The architecture of cross-hemispheric communication in the aging brain: linking behavior to functional and structural connectivity

Abstract

Contralateral recruitment remains a controversial phenomenon in both the clinical and normative populations. To investigate the neural correlates of this phenomenon, we explored the tendency for older adults to recruit prefrontal cortex (PFC) regions contralateral to those most active in younger adults. Participants were scanned with diffusion tensor imaging and functional magnetic rresonance imaging during a lateralized word matching task (unilateral vs. bilateral). Cross-hemispheric communication was measured behaviorally as greater accuracy for bilateral than unilateral trials (bilateral processing advantage [BPA]) and at the neural level by functional and structural connectivity between contralateral PFC. Compared with the young, older adults exhibited 1) greater BPAs in the behavioral task, 2) greater compensatory activity in contralateral PFC during the bilateral condition, 3) greater functional connectivity between contralateral PFC during bilateral trials, and 4) a positive correlation between fractional anisotropy in the corpus callosum and both the BPA and the functional connectivity between contralateral PFC, indicating that older adults' ability to distribute processing across hemispheres is constrained by white matter integrity. These results clarify how older adults' ability to recruit extra regions in response to the demands of aging is mediated by existing structural architecture, and how this architecture engenders corresponding functional changes that allow subjects to meet those task demands.

Figure 1.

Illustration of experimental stimuli used in the word matching task.

Figure 2.

Neuroimaging findings. Brain figure represents both the activations observed in the age group × condition interaction (heat blobs), LMFG and RMFG ROIs (red and blue), and LMFG seed for the PPI analysis (greed dot), displayed upon a canonical brain. (A) Activity in younger and older adults during the bilateral condition, depicting an age-related increase in contralateral activity. (B) LMFG and RMFG beta estimates for the age group × condition interaction, showing greater activity in the bilateral condition in right hemisphere for older adults. This interaction contrast shows that the effects observed in (A) are specific to the bilateral condition. Bars represent activity summated over the entire MFG ROI. (C) PPI parameter estimates for the right MFG again summated over the entire ROI. YA: younger adult; OA: older adult.

Figure 3.

Relationship between residual variance in bilateral accuracy (removing the variance associated with unilateral trials) and functional imaging measures across individuals. (A) Residual bilateral accuracy is predicted by both (A) activity in the right MFG, (B) functional connectivity between a seed region in left MFG and right MFG, and (C) structural connectivity within the genu of the corpus callosum. All 3 relationships represent a significant age-related difference between correlations (P = 0.03, P = 0.006, and P = 0.03, respectively).

Figure 4.

The correlation between functional connectivity (β estimates from PPI model for RMFG, based on a LMFG seed) and the BPA (residualized bilateral accuracy) is reduced by partialing out the variance associated with FA (A) and RD (B) in the genu. Parentheses indicate correlations after partialing out the variance associated with the corresponding white matter measure. This result suggests that white matter acts as a moderator by weakening the effect of functional connectivity on bilateral accuracy.