Hand preference and the corpus callosum: Is there really no association?

Abstract

Originating from a series of morphometric studies conducted in the 1980s, it appears a widely held belief in cognitive neuroscience that the corpus callosum is larger in left or mixed handers than in right handers (RH). However, a recent meta-analysis challenges this belief by not finding significant differences in corpus callosum size between handedness groups. Yet, relying on the available published data, the meta-analysis was not able to account for a series of factors potential influencing its outcome, such as confounding effects of brain size differences and a restricted spatial resolution of previous callosal segmentation strategies. To address these remaining questions, we here analysed N = 1057 participants' midsagittal corpus callosum of from the Human Connectome Project (HCP 1200 Young Adults) to compare handedness groups based on consistency (e.g., consistent RH vs. mixed handers, MH) and direction of hand preference (e.g., dominant RH vs. dominant left handers). A possible relevance of brain-size differences was addressed by analysing callosal variability by both using forebrain volume (FBV) as covariate and utilising relative area (callosal area/thickness divided by FBV) as a dependent variable. Callosal thickness was analysed at 100 measuring points along the structure to achieve high spatial resolution to detect subregional effects. However, neither of the conducted analyses was able to find significant handedness-related differences in the corpus callosum and the respective effect-sizes estimates were small. For example, comparing MH and consistent RH, the effect sizes for difference in callosal area were below a Cohen's d = 0.1 (irrespective of how FBV was included), and narrow confidence intervals allowed to exclude effects above |d| = 0.2. Analysing thickness, effect sizes were below d = 0.2 with confidence intervals not extending above |d| = 0.3. In this, the possible range of population effect sizes of hand preference on callosal morphology appears well below the effects commonly reported for factors like age, sex, or brain size. Effects on cognition or behaviour accordingly can be considered small, questioning the common practise to attribute performance differences between handedness groups to differences in callosal architecture.

Keywords: Brain asymmetry; Corpus callosum; Handedness; Human connectome project.

Fig. 1

Frequency distribution of the newly calculated EHI LQ (N = 1057). The LQ ranges from −100 (extreme preference for the left hand) to 100 (extreme preference of the right hand). The vertical dashed lines indicate the LQ = 80 cut-off used by Habib et al. (1991) to define consistent left- and right-hand preference, respectively.

Fig. 2

Segment-wise color-coded presentation of the effect size (Cohen's d) of the Handedness Group main effect for the consistency-based definition of handedness. The colour orange codes differences in favour of the MH group, while blue codes difference in favour of the cRH group. The location of the minimum and maximum effect size are marked by black dots on the respective segment. For the absolute thickness analysis (top), the value of minimum and maximum d are additionally provided together with their 95 confidence intervals. Of note, for neither of the analyses did the main effect survive FDR adjustment. For a small number of individual segments, uncorrected p-values were below 0.05 and respective segments are surrounded by a dotted line. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

Segment-wise presentation of the effect size of the Handedness Group main effect for the direction-based definition of handedness. Orange indicates mean deviations in favour of the LH group (dLH/cLH), while blue codes difference in favour of the RH group (dRH/cRH). Location of minimum/maximum d are marked with black dots. For the absolute thickness analysis (top), the minimum/maximum values (and 95% confidence limits) are additionally provided. For neither of the analyses did the main effects survive FDR adjustment. For some segments the uncorrected p-values were below 0.05. These are surrounded by a dotted line. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4

Segment-wise presentation of the effect size d of the main effect of Sex for the three analyses conducted in relation to comparison C. Orange codes thicker female, blue thicker male segments. Minimum/maximum d values (and 95% confidence limits) are additionally provided). Dark orange and dark blue indicate significant segments after FDR adjustment to 5%. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)