Connectivity-Based Topographical Changes of the Corpus Callosum During Aging

Abstract

Background: The corpus callosum (CC) is the most prominent white matter connection for interhemispheric information transfer. It is implicated in a variety of cognitive functions, which tend to decline with age. The region-specific projections of the fiber bundles with microstructural heterogeneity of the CC are associated with cognitive functions and diseases. However, how the CC is associated with the information transfer within functional networks and the connectivity changes during aging remain unclear. Studying the CC topography helps to understand the functional specialization and age-related changes of CC subregions. Methods: Diffusion tractography was used to subdivide the CC into seven subregions from 1,086 healthy volunteers within a wide age range (21-90 years), based on the connections to the cortical parcellations of the functional networks. Quantitative diffusion indices and connection probability were calculated to study the microstructure differences and age-related changes in the CC subregions. Results: According to the population-based probabilistic topography of the CC, part of the default mode network (DMN) and limbic network (LN) projected fibers through the genu and rostrum; the frontoparietal network (FPN), ventral attention network (VA) and somatomotor networks (SM) were interconnected by the CC body; callosal fibers arising from the part of the default mode network (DMN), dorsal attention network (DA) and visual network (VIS) passed through the splenium. Anterior CC subregions interconnecting DMN, LN, FPN, VA, and SM showed lower fractional anisotropy (FA) and higher mean diffusivity (MD) and radial diffusivity (RD) than posterior CC subregions interconnecting DA and VIS. All the CC subregions showed slightly increasing FA and decreasing MD, RD, and axial diffusivity (AD) at younger ages and opposite trends at older ages. Besides, the anterior CC subregions exhibited larger microstructural and connectivity changes compared with the posterior CC subregions during aging. Conclusion: This study revealed the callosal subregions related to functional networks and uncovered an overall "anterior-to-posterior" region-specific changing trend during aging, which provides a baseline to identify the presence and timing of callosal connection states.

Keywords: aging trajectory; atlas; diffusion MRI; functional networks; segmentation; tractography.

FIGURE 1

The probabilistic connection maps of 7 CC subregions. (A) The atlas of 7 functional networks with color-coding; (B) Hard segmentation map of the highest connection probability to corresponding functional networks with the same color-coding of panel (A); (C) The probabilistic maps of the midsagittal CC, showing P(v,n) in each CC subregion.

FIGURE 2

The diffusion indices of 7 CC subregions. The mean and standard deviation of the regressed DI(s,n) of all participants are plotted in the order of the spatial locations of all CC subregions, generally from the anterior to posterior part in the CC. RM-ANOVA showed significant differences in FA, MD, AD and RD between the CC subregions. FA values of the CC subregion connecting to VIS and DA are statistically higher than other subregions. MD and RD values of the CC subregions connecting to VIS and DA are statistically lower than those of other subregions. See Supplementary Table 1 for the details of the statistical analysis of paired comparisons. VIS, CC subregion connecting to the visual network; SM, CC subregion connecting to the somatomotor network; DA, CC subregion connecting to the dorsal attention network; VA, CC subregion connecting to the ventral attention network; LN, CC subregion connecting to the limbic network; FPN, CC subregion connecting to the frontoparietal network; DMN, CC subregion connecting to the default mode network; FA, fractional anisotropy; MD, mean diffusivity; RD, radial diffusivity; AD, axial diffusivity.

FIGURE 3

The connection probability of 7 CC subregions across all 7 age groups. The mean and standard deviation of the regressed P(s,n) of each age group are plotted in the order of the spatial locations of all CC subregions, generally from the anterior to posterior part in the CC. ANCOVA showed significant differences among the 7 age groups for the subregions to DMN, FPN, VA, SM, DA, and VIS, which exhibited an anterior-to-posterior changing gradient. The subregions to DMN, FPN, VA and SM that are located in the anterior and middle part of the CC showed an earlier and severer decrease in the connection probability than the subregion to DA and VIS in the posterior CC. Statistical analysis of the paired comparisons is shown in Supplementary Table 2. VIS, CC subregion connecting to the visual network; SM, CC subregion connecting to the somatomotor network; DA, CC subregion connecting to the dorsal attention network; VA, CC subregion connecting to the ventral attention network; LN, CC subregion connecting to the limbic network; FPN, CC subregion connecting to the frontoparietal network; DMN, CC subregion connecting to the default mode network.

FIGURE 4

Best-fitting curves for regressed diffusion indices vs. age. Ages of peak FA values and minimum MD, RD and AD values from the quadratic fits are labeled by colored squares for all the subregions. Fit curves of all diffusion indices in the subregion to VIS are flat, which indicated that slight changes occurred in the microstructure of the CC fibers interconnecting the VIS during the aging process. On the contrary, fit curves of the rest subregions exhibited a U shape of the fit curves of MD, RD, and AD, and an inverted U shape of the fit curve of FA. The scatter plots of DTI indices vs. age are shown in Supplementary Figure 3. VIS, CC subregion connecting to the visual network; SM, CC subregion connecting to the somatomotor network; DA, CC subregion connecting to the dorsal attention network; VA, CC subregion connecting to the ventral attention network; LN, CC subregion connecting to the limbic network; FPN, CC subregion connecting to the frontoparietal network; DMN, CC subregion connecting to the default mode network; FA, fractional anisotropy; MD, mean diffusivity; RD, radial diffusivity; AD, axial diffusivity.