Loneliness is associated with distinct divergences in hippocampus (HC)-default network (DN) covariation. We explored the structural covariation between the 38 subregions of the HC and 91 subregions of the DN, by means of a codecomposition based on a canonical correlation analysis (CCA). We subsequently determined how the ensuing subregion patterns diverged in individuals with loneliness. Shown here are the subregion divergences in mode 1 of HC-DN covariation. Mode 1 of the CCA solution achieves the most explanatory covariation, with a canonical correlation of rho = 0.51. A shows the HC subregion pattern (left, one canonical vector of mode 1) with parameter weights that robustly diverge between lonely and nonlonely groups in mode 1; mapped onto 8 consecutive coronal slices of the left and right HC in the anterior (top)-to-posterior (bottom) direction. B shows the DN subregion pattern (right, other canonical vector of mode 1) that robustly diverged between the lonely and nonlonely groups. Overall, within the dominant structural covariation pattern between the HC and the DN there are specific subregions whose volumes systematically diverge in lonely individuals. The most pronounced structural divergences are in bilateral CA1 body and hippocampal fissure, as well as subregions bilaterally in the posterior cingulate cortex (PCC), retrosplenial cortex (RSC), and dorsomedial prefrontal cortex (dmPFC). Thus, specific HC and DN anatomical subregions are preferentially linked loneliness. ML, molecular layer; PrS, presubiculum; STS, superior temporal sulcus; Sub, subiculum; TPJ, temporoparietal junction.