Grandmaternal caregiving is associated with a distinct multi-voxel neural representation of grandchildren in the parental motivation circuit

Abstract

Grandmothers enhance grandchild survival and maternal health through caregiving. Comparative evidence suggests that human grandmotherhood reflects a unique life history strategy promoting the inclusive fitness of post-reproductive females. Despite its evolutionary importance, the proximate neural mechanisms supporting grandmaternal caregiving remain unclear. This study uses functional magnetic resonance imaging (fMRI) and multivariate approaches to investigate how grandmaternal brains encode information about grandchildren and translate it into caregiving. Forty-seven grandmothers (age = 59.1 ± 7 years) completed an fMRI task viewing photos of a grandchild, the grandchild's parent, unfamiliar individuals, and nonhuman objects. Multi-voxel activation patterns associated with these stimuli were analyzed using representational similarity analysis, focusing on the hypothalamic and mesolimbic regions critical for mammalian parenting. Results reveal that grandchildren had the most distinct multi-voxel pattern of activation within these regions, potentially reflecting the grandmothers' motivational readiness to engage in grandmaternal caregiving. Indeed, greater neural dissimilarity between the grandchild and other social categories correlated with higher self-reported affection and supportive behaviors towards grandchildren, particularly in paternal grandmothers. Our findings provide novel insights into the mechanisms of grandmaternal caregiving that enhances inclusive fitness.

Keywords: caregiving; grandmother hypothesis; grandmothers; motivation; multivariate analysis.

Figure 1.

Schematic representation of the fMRI task. Participants were instructed to pay attention to stimuli and share the emotions of the targets, whenever applicable. AI-generated faces were used for creating the figure (https://generated.photos).

Figure 2.

Motivation ROI mask and model RDMs. (a) The Motivation ROI mask was defined using a binarized meta-analytic association map derived from NeuroSynth (Search term: ‘Motivation’, FDR q < 0.01). The mask encompassed subregions in the striatum, hypothalamus, and midbrain, such as the NAcc, medial MPOA, and VTA/SN. A schematic representation of RSA is illustrated with UP and OGC as examples. (b) Five categorical model RDMs were created to capture theoretical representational structure sensitive to familiarity, human face, kinship, and strangers.

Figure 3.

Representational structure of social and non-social categories in the Motivation ROI. (a) Neural RDM extracted from Motivation ROI. Different colours represent the pairwise correlation distance, with warmer colours representing greater dissimilarity. (b) Multidimensional scaling plot and (c) hierarchical clustering analysis showing the distinct multi-voxel neural representation of OGC. (d) Mean BOLD activation in the Motivation ROI associated with five stimulus categories.

Figure 4.

Whole-brain searchlight RSA. Outside regions originally included in the motivation ROI, a significant cluster was found in and around brain areas previously implicated in emotional empathy and mentalizing. Results are thresholded with cluster-level correction (Z > 3.7, cluster-level familywise error corrected P < .05) (AI = anterior insula, IFG = inferior frontal gyrus, STS = superior temporal sulcus, LG = lingual gyrus, VTA/SN = ventral tegmental area/substantia nigra, SMA = secondary motor area, dACC = dorsal anterior cingulate cortex, mOFC = medial orbitofrontal gyrus).

Figure 5.

Distinct neural representation of OGC was associated with a specific dimension of grandmaternal caregiving. The average correlational distance between OGC and other social categories (FP, UGC, and UP) was positively associated (Spearman’s ρ) with the Affective and Behavioral Support dimension of the grandmaternal caregiving but not Instrumental and Financial Support.

Figure 6.

The association between the neural representation of OGC and grandmaternal caregiving was more pronounced among paternal grandmothers. Distinct multi-voxel representation of OGC was positively correlated with Affective and Behavioral Support stronger in paternal grandmothers than in maternal grandmothers, who generally engaged in greater caregiving, regardless of the distinct neural signatures of OGC.