When two people talk, their brain activity tends to synchronize. Does feeling close to someone cause that inter-brain synchrony? A new study put the idea to the test. Earlier work found that couples and friends sync more than strangers. But those studies compared existing relationships. Closeness came bundled with shared history and familiarity. A team at Vrije Universiteit Amsterdam broke that bundle apart. They brought in 123 pairs of strangers and controlled how close each pair became. Then they measured what that did to inter-brain synchrony.
Mobile EEG lets the conversation flow freely
The study included 123 pairs of participants, aged 18 to 35. Nobody was paired with a friend. Every pair shared a gender. Half the pairs spoke Dutch. The rest were international pairs speaking English.
To record EEG, the researchers used two mobile EEG systems. Each amplifier streamed data to a laptop over Bluetooth. That is what made the design work. Partners sat side by side in chairs, at a normal social distance. They gestured, shifted and looked at each other. The equipment stayed out of their way.
Two frames from the video recordings tell the whole design at a glance. On the left, a talking pair sits angled toward each other with question sheets on their laps. On the right, a divider stands between the silent pair and blocks eye contact.Both laptops ran mbtStreamer. The software takes the data from each amplifier and pushes it onto the network as an LSL stream. Every other computer on the network can see it. One laptop recorded both streams into a single file, already in sync.
What the pairs actually did
Every pair drew one of three conditions at random. Those three conditions are the manipulation, the one thing the study changed on purpose. In No-Interaction, a divider blocked eye contact, and partners sat in silence thinking about superficial questions. In Small-Talk, they asked each other those same questions out loud. In Fast-Friends, the questions grew steadily more personal, a procedure that reliably builds closeness between strangers.
Every condition lasted the same 24 minutes, split into three eight-minute blocks. Experimenters handed over each question set and then left the room.
Everyone rated closeness, liking and similarity twice, once before the manipulation and once after. The silent condition acted as a baseline. It separated the effect of interaction from the effect of closeness.
How synchrony was computed
Each recording was band-pass filtered into delta, theta, alpha and beta. A Hilbert transform then gave the instantaneous phase of every signal. The analysis ran separately in each band.
Inter-brain synchrony came from the circular correlation coefficient, the circular analogue of Pearson’s r. It was computed between homologous electrodes (Fp1 against Fp1, Cz against Cz), inside four-second windows. The authors had preregistered one second, but short windows inflate phase-coupling estimates. Four seconds is the safer choice, especially at low frequencies.
The field has no single agreed measure of inter-brain synchrony. Different metrics can disagree on the same data. Circular correlation is among the more conservative — it resists coupling that appears by chance. Inter-brain synchrony was then averaged within each block, and pooled across three brain regions: frontal, central and posterior.
Strangers got closer, whatever they talked about
The manipulation did its job. Closeness jumped in both conversation conditions. The silent condition barely moved. Fast-Friends pairs ended up feeling closer than small-talk pairs. Perceived similarity followed the same pattern. Liking rose in both talking conditions, but did not separate them.
Reported closeness starts at the same low level everywhere. After 24 minutes the two talking groups sit far above the silent group. The intimate-question group sits highest of all. The gap between the two talking groups is real but small. It is dwarfed by the gap between talking and not talking.Interaction lifted inter-brain synchrony, closeness did not
Alpha carried the primary hypothesis, thanks to its links with gaze, turn-taking and imitation. It showed nothing at all. No effect of condition, no effect of region, no sign of the two combined.
The other bands had their own reasons to show something. Delta came in from work on emotional engagement, theta and beta from earlier studies comparing couples, friends and strangers. Theta and beta gave nothing either. That leaves delta. Condition mattered there, but not in the way the study predicted. The gap was between talking and not talking, not between intimate questions and small talk.
Four bands, three tests each, and almost every cell comes back empty. Alpha, the band the study bet on in advance, holds the smallest numbers in the table. Beta moves with region but not with condition. Only the delta row breaks the pattern.Delta inter-brain synchrony rose in both conversation conditions. It held across the whole scalp. Frontal and central sites showed the strongest coupling.
Delta inter-brain synchrony sits lowest in the silent condition. Both talking conditions rise above it, and they sit level with each other. The middle panel shows the frontal and central advantage over posterior sites. The third panel covers beta. There the only difference appeared over posterior electrodes, and it ran the wrong way. Coupling was higher without conversation. The effect is small and narrow, so the authors treat it as a likely fluke.Motor synchrony came from the video, via motion energy analysis. It mirrored the delta pattern exactly. Interacting pairs coordinated their movements more than silent pairs. Intimate conversation and small talk did not differ. Coordination also grew over the session and peaked in the final block.
Movement coordination sits well above the silent condition in both talking groups. The two talking groups overlap almost completely. The shape of the plot echoes the delta inter-brain synchrony result.So does moving together drive brains together? Not here. Pairs who coordinated their movements more showed no more delta inter-brain synchrony than the rest. That still left one possibility: that the delta difference came from movement, not from conversation itself. Taking movement out of the delta analysis left that effect untouched. Taking delta out of the movement analysis did the same. Both rise during interaction, but neither explains the other.
With movement ruled out, the authors offer two other reasons why delta rose during conversation. The first is speech tracking. Delta oscillations lock onto the acoustic envelope of speech. Speaker and listener both hear the same voice. Shared emotional engagement is the second. Neither could be tested here, since the study measured neither speech nor engagement.
Brain coupling and feelings went separate ways
Delta inter-brain synchrony rose whenever people talked. But did the pairs with the most of it end up feeling closest? No. It predicted nothing about closeness, liking or similarity. Motor synchrony did a little better, modestly predicting perceived similarity.
Some of the feeling was settled before anyone spoke. Dutch-speaking pairs liked each other more than international pairs at the very start. Talking closed that gap. Among pairs who never spoke, it held. A shared background gives affiliation a head start, and interaction levels it out.
What this means for hyperscanning research
This is the first causal test of closeness as a driver of inter-brain synchrony. The answer is no. Pairs who felt closer did not synchronize more, in any band. What moved synchrony was the act of interacting, which makes delta inter-brain synchrony look like a marker of social interaction rather than of affection.
Read the full paper here:Testing the causal relationship between interpersonal closeness and inter-brain synchrony (Scientific Reports, 2026)
