text{LH: 2-category}} = 0.04 \, \left( {0.20} \right), \, {\text{3-category}} = 0.54 \, \left( {0.91} \right),{\text{ 4-category}} = 0.75 \, \left( {1.15} \right) \hfill \\ \end{gathered}$$

ANOVA was conducted with Hand as the group measure, and switch error as the dependent measure. Effects of Hand (RH, LH) were not significant (F = 1.60, df = 1,143, p = 0.21, ns) (see Fig. 5).

Fig. 5

Dotplot of mean total errors per individual per group (RH, LH)

Error type: Errors were analysed as to type according to a simplified procedure described in Gurd (1993, unpublished DPhil thesis), and reported as overall percentage of errors per handedness group, and were produced in comparable proportions:

$$\begin{aligned} {\text{RH:}} & {\text{ Within category sequencing errors: }}13/21 = 61.90\;\% {\text{ of RH errors}} \\ & {\text{Between category sequencing errors: }}6/21 = 28.57\;\% {\text{ of RH errors}} \\ & {\text{Other errors: }}2/21 = 9.52\;\% {\text{ of RH errors}} \\ {\text{LH:}} & {\text{ Within category sequencing errors: }}26/34 = 76.47\;\% {\text{ of LH errors}} \\ & {\text{Between category sequencing errors: }}6/34 = 17.65 \;\% {\text{ of LH errors}} \\ & {\text{Other errors: }}2/34 = 5.88 \;\% {\text{ of LH errors}} .\\ \end{aligned}$$

Notes

Complete data sets (i.e. full runs) were available for 23 of the 25 twin pairs for analysis of switch rate (missing data for: RH_twin25 and LH_twin15 for switch rate 4). For switch cost analyses, 22 pairs had complete data sets (missing data for: RH_twin4, LH_twin15, RH_twin25, LH_twin25 and for switch cost 4). (Results from one further twin pair (tw_16) has already been reported (Lux et al. 2008). Their pattern of fMRI activations were unique and different, hence their data are not reproduced here.)