how connection developsMuch of the body's work continues without becoming the subject of a conversation. Glucose changes after a meal. Skin temperature responds to a cold room. A sudden stimulus brings a brief rise in arousal. Another person may be sitting nearby while all of this happens.
Monia Masalha and colleagues asked whether that social presence could be relevant to these ordinary regulatory processes. Their 2026 paper on Social Physiology reports three experiments with adult romantic partners, which the authors describe as preregistered, developed after an exploratory study with infants and their mothers.
The researchers compared responses to physiological challenges when participants were alone and when a close partner was present. Their central question concerned how far a response moved from its reference level and how long that disturbance lasted. Across the adult experiments, responses were generally smaller or resolved more quickly in the social condition. The strength of that pattern varied by system and analysis.
What regulatory efficiency meant here
The authors use efficiency to describe a measured response over time. They combined the size of a response with how long it lasted, using an area-under-the-curve measure calculated differently for each physiological system. A smaller area indicated a smaller overall response under their chosen definition. A brief rise and a response that remained elevated could therefore contribute differently, even if both reached a similar peak.
This did not involve directly measuring how many calories the body used. The proposal that improved regulation could conserve energetic resources is an interpretation of the physiological patterns. Keeping those two levels distinct helps explain both what was measured and why the researchers found it interesting.
The adult experiments involved 180 different participants in total. Some took part in more than one experiment. Participants were young adults in heterosexual romantic relationships of at least nine months, and the health criteria excluded several conditions, including diabetes.
Each person completed both an alone and a social condition, with the order randomly assigned and counterbalanced. Partners could talk and touch naturally during the social condition. Researchers also compared the first condition completed by different participants. That comparison reduced concerns about a preceding condition influencing the next response, but it was an analysis of the same research sample rather than a separate replication.
Glucose after a standardized meal
In the glucose experiment, 108 participants ate a standardized portion of bread containing 50 grams of available carbohydrate. The researchers measured glucose before eating and repeatedly during the following two hours.
For glucose, the researchers calculated the area above a fixed reference of 100 milligrams per deciliter over the two hours. This was not simply the rise above each person's own starting level. The social condition showed a smaller area, with a modest effect in the comparison within individuals and a somewhat larger effect in the first-condition comparison between participants. It concerned the shape and duration of the response after the meal, rather than evidence that partners had changed participants' long-term glucose control.
This experiment was particularly important to the authors because eating the meal did not produce an increase in their measure of physiological arousal. Skin conductance, which reflects activity associated with sweat glands, did not explain the glucose difference in the analyses reported. The association remained after that measure was included in the analysis.
That result suggests the glucose finding was not simply explained by reduced stress as assessed through skin conductance. It does not eliminate every possible stress-related pathway. Other psychological, hormonal, or autonomic measures might provide information the study did not collect.
Skin temperature after brief cold exposure
The temperature experiment included 116 participants. Each condition involved a baseline period, five minutes in a cold room set to 12 degrees Celsius, and recovery in a warmer room.
Wrist sensors recorded peripheral skin temperature. The researchers found a smaller overall temperature disturbance in the social condition, with a modest within-person effect and a larger first-condition comparison. These measurements concern skin temperature, not a demonstrated change in core body temperature.
Touch was relevant here. More physical contact was associated with a smaller disturbance in the social condition. However, touch did not consistently explain the pattern across the different experiments. The temperature result after adjustment for touch was also sensitive to the statistical convention used, meeting the study's planned one-tailed test but not the corresponding two-tailed threshold.
One-tailed tests assess a predicted direction of effect. The paper used them for several planned comparisons. That is worth recognizing when interpreting small effects close to a conventional significance threshold.
Arousal in adults and infants
The adult stress experiment used two mild electrical shocks and measured skin conductance. Among 115 participants, the first-condition comparison showed a smaller response with a partner present. The within-person comparison pointed in the same direction but did not reach the conventional threshold for statistical significance, with a reported p value of .09.
The adult arousal evidence is therefore more qualified than a claim that every analysis clearly showed a benefit. When the researchers combined the three adult systems in a model that accounted for repeated participation and couple membership, they found an overall social-condition effect. That pooled finding does not remove the differences among the individual experiments.
The infant study had a different role. It was the exploratory pilot that preceded the preregistered adult experiments. The final sample included 56 infants, aged roughly two and a half to thirteen months. Researchers measured skin conductance around a brief loud sound, comparing the mother's presence with her absence.
Responses showed a smaller overall disturbance in the social condition under the study's analyses. The pilot concerned arousal only. It did not test infant glucose regulation or temperature, and it cannot establish that all the adult findings extend to infancy.
A social context for regulation
The authors consider several possible explanations. Physical contact and heat exchange may contribute in some circumstances. A familiar person may affect perceived threat. Social information may also become part of the brain's ongoing anticipation of bodily demands.
The experiments did not identify one mechanism that explains every result. Natural interaction allowed several channels to operate together, including conversation, attention, touch, and quiet company. This gives the social condition some resemblance to ordinary life, while making it harder to isolate exactly what produced a difference.
There was also meaningful variation between individuals. Some participants showed more favorable regulation with their partner, some showed little difference, and some showed the opposite pattern. The authors explicitly describe the findings as a demonstration in close bonds, where social information was expected to matter. They do not establish that proximity to any person is beneficial. The study compared being with an existing close partner and being alone. It did not compare that partner with a friend, a stranger, or another source of company, so it cannot tell us how much depended on familiarity, the particular relationship, or having someone present at all.
For someone interested in sharing more ordinary time with a close person, the social condition is a useful detail. Partners could talk, touch, or remain quietly together. They were not required to have an intimate conversation or actively reassure one another. The experiment leaves those forms of company together, so it offers no recipe for which one to choose. It does make everyday presence a more specific subject of inquiry, alongside the support people deliberately give.
The study followed brief responses within laboratory sessions. Whether these differences persist, accumulate, or contribute to disease prevention remains an open question. The results are not evidence for treating a metabolic condition through company at meals.
What they add is a more immediate question about companionship. A familiar person may be part of the setting in which ordinary physiological regulation occurs, even when neither person is actively offering comfort or discussing a problem. The body and the social situation can be studied together at that small, everyday scale.
Sources
Masalha, M., Cohen, M., Aloni, M., Toledo, E., Stahl, E., Shachar, D., Eliav, O., Fuchs, S., & Atzil, S. (2026). Social Physiology: Human homeostasis is more efficient in social proximity. Science Advances, 12(39), eaeg4937. Open-access full text.