The Neuroscience Of In-Group And Out-Group Perception

Before someone gets hurt, before a policy gets enacted, before a mob forms — a quieter thing happens. A classification system deep in the brain makes a determination: person, or not-person?

Most of the time you don't notice it happening. You walk through a city and your brain is constantly running social cognition on everyone around you — assigning likely mental states, reading facial expressions, imagining what they might want or fear. You do this automatically, without effort, thousands of times a day. This is the social brain doing what it evolved to do.

But there's a threshold. When someone is categorized as extreme out-group — far enough outside the circles of perceived kinship, status, or humanity — this system can disengage. And when it does, the nature of perception itself changes.

This isn't philosophy. It shows up on brain scans.

In a landmark series of studies at Princeton, social psychologist Susan Fiske and her colleague Lasana Harris examined BOLD fMRI responses as participants viewed photographs of different social groups. Their framework — the Stereotype Content Model — proposed that social groups are perceived along two axes: warmth and competence. Groups rated low on both axes (homeless people, drug addicts in some studies) occupy a unique position in the social perception landscape.

What Harris and Fiske found was that for these extreme low-warmth, low-competence groups, the medial prefrontal cortex (mPFC) — a region consistently associated with person-perception and social cognition — showed significantly reduced activation compared to other social groups. The neural signature for thinking about these individuals resembled the response to objects more than to people.

The mPFC is not the only player here. The social cognition network includes the temporoparietal junction (TPJ), posterior superior temporal sulcus (pSTS), and anterior temporal cortex — structures collectively involved in mentalizing, theory of mind, and social inference. When this network activates, you are doing the work of considering another person's inner life. When it doesn't, you're not.

The implications are significant: dehumanization, in its neurological form, is not only or primarily a matter of explicit negative attitude. It can occur even without conscious hostility. The failure may be perceptual — a failure to fully complete the cognitive act of seeing someone as a subject with an inner life.

This aligns with philosopher Frantz Fanon's observation that colonialism operates not just through oppression but through a prior act: the stripping of the colonized person from the category of the fully human. What Fanon described phenomenologically, Harris and Fiske demonstrated neurologically, albeit in a contemporary laboratory setting.

The dominant model of prejudice for much of the twentieth century was attitudinal. You hold a negative belief about a group. That belief produces discriminatory behavior. Change the belief, change the behavior.

This model is not wrong, but it's incomplete in a way that matters enormously.

If prejudice lived only in explicit beliefs, then we could reasonably expect that well-intentioned, well-educated, consciously egalitarian people would be largely free of it. But we know this isn't true. Implicit Association Tests (IAT) reveal that people who sincerely disavow prejudice often still show automatic associations between certain social groups and negative valence — faster sorting of outgroup faces with "bad" words, slower sorting with "good" words.

More importantly: the neural findings suggest that dehumanization can occur at a perceptual level prior to judgment. You're not deciding someone is less human. Your brain is, partially, failing to complete the process that would register their humanity in the first place.

This is a more uncomfortable finding. It means the work of repair can't happen only in the realm of opinion-changing. It has to happen in the realm of perception-restructuring.

The categorization system is not a rigid file cabinet. It's dynamic, cue-sensitive, and modifiable in real time.

Research in social neuroscience has identified several conditions that can shift a person from out-group to in-group processing:

Shared identity cues. The minimal group paradigm, pioneered by Henri Tajfel, showed that people form in-groups based on arbitrarily thin criteria — a shared color preference, a coin-flip, even being told you're both "overestimators." The brain is primed to find shared category membership and use it as a trigger for social processing. This can be exploited in both directions — toward inclusion or exclusion depending on which categories are made salient.

Personal individuation. When you learn specific information about a person — their name, something they care about, a detail from their life — this pulls them out of the category-processing mode and into the person-processing mode. The social cognition network comes back online. Studies by Fiske and others show that when participants are asked to consider what a photographed person might like to eat or drink, even for stereotyped out-group members, mPFC activation increases.

Shared goals or common fate. Classic contact theory research, and its more recent neural extensions, shows that working toward the same objective — particularly under conditions where the shared goal cannot be achieved without cooperation — rapidly reorganizes who counts as "us." The Robbers Cave experiment showed this in an adolescent cohort. Military combat research has shown this in adults under high-stakes conditions.

Expressed vulnerability or need. Tears, visible pain, direct appeals for help — these can bypass the categorization system by activating the caretaking circuitry, which is more primitive and harder to suppress than deliberate social evaluation. This is partly why a photograph of a single suffering child can generate more moral response than statistics about thousands.

Sustained proximity without threat. Extended, positive-or-neutral contact — the kind that allows the brain to accumulate individuating information — gradually retrains categorical response. This is one mechanism behind the well-documented finding that people with more cross-group friendships show reduced implicit bias.

The traditional approach to prejudice reduction emphasized attitude change: diversity training, counter-stereotypic information, exposure to arguments. The evidence that these work is, charitably, mixed. Brief interventions that change stated attitudes often fail to reduce implicit bias or alter behavior in high-stakes situations.

A perceptual-level account suggests a different emphasis:

Contact must be individuating. Not mere exposure to out-group members, but contact that generates specific personal knowledge. Research by Patricia Devine distinguishes between passive diversity exposure (which can actually reinforce stereotypes by keeping categories salient) and meaningful contact that generates person-level data.

Structure matters more than attitude. Intergroup contact theory's conditions — equal status, cooperation toward common goals, institutional support — aren't soft social engineering. They're engineering the specific conditions that pull people out of category-processing and into person-processing mode. The institutional part matters because it changes who sits in what room together, which changes whose individuating information you accumulate over time.

Repeated activation is required. A single positive contact experience doesn't retrain a perceptual system. But repeated exposure, over time, with sufficient positive valence — this actually does shift baseline categorical responses. The brain learns who is safe, who is familiar, who has the markers of ingroup.

Narrative and story work at the neural level. Fiction has long been suspected of increasing empathy, but there's now neural evidence that reading literary fiction improves theory of mind performance. The mechanism appears to be that deep character fiction forces sustained engagement with another person's inner life — which is exactly what activates and exercises the social cognition network.

Lasana Harris has pointed out that the societies most at risk of mass atrocity are those in which the dominant group has undergone sustained propaganda campaigns designed to reclassify a target population. The actual violence follows perceptual dehumanization, not the other way around.

This is not merely of historical interest. Right now, in multiple places on earth, media and political rhetoric are systematically working to suppress social cognition for particular groups. The target populations shift by region — migrants, minorities, political opponents, the poor, the mentally ill. The mechanism is the same. Flood the informational environment with images and language that strip the target of individuality, interiority, and relatable need. Get the social cognition network to quiet. Then the rest follows.

The countermoves are also predictable now that we understand the mechanism:

Make the individual visible. Name them. Show their face in a moment of recognizable human emotion. Give the brain the inputs it needs to run the person-processing subroutine. This is why human rights journalism that centers specific stories is not merely emotionally manipulative — it's neurologically strategic. It's counteracting the perceptual machinery of dehumanization.

The law says: We Are Human. The neuroscience says: the brain sometimes needs to be reminded of this. Not by being told a belief, but by being given perceptual conditions under which it can actually see it.

This is retrainable. The architecture is not destiny. But it requires deliberate work — in media, in policy, in urban design, in education, in the daily choices about whose names we learn and whose stories we read.

That work is not optional. It is, in the most literal sense, the difference between a species that eats itself and one that figures out how to survive.

The Name Practice. This week, in any context where you encounter someone you'd normally pass without personalization — a service worker, a neighbor you've never spoken to, a person who looks categorically different from you — learn their name. Engage one genuine question. Notice what shifts in your perception of that person over the days that follow. You are manually triggering individuation.

Media Audit. For two weeks, track which groups in your media diet are represented as categories (migrants, criminals, addicts, the homeless) versus individuals with names, backstories, and interior lives. Notice the direction of the asymmetry. Then deliberately consume at least one piece of long-form journalism or documentary that humanizes the most categorized group in your feed.

The Contempt Inventory. Make a private list of the groups toward which you feel contempt — not just dislike, but the particular flavor of dismissal that includes the sense that they're beneath engagement. That feeling is the phenomenological marker of the dehumanization circuit. You don't have to agree with anyone. You do have to restore their membership in your social cognition network. Identify one individuating fact about someone from each group on your list.

Contact by Design. Examine your life and identify where structural segregation has prevented you from accumulating individuating information about specific out-groups. Housing, workplace, social circles, consumption habits. Pick one structure to change. Not a one-time experience — sustained change in who you are physically proximate to, over time, with the goal of knowing names and lives.

---

Sources and further reading: Harris, L.T. & Fiske, S.T. (2006). "Dehumanizing the Lowest of the Low: Neuroimaging Responses to Extreme Out-Groups." Psychological Science. Fiske, S.T., Cuddy, A.J.C., Glick, P., & Xu, J. (2002). "A Model of (Often Mixed) Stereotype Content." Journal of Personality and Social Psychology. Tajfel, H. & Turner, J.C. (1979). "An Integrative Theory of Intergroup Conflict." Devine, P.G. (1989). "Stereotypes and Prejudice." Journal of Personality and Social Psychology. Fanon, F. (1952). Black Skin, White Masks. Pettigrew, T.F. & Tropp, L.R. (2006). "A Meta-Analytic Test of Intergroup Contact Theory." Journal of Personality and Social Psychology.