Social interaction sequences form relationship trajectories along abstract dimensions of affiliation and power.

(A) An example of a power interaction. Participants read text that describes the narrative and on decision trials choose between two options. Based on their choice, the character moves −1 or +1 along the active dimension (power or, on other trials, affiliation). (B) The participant forms relationships with different characters through sequences of interactions in the narrative. The decisions the participant makes in the implicit affiliation and power interactions change the character’s location in social space, forming a relationship trajectory (participants are unaware of the dimensions). Participants interact with 6 characters: 5 each with 6 affiliation and 6 power trials and 1 with 3 neutral trials. 4 characters are shown for illustration. (C) Schematic of post-task self-report placements. The mapping error between the behavioral and subjective placements was calculated as the average character-wise Euclidean distance between the locations; this was significantly smaller than expected by chance. (D) Mapping error was negatively correlated with task memory, suggesting the subjective maps depend on memory. 95% confidence intervals for regression line are indicated by the shaded region and p-value significance is indicated by asterisks: * < 0.05. Validation sample only (n = 32).

The left hippocampus represents affiliation and power decision trials differently.

Decision trials of the same dimension had more similar neural patterns than trials of different dimensions. A) An analysis restricted to an anatomically defined hippocampus showed a significant cluster in the left hemisphere. B) An exploratory whole-brain analysis showed additional clusters in the bilateral middle temporal gyrus (MTG) and angular gyrus (AG). Participants were pooled across samples (n = 50); sample included as covariate. Permutation-derived pFWER < 0.05.

The hippocampus represents dynamically changing abstract social locations.

A) We used representational similarity analyses (RSA) to study location-like effects. We used the behavioral locations implied by the participants choices to compute behavioral distance matrices. The distance matrix shown is a small, toy schematic to represent the logic; there were 60 decision trials in the task. We ran regressions and included control distances for task- and time-related variables to isolate the contribution of abstract location-related distances. B) We ran a within-character version of the RSA, where distances were within-character, to test for the brain dynamically tracking locations over relationship-specific social interactions. C) The anatomically defined left and right hippocampi both showed within-character location-like effects, in a region-of-interest analysis. The left show effects from the mid- to posterior hippocampus, whereas the right was mainly in mid-hippocampus. D) An exploratory whole-brain analysis also showed clusters in the posterior cingulate cortex (PCC) and putamen (PUT), as well as a small cluster in the right inferior temporal cortex (not shown). E) We ran an additional RSA, including all trial pairs, to test for evidence of a shared map-like representation. F) In a hippocampal specific analysis, there was an effect in the mid-to-posterior right hippocampus. In these analyses, participants were pooled across samples (n = 50); sample included as covariate. Permutation-derived pFWER < 0.05.

Examples of affiliation and power interaction decisions.

Several examples of affiliation and power interactions are shown. The slide preceding the choice trial is shown to give the context for the interaction, along with the decision that would increase and the decision that would decrease the location along the relevant dimension.