Functional anatomy of interhemispheric cortical connections in the human brain

Abstract

Interhemsipheric interaction between the human cerebral hemispheres is served by abundant white-matter fibres in the human corpus callosum (CC). Damage to these fibres has notable behavioural and cognitive sequelae that depend on the exact location of the fibre loss. Until now, correlations between fibre loss and neurological disorders have been limited to post-mortem studies. Here we used probabilistic diffusion magnetic resonance imaging tractography to produce a two-dimensional map of the CC in the mid-sagittal plane. We observed an antero-posterior topography of interhemispheric tracts within the CC, consistent with our current neuroanatomical understanding of post-mortem studies in human. Callosal tract to the left and right hemispheres had comparable volume. Gender, a factor that is often reported to affect CC shape and geometry, also had no effect on the volume of the tracts. Our map showed high consistency across individuals. We propose that this map might be useful in the study of the effects of damage to human CC in neurodegenerative and cognitive disorders.

Figures

Fig 1

Mid-sagittal group probability maps showing the location of cortical connections within the corpus callosum (CC). The organization of connections was very similar for cortical targets in the left and right hemispheres: results for tracking to cortical targets in the left hemisphere are shown in the left column, and the right hemisphere in the right column. The colour scale represents the population probability of a given CC voxel connecting to a particular target from dark red (10%) to white (100%). The map shows clear antero-posterior topographic organization. Prefrontal cortical connections occupied a large part of the volume of the CC. Abbreviations: PFC, prefrontal; M1, primary motor; S1, primary somatosensory; PPC, posterior parietal; PMC, premotor; Occ, occipital; Tmp, temporal.

Fig 2

Map of the corpus callosum according to the connectivity of voxels from the mid-sagittal section to seven cortical regions in each hemisphere. Results for tracking to targets in the right hemisphere are shown at the top, and for the left hemisphere at the bottom. The map demonstrates the overlap between callosal regions projecting to adjacent cortical target areas. Borders between regions are derived by thresholding the population probability maps at a probability of 30%. Abbreviations as for Fig. 1.

Fig 3

Hard segmentation maps for each individual studied showing a high reproducibility and consistency across subjects and across hemispheres. Hard segmentations are created by classifying each callosal voxel according to the cortical target with which it has the highest probability of connection. Abbreviations as for Fig. 1.

Fig 4

Absolute (a) and relative (b) volume of callosal voxels connected to the cortical target regions. The figures show a significant effect of cortical region for both measures (absolute: F = 123.5, P < 0.001; relative: F = 36.4, P < 0.001). No effect of hemisphere was found for absolute volume, but a trend was found for the relative volumes, reflecting the tendency for greater CC volumes to be found for tracking to left vs. right hemisphere targets (F = 4.9, P = 0.05). Abbreviations as for Fig. 1.