The Neuroscience of First Impressions

A new classmate walks in on the first day of term. Before they have said a word, you have already decided something about them: friendly, a bit intimidating, maybe someone you would trust with a secret. You did not choose to make that judgment. It simply arrived.

Psychologists have studied this moment for decades, and the findings point in two directions at once. Our brains are remarkably quick and consistent at forming impressions from faces. Those impressions are also, quite often, wrong.

A judgment in a tenth of a second

In a well-known set of five experiments, researchers showed people photographs of unfamiliar faces for different lengths of time and asked them to rate traits such as attractiveness, likeability, trustworthiness, competence and aggressiveness. Judgments made after seeing a face for just 100 milliseconds correlated highly with judgments made in the absence of time constraints. For all five traits, increased exposure time did not significantly increase the correlations.

Looking longer did change something, though. When exposure rose from 100 to 500 milliseconds, participants’ judgments became more negative, response times for judgments decreased, and confidence in judgments increased. In other words, a little more time mostly made people more sure of the impression they had already formed.

It is worth being precise about what this study measured. It showed that brief glances produce the same judgments as long looks. It did not test whether those judgments were true of the people in the photos.

What happens in the brain

One region that keeps appearing in this research is the amygdala, a pair of almond-shaped structures deep in the brain that are involved in detecting threats and emotionally important events. A 2016 systematic review and meta-analysis pooled brain-imaging studies of faces that look more or less trustworthy. Its pooled analysis supported the evidence that the (right) amygdala responds preferentially to untrustworthy faces, and that the amygdala, insula and part of the thalamus showed negative correlation with trustworthiness. The same analysis also found high heterogeneity between studies, meaning results varied a lot from one study to the next, so the size of this effect is uncertain.

Why would the brain bother making such fast guesses at all? One influential explanation is the “overgeneralization” idea. According to this view, facial cues that normally carry useful information, such as an angry expression or the features of a baby, end up shaping impressions of people who merely resemble those categories. For example, higher dominance and lower warmth are perceived not only in angry faces, but also in neutral expression faces that show more resemblance to angry expressions. The same review notes remarkable consensus in first impressions of traits like warmth and dominance, including agreement across very different cultures.

Impressions that change real outcomes

These snap judgments do not stay in the lab. In one study, people saw pairs of photographs of political candidates and judged who looked more competent. Those competence judgments predicted the outcomes of U.S. congressional elections better than chance (e.g., 68.8% of the Senate races in 2004), and the inferences occurred within a 1-second exposure to the faces of the candidates. That shows impressions from faces may influence how people vote. It does not show that the candidates who looked competent were actually more competent.

Fast and shared, but not very accurate

So are first impressions right? The honest answer is “sometimes, a little, and much less than we feel they are.”

A review of the field concluded that although some first impressions of traits show better than chance accuracy, impressions can also be wrong. A study that compared people’s self-reported personality with how strangers rated their faces found a similar pattern: several personality traits can be read above chance from a face, yet the researchers’ prediction model fails to reliably infer personality traits from either facial features or first impressions.

Another problem is that a face is not a fixed thing. People look different across photos, lighting, angles and moods. When researchers collected trait ratings for many different images of the same people, they found that within-person variability in social trait evaluations is at least as great as between-person variability. Put simply, two photos of the same person can produce impressions as different as photos of two different people. That is hard to square with the idea that a face reliably reveals a stable personality.

A 2026 speed-dating study added a real-world angle. After 3-minute conversations, participants rated each other, and separate raters judged the same people from photos alone. Physical traits, such as attractiveness and athleticism, showed strong correspondence between face-only and in-person ratings, and there was also correspondence between some non-physical traits. Note what this compares: impressions from photos with impressions after a short meeting. Agreement between two impressions is not the same as accuracy about who someone really is.

Does the pattern hold everywhere?

Much of this research was done with Western participants. A large preregistered project, run as a Registered Report, tested a leading model of face impressions, which says judgments boil down to two dimensions (roughly, trustworthiness and dominance). It repeated the original method across 11 world regions, 41 countries and 11,570 participants. Using the original analysis, the valence–dominance model generalized across regions. With an alternative method that allowed the dimensions to be related, the researchers observed much less generalization. So the broad finding replicated, but how universal the exact structure is depends on how the data are analyzed.

What we still do not know

Several big questions remain open. Most studies use posed, neutral photographs, which differ from the moving, talking people we meet. Brain-imaging studies show which regions respond to faces that look trustworthy, but they are correlational and vary widely in their results. And researchers still debate how much of the small accuracy that does exist comes from faces themselves versus from expressions, grooming or context.

The takeaway

Your brain forms an impression of a face within a fraction of a second, and other people often form the same impression. That speed and agreement can make first impressions feel trustworthy. The evidence suggests they deserve less confidence than they inspire: they shift from photo to photo, and they predict actual personality only weakly. A reasonable inference from this research, rather than a tested rule, is that a first impression is best treated as a first guess, one worth checking against how a person actually behaves over time.