The Contribution of the Corpus Callosum to Language Lateralization

Abstract

The development of hemispheric lateralization for language is poorly understood. In one hypothesis, early asymmetric gene expression assigns language to the left hemisphere. In an alternate view, language is represented a priori in both hemispheres and lateralization emerges via cross-hemispheric communication through the corpus callosum. To address this second hypothesis, we capitalized on the high temporal and spatial resolution of magnetoencephalographic imaging to measure cortical activity during language processing, speech preparation, and speech execution in 25 participants with agenesis of the corpus callosum (AgCC) and 21 matched neurotypical individuals. In contrast to strongly lateralized left hemisphere activations for language in neurotypical controls, participants with complete or partial AgCC exhibited bilateral hemispheric activations in both auditory or visually driven language tasks, with complete AgCC participants showing significantly more right hemisphere activations than controls or than individuals with partial AgCC. In AgCC individuals, language laterality positively correlated with verbal IQ. These findings suggest that the corpus callosum helps to drive language lateralization.

Significance statement: The role that corpus callosum development has on the hemispheric specialization of language is poorly understood. Here, we used magnetoencephalographic imaging during linguistic tests (verb generation, picture naming) to test for hemispheric dominance in patients with agenesis of the corpus callosum (AgCC) and found reduced laterality (i.e., greater likelihood of bilaterality or right hemisphere dominance) in this cohort compared with controls, especially in patients with complete agenesis. Laterality was positively correlated with behavioral measures of verbal intelligence. These findings provide support for the hypothesis that the callosum aids in functional specialization throughout neural development and that the loss of this mechanism correlates with impairments in verbal performance.

Keywords: corpus callosum; hemispheric specialization; language; magnetoencephalography.

Copyright © 2016 the authors 0270-6474/16/364522-12$15.00/0.

Figures

Figure 1.

Stimulus-locked (0 ms = auditory noun onset) group analyses of changes in beta (12–30 Hz) oscillatory power during auditory verb generation. A, Changes in beta power over the left hemisphere (one-sample t test). Robust reductions in beta power are seen over the left hemisphere in both the control and AgCC cohorts. B, Changes in beta power over the right hemisphere. Changes in frontal and temporal beta power are present only in AgCC. C, Comparison (unpaired nonparametric t test) between the control and AgCC groups. Significant increases in activity (in blue) over frontal and temporal regions are present in only the right hemisphere. Minor decreases in activity (in red) are present in the posterior regions of the temporal lobe and occipital lobe. Abbreviations are as in Table 5. All images are statistically thresholded and superimposed on a MRI template brain using MRICro.

Figure 2.

Time course plots of beta power (in dB) for regions of interest derived from the stimulus-locked group comparison in Figure 1C. Green, NC; red, pAgCC; blue, cAgCC. Error bars indicate SEM.

Figure 3.

Response-locked (0 ms = speak verb) group analyses of changes in beta (12–30 Hz) oscillatory power during auditory verb generation. A, Changes in beta power over the left hemisphere (one-sample t test). Robust reductions in beta power are seen over the left hemisphere in both the control and AgCC cohorts before response generation and during execution. B, Changes in beta power over the right hemisphere. Changes in frontal and temporal beta power before the response are present only in AgCC, with bilateral changes in both groups after 0 ms. C, Comparison (unpaired nonparametric t test) between the control and AgCC groups. Significant increases in beta activity (blue) before response onset are present only in the right hemisphere, with greater activity in AgCC. Decreases in beta activity for the AgCC cohort (red) are present in the posterior regions of the left hemisphere outside of the language network. Dashed vertical line indicates 0 ms. Conventions are as in Figure 1.

Figure 4.

Time course plots of beta power (in dB) for regions of interest derived from the response-locked group comparison in Figure 3C. Conventions as in Figure 2.

Figure 5.

Response-locked group analyses of changes in beta (12–30 Hz) oscillatory power during picture naming. A, Changes in beta power over the left hemisphere (one-sample t test). Like auditory verb generation, robust reductions in beta power over frontal regions are seen over the left hemisphere in both the control and AgCC cohorts (A), whereas only changes in beta power are seen in the right hemisphere preresponse in AgCC (B). C, Comparison (unpaired nonparametric t test) between the control and AgCC group showing that increased beta activity is present in the right frontal regions in AgCC. Conventions are as in Figure 3.

Figure 6.

Time course plots of beta power (in dB) for regions of interest derived from the response-locked group comparison in Figure 5C. Conventions as in Figure 2.

Figure 7.

LI in AgCC. A, Scatterplot of LI derived from the auditory verb generation task in the NC (green), pAgCC (red), and cAgCC (blue) groups. Significantly (*p < 0.05) lower LI was identified in the cAgCC group compared with the NC group, indicating a greater likelihood of right hemisphere lateralization for language. Black (AgCC) and colored horizontal bars indicate the group mean. B, Correlation between LI and VIQ for the NC (green) and AgCC (black) groups. A significant (r = 0.55, p < 0.005) positive correlation was identified in the AgCC group, but not in the NC group (r = 0.25, p > 0.05).