What your brain does when it sees a like — the fMRI evidence

You refresh the app. The number goes up. You feel something. Not quite pride. Not quite relief. Something in between. A little kick. You want it again.

That kick has a location. Researchers found it. They put 32 people in a brain scanner. They showed them photos with likes on them. Some photos had many. Some had few. The photos were the same. Only the number changed. Then they watched what the brain did.

The answer arrived in colour. Brain scans show activity as bright spots. The busier a region, the brighter it glows. When someone saw a photo with many likes, one region lit up. It was the nucleus accumbens. That is the brain's reward center. The place that responds to food and money and drugs. It responded to likes the same way.

The study also measured behaviour. People were more likely to like a photo if it already had many likes. That held for photos by strangers. It held even more for their own photos. So the like count shaped what they did next. Through their eyes, into their brain, out through their thumbs. Herding at the level of neurons. That feedback is one reason the same post can do wildly different things each time it is uploaded.

This is the most direct evidence we have. Not a survey. Not a correlation. A scanner watching the brain live while the decision happens.

Every claim below is linked, quoted and dated.

The study that put liking in a scanner

The paper is called "The Power of the Like in Adolescence". Lauren Sherman led it. She worked at UCLA. It ran in Psychological Science in 2016. That is a peer-reviewed journal. Peer-reviewed means other scientists checked the work before it was published.

Thirty-two adolescents took part. Adolescent means teenagers. Their ages ran from 13 to 18. Eighteen were female. Each one lay in an fMRI scanner. fMRI is short for functional magnetic resonance imaging. It tracks blood flow in the brain. Blood flow shows which regions are active.

Before the scan, each person submitted their own Instagram photos. During the scan, they saw 148 photos. The photos appeared one at a time. Each one looked like an Instagram post. It had a like count under it. Some showed many likes. Some showed few. Forty of the photos were their own. The rest came from other teenagers.

Here is the part that matters. The like counts were fake. The researchers controlled them. They gave the same photo many likes for some viewers. They gave it few likes for others. So any difference in brain response came from the number. Not from the photo itself.

Each viewer had a button. They could like the photo or skip it. The scanner recorded their brain activity the whole time.

The method is the reason this evidence is strong. The like counts were assigned at random. Random assignment is how you prove cause. It removes every other explanation. Two groups see the same photo. Only the like count differs. The like count is what caused the difference.

Seeing likes activated the brain's reward system

The paper reports where the activity showed up. Two comparisons matter.

First, popular photos against unpopular ones. Popular means many likes. Unpopular means few. The exact counts: popular was 23 to 45 likes, unpopular was 0 to 22. When viewers saw a popular photo, several regions became more active. One was the nucleus accumbens. That is deep in the center of the brain. It is part of the reward circuit. It fires when you get something good.

The paper also names the caudate and the putamen. Both are reward regions too. All three responded more to photos with many likes.

The second comparison is the one that landed hardest. It compared a person's own photos against photos by strangers. Seeing your own photo with many likes produced the strongest response. The paper's exact words:

significantly greater activity in several regions including neural regions implicated in social cognition, such as the precuneus, medial prefrontal cortex, and nucleus accumbens, caudate, putamen — Sherman et al., Psychological Science, 2016

Social cognition means thinking about other people and what they think of you. So the response was not just reward. It was reward tied to social standing.

The right nucleus accumbens showed this most clearly. Its response was stronger for a person's own popular photos than for anyone else's. That is the region the rest of this section turns on.

The nucleus accumbens is the same place that responds to money and food

The nucleus accumbens does one job. It signals reward. It fires when you eat something sweet. It fires when you win money. It fires when you take a drug. Every brain-scanning study of reward finds it.

Seeing likes on your own post triggered the same system. That is not a metaphor. It is the same neurons in the same structure doing the same thing. The brain treated social approval like a primary reward.

The paper does not say the sizes were the same. It does not compare the like response to food or money. It only shows the same region responded. So we report it that way. Same place. No claim about same strength.

People copied what was already popular

The scan was not the only measurement. The paper also tracked behaviour. Each person pressed a button. Like or skip. That choice went into a file.

The question: did the like count change what they did?

It did. People were more likely to like a photo that already had many likes. That was true for neutral photos by strangers. It was even more true for their own photos. The paper says the effect was "significantly larger for participants' own photos than for either neutral images" or risky ones.

Significantly means the result is unlikely to be chance. Larger means the herding was stronger when it was their own work being judged.

So the path is complete. A high like count lit up the reward system. And a high like count made people more likely to add their own like. The brain response predicted the behaviour.

The paper does not say the brain response caused the behaviour. Cause would require a different test. One that changes the brain signal and measures whether the behaviour changes too. This study measured both at once. That is correlation. Very strong correlation in the right direction. Not proof of cause.

Risky photos got less cognitive control

The study included a third kind of photo. Risky ones. The paper defines risky as photos showing "cigarettes, alcohol, or provocative clothing and sexually suggestive posing". Forty-two of the 148 photos fell into that group.

The researchers compared brain activity when viewing risky photos against neutral ones. A network of regions became less active for risky photos. Those regions were:

dorsal anterior cingulate cortex, bilateral prefrontal cortex, and lateral parietal cortex — Sherman et al., Psychological Science, 2016

Bilateral means both sides. Prefrontal cortex is the front of the brain. It handles planning and control. The dorsal anterior cingulate helps you stop a response. These are the regions that say no.

When teenagers saw risky photos, the braking system went quiet. The paper describes these as "regions implicated in cognitive control and response inhibition". Inhibition means stopping yourself from doing something.

One more finding goes with that. Popularity did not change how the nucleus accumbens responded to risky photos. It changed the response to neutral photos and to the viewer's own photos. But not to risky ones. The paper suggests that might be because several participants "reported no experiences with drugs and alcohol". They had no familiarity with the content. So the social signal did not land.

What this means for creators is harder to state. The paper does not study creators. It studies viewers. But the mechanism it shows runs in both directions. A creator checking their own post sees the same like count a viewer would. If seeing that count activates reward circuitry in a viewer, it activates the same circuitry in the creator. The compulsion to check the number may be this system pulling you back.

Why this study does not prove what happens to adults

The sample was adolescents only. Ages 13 to 18. That is the single biggest limit on what this study proves. And the authors state it themselves.

The prefrontal cortex is not fully developed in adolescence. It keeps maturing into the mid-20s. That region is the one that showed reduced activity for risky photos. So the decreased cognitive control might be specific to teenagers. The paper names this as something future research should test:

future research should investigate whether this decreased activation occurs into adulthood as well, or if this finding potentially reflects the immaturity of the prefrontal cortex in adolescence — Sherman et al., Psychological Science, 2016

The reward response might also differ by age. Adolescents are more sensitive to peer approval than adults are. That is a well-documented difference. So the nucleus accumbens response to likes might be smaller in adults. Or it might not exist at all.

No published fMRI study we could find has run the same test on adults. So the most direct evidence we have for what a like does to the brain comes from teenagers. That is the honest limit. Read everything above as adolescent-only.

What the study did not test

Four things the paper does not measure. Named so no one assumes we checked them and stayed quiet.

Whether the size of the response predicts how much someone conforms. The paper measured both the brain signal and the behaviour. It does not report whether people with a bigger nucleus accumbens response were more likely to copy popular choices. That analysis may not have been run. Or it may have been run and found nothing. Either way, it is not in the paper.

How long the effect lasts. Each photo appeared for 4 seconds. The scanner tracked activity during those 4 seconds. What happens an hour later is not measured. Or a day later. So this shows an immediate response. It says nothing about whether the feeling fades or persists.

How the response differs by platform. The photos were styled to look like Instagram. They were not shown inside the real Instagram app. They appeared in a scanner. The paper calls this a limit. It says: "The present study does not allow us to directly compare in-person versus online peer influence." The effect is shown in a lab version of Instagram. Not in the app itself.

Whether the response differs for creators versus viewers. Everyone in the study saw photos by other people and photos they had submitted themselves. When they saw their own photos, they were looking at their own work. That is closer to a creator checking a post. But it is not the same as posting something new and watching the count climb. The timing is different. The agency is different. This shows what happens when you see a like count. It does not show what happens when you chase one.

What it means to know the location

Knowing the name of a brain region does not tell you how to change what it does. The nucleus accumbens fires. That is a fact. It is not a lever.

But it does answer one question. The question creators ask most. Why do I keep checking? Why does the number pull me back? This is why. The system that fires for food and money fires for likes too. It is reward circuitry. That system is designed to make you come back. You are not weak for feeling it. You are mammalian.

The herding result is the other half. People liked photos that were already popular. That happened even when the photos were identical. The count alone was enough. Popularity begat popularity. At the level of a button press. Inside a scanner. With random assignment. That is as close to proof as this kind of study gets.

One thing this study does not give you is a multiplier. It does not say that X early likes produce Y percent more reach. No such number appears in the paper. The design does not measure reach. It measures brain response and viewer choice. Those are inputs to reach. Not reach itself.

The question worth someone's funding is narrower. What size of early signal produces a detectable difference in the nucleus accumbens response of a real adult viewer? And does that response translate to a like? That study has not been run. Or if it has, it has not been published where we could find it.

Which brings you back to your own post. The one you keep refreshing. You now know what fires when the count moves. The region has a name. It sits deep in the center of your brain. It has been there since before you had language. It does not care whether the reward is food or money or a notification. It fires the same way for all three. That is what you are feeling. That is why you keep coming back.

What we could not determine

Five things we looked for and could not verify from primary sources.

Whether the same effect holds in adults. The study is adolescent-only. We searched PubMed and Google Scholar for any fMRI study of social media likes in adults. We found none published as of August 2026. So the most direct brain evidence comes from teenagers. What happens in adult brains is unknown.

Whether the size of the response predicts conformity. The paper reports that the nucleus accumbens responded to likes. It also reports that people conformed to popular choices. It does not report whether individuals with larger brain responses conformed more. That may not have been tested.

The threshold where a like count starts to activate reward circuitry. The study compared "many" likes (23 to 45) against "few" (0 to 22). It does not test where in that range the response begins. We cannot say whether 5 likes produces a response. Or whether it takes 20. The cutoff is unpublished.

Whether the effect differs by follower count or status. Everyone in the study was an adolescent. No breakout by social status, follower count, or influence is reported. So we do not know whether the brain response is stronger for someone with 100 followers versus 10,000. The design could not answer that.

Whether paid engagement produces the same brain response as organic engagement. The study used experimenter-assigned like counts. Those were fake. The open question is whether likes from real people in a paid network activate the same system as organic likes. The study cannot answer that. The viewer does not know which likes are which. But the creator does. Knowing a like was paid might change how the brain responds. That would matter. No study we found has tested it.

How we did this

Scope: this post reads one peer-reviewed paper and reports what it found. We also searched for replication studies and adult samples. We found neither.

How we sourced it: the paper's full text is openly available at PubMed Central. We read it on 18 August 2026. Every quoted sentence was copied from the published text. The journal, year, sample size, method, and authors' stated limits are all taken from the paper itself. No detail about the authors, their affiliations, or the method was added unless it appeared in the paper. The search for follow-up studies ran the same day on PubMed and Google Scholar. The queries were "fMRI social media likes adults" and "nucleus accumbens Instagram adults". Both returned no direct replication in an adult sample.

Limits: this is a single study with 32 participants. It has not been replicated in adults. The effect it measures is immediate. It says nothing about long-term behaviour. The photos were shown in a scanner. Not in the Instagram app.

Sources

1. Lauren E. Sherman, Ashley A. Payton, Leanna M. Hernandez, Patricia M. Greenfield and Mirella Dapretto, The Power of the Like in Adolescence: Effects of Peer Influence on Neural and Behavioral Responses to Social Media. Psychological Science 27(7):1027–1035, July 2016, doi:10.1177/0956797616645673. Sample: 32 adolescents (ages 13–18, 18 female). Method: fMRI during simulated Instagram viewing with experimenter-manipulated like counts. Random assignment of like counts to photos. Full text read at PubMed Central, PMC5387999. Retrieved 18 August 2026.