Same bread, same person, same 2 hours of finger-prick tests.
The only difference was who sat on the sofa. When a romantic partner was in the room, blood sugar rose less after the meal and fell back faster.
That’s the headline result of a set of experiments just published in Science Advances. The researchers then tried the same idea with cold, with mild electric shocks and with babies.
The pattern held each time, though some effects were tiny.
- In a lab, young adults’ blood sugar response to white bread was about 15% smaller with their partner present than alone.
- Their skin cooled less in a cold room, and their stress response to mild shocks was a little smaller, with a partner nearby.
- Babies reacted slightly less to a loud noise when their mother was with them.
Is being alone bad for your health?
Long-term loneliness and isolation are linked to real harm. That part isn’t new.
The World Health Organization says 1 in 6 people worldwide is affected by loneliness. It links loneliness to more than 871,000 deaths a year, about 100 deaths every hour.
The WHO also says loneliness and social isolation raise the risk of stroke, heart disease, diabetes, memory decline and early death.
But most of that evidence comes from following people over years. It shows links, not what happens inside the body from moment to moment.
That’s the gap this new study tries to fill. Not “do lonely people get sicker?”, but “does the body work differently when someone close is there?”
The study: four tests, alone and together
Who took part
The work comes from the Atzil Lab at the Hebrew University of Jerusalem. The adult experiments were preregistered, meaning the team published its plan and predictions before collecting data.
The 180 adults were in heterosexual couples, aged 18 to 37, with an average age of about 25. Half were women.
Couples had been together for at least 9 months.
Everyone was healthy, with no diabetes and a normal body weight range. Some people took part in more than one experiment.
| Test | People | What they faced | What was measured |
|---|---|---|---|
| Blood sugar | 108 adults | 102 g of white bread (50 g of carbs) | Finger-prick glucose for 2 hours |
| Cold | 116 adults | 5 minutes in a 12°C room | Skin temperature |
| Stress | 115 adults | 2 mild electric shocks to the forearm | Skin sweat response |
| Babies | 56 babies, aged 2.5 to 13 months | A 2-second loud noise (82 dB) | Skin sweat response |
How “alone” and “together” worked
Each person did the test twice on the same day. Once alone, once with their partner in a comfortable room with a sofa and a carpet.
The order was random. Partners could talk or touch as they liked, and sessions were filmed.
For the babies, the mother either held and played with her child, or waited in another room, out of sight and hearing.
The team measured the size and length of each body response. A smaller, shorter response to the same challenge counted as more efficient regulation.
What the study found
Blood sugar rose less with a partner
This is the most surprising result. Blood sugar after a meal isn’t something we usually think of as social.
When people ate the bread with their partner present, their blood sugar response was about 15% smaller than when they ate alone.
Glucose levels started at similar levels in both sessions. After eating, they were lower with a partner at every check from 45 minutes to 2 hours.
The team also compared only each person’s first session, to rule out any carry-over from doing the test twice. That gap was bigger: about 24%.
In a lab, the same bread raised blood sugar about 15% less when a close partner was in the room. The cold and stress tests showed the same direction, but the effects were smaller.
Cold and stress showed the same pattern
In the cold room, skin temperature dropped less and recovered faster with a partner. The response was about 7% smaller.
The stress test was the weakest. The response to the shocks was about 2% smaller with a partner, and that difference wasn’t statistically significant.
Comparing only first sessions, the stress gap was about 6%, which was significant.
| Test | Smaller response with partner (same people, twice) | Smaller response with partner (first session only) |
|---|---|---|
| Blood sugar | 14.6% | 23.9% |
| Cold | 6.8% | 16.2% |
| Adults’ stress | 1.9% (not significant) | 6.0% |
| Babies’ stress | 2.6% | 4.2% |
Pooled across the three adult tests, responses were about 8% smaller with a partner. Looking only at first sessions, they were about 12% smaller.
Babies reacted a little less with their mother
The baby test was an early pilot, not preregistered. Still, babies’ stress response to the loud noise was about 3% smaller with their mother nearby.
The effect didn’t depend on the baby’s age.
It wasn’t just calm or cuddles
An obvious guess is that a partner simply makes you less stressed. The team checked that in the blood sugar test.
They found stress levels didn’t rise during the meal test, and didn’t differ between alone and together. The blood sugar effect held after accounting for stress.
Touch didn’t explain it either. More touch went with a smaller temperature drop in the cold room, but touch wasn’t a reliable explanation across the tests.
The results also held after accounting for test order, time of day, body weight and how long people took to eat.
Why the result needs caution
This is a careful study, but it’s a lab study of short tests. It doesn’t show that company prevents diabetes or helps you live longer.
The people were young, healthy couples. We don’t know if the same thing happens with friends, older adults or people with diabetes.
There was no stranger or friend condition. So we can’t tell if a partner matters, or just having anyone in the room.
People weren’t blind to the condition, and couldn’t be. They knew whether their partner was there.
“More efficient” is the authors’ reading of a smaller response. The study didn’t measure the body’s energy use directly.
Some people actually did worse with their partner present. The authors call this a “social loss” and say company isn’t always helpful.
The work was funded by a European Research Council grant. The lead scientist is exploring a patent related to the findings; the other authors declared no competing interests.
What earlier research found
Strong relationships and survival
A 2010 meta-analysis in PLOS Medicine pooled 148 studies with 308,849 people.
People with stronger social relationships were 50% more likely to survive during follow-up. The authors said that link was as strong as well-known risk factors for early death.
Holding a partner’s hand under threat
A 2006 study in Psychological Science scanned the brains of 16 married women facing the threat of a shock.
Holding their husband’s hand calmed threat-related brain activity the most. A stranger’s hand helped less.
Happier marriages showed a bigger effect.
Loneliness and type 2 diabetes
A 2025 meta-analysis in Diabetes Research and Clinical Practice pooled 9 studies with over 1.1 million people, followed for about 11 years.
Loneliness was linked to a 32% higher risk of type 2 diabetes. Social isolation was linked to a 20% higher risk.
Not every measure moves
A 2025 meta-analysis in Health Psychology asked whether people who feel well supported have calmer heart responses to stress.
It found no clear link with blood pressure or heart rate reactions. So social support doesn’t show up in every body measure.
How much should you trust this?
Promising. The design is fair and the results point the same way, but the effects are short-term and mostly small.
What makes it convincing
- The adult experiments were preregistered, with predictions published first.
- Each person served as their own comparison, in random order.
- The pattern repeated across four different body systems and two age groups.
- The data and code are publicly available.
What makes me cautious
- These are short lab tests, not long-term health outcomes.
- The participants were young, healthy couples.
- No stranger or friend condition was tested.
- Some effects were tiny, and one wasn’t significant.
- The lead scientist is exploring a patent related to the work.
| The study shows | The study does not show |
|---|---|
| Smaller blood sugar responses with a partner present | That company prevents diabetes |
| Similar, smaller effects for cold and stress | That any person in the room has the same effect |
| The effect wasn’t explained by stress or touch | What actually causes it |
| Babies reacted slightly less with their mother | That these short effects add up over years |
What this means for you
- Don’t change any diabetes treatment because of this. It’s a lab finding in healthy people. Follow your doctor’s advice.
- Shared meals may matter beyond the food. This study can’t prove it, but it gives one possible reason why company might help the body.
- Take loneliness seriously. The long-term links are strong. The WHO’s report on social connection has practical ideas for staying connected.
- Company isn’t magic. Some people did worse with their partner. Good company matters more than any company.
In this TED talk, psychiatrist Robert Waldinger shares what one of the longest studies of adult life learned about relationships and health:
What we still don’t know
- Does it work with friends or strangers? Only partners and mothers were tested.
- What drives it? The authors suggest the brain may plan ahead differently when someone close is near. That isn’t tested yet.
- Does it help people with diabetes? Everyone in the study had normal blood sugar.
- Do small, daily effects add up? Nobody has followed this over months or years.
- Who gets the “social loss”? Why some people did worse with their partner is still open.
My take: the body notices who’s in the room
What I like about this study is the blood sugar test. Nobody expects a partner on the sofa to change how you handle a slice of bread.
I’m less convinced by the big claims about evolution and energy. A 2-hour lab session can’t carry that much weight, and some effects were very small.
Still, it adds a missing piece. We’ve long known that lonely people tend to be less healthy.
This study suggests the body may respond to good company within minutes.
Paper: Social Physiology: Human homeostasis is more efficient in social proximity
Published: Science Advances, 2026-09-23
Study: Four lab experiments comparing each person alone and with a close partner; three preregistered
Who: 180 healthy adults aged 18 to 37 in romantic couples, and 56 babies with their mothers, in Israel
Funding: European Research Council Starting Grant; the lead scientist is exploring a patent related to the work, other authors declared no competing interests
Evidence: Promising — consistent, preregistered results, but short-term lab tests with mostly small effects
Comments
No comments yet. What did you think?