Both oxytocin (OT) and touch are key mediators of social attachment. In rodents, tactile stimulation elicits endogenous release of OT, potentially facilitating attachment and other forms of prosocial behavior, yet the relationship between endogenous OT and neural modulation remains unexplored in humans. Using serial sampling of plasma hormone levels during functional neuroimaging across two successive social interactions, we show that contextual circumstances of social touch influence not only current hormonal and brain responses but also later responses. Namely, touch from a male to his female romantic partner enhanced her subsequent OT release for touch from an unfamiliar stranger, yet females’ OT responses to partner touch were dampened following stranger touch. Hypothalamus and dorsal raphe activation reflected plasma OT changes during the initial social interaction. In the subsequent interaction, precuneus and parietal-temporal cortex pathways tracked time- and context-dependent variables in an OT-dependent manner. This OT-dependent cortical modulation included a region of medial prefrontal cortex that also covaried with plasma cortisol, suggesting an influence on stress responses. These findings demonstrate that modulation between hormones and the brain during human social interactions can flexibly adapt to features of social context over time.

Plasma samples for OT analysis were extracted using acetonitrile precipitation (Merck Millipore: Human Neuropeptide Magnetic Bead Panel 96-Well Plate Assay Cat. # HNPMAG-35K) and OT concentrations were then determined using the Oxytocin ELISA kit (Enzo Life Sciences; sensitivity > 15.0 pg/ml, intra-assay precision 10.2-13.3 % CV, inter-assay precision 11.8–20.9 % CV).

Plasma cortisol levels were determined using the Cortisol Parameter Assay Kit according to the manufacturer’s recommendations (R&D Systems, Minneapolis, Minnesota, USA) (sensitivity 0.071 ng/mL, precision 10.4%). Cortisol analysis serum samples were diluted 60 times preceding analysis. Pretreatment steps of the serum samples resulted in a dilution factor of three and the pretreated serum samples required an additional 20-fold dilution in Calibrator Diluent RD5-43.

For both the OT and cortisol analyses, standards and controls were implemented according to manufacturer recommendations. Washing procedures were performed using a Wellwash™ Microplate Washer (ThermoFisher Scientific, Waltham, Massachusetts, USA) and the absorbance was read using a Multiskan™ FC Microplate Photometer (ThermoFisher Scientific, Waltham, Massachusetts, USA). The color development of the samples was read for OT at 405 nm (background correction at 571 nm) and for cortisol at 450 nm (background correction at 571 nm). SkanIt™ Software was used for creation of standard curves, curve fitting and calculation of concentrations (ThermoFisher Scientific, Waltham, Massachusetts, USA).