How to Read Micro-Expressions

Someone tells you they are fine. Their face agrees for a fraction of a second, then something flickers underneath it, and by the time you register the change it is gone. You cannot name what you saw, but you walk away certain that fine was not the whole story.

That flicker has a name. Micro-expressions are involuntary facial movements that appear and disappear in roughly a twentieth to a fifth of a second, far faster than the expressions we consciously produce. They are the moment before self-control catches up, and because they arrive faster than deliberate thought, they are much harder to fake than the expression that follows them.

This guide covers what micro-expressions actually are, the seven core patterns worth learning, how to read them without turning into a suspicious amateur profiler, and where the limits are. Because those limits matter, and most articles on this subject skip past them entirely.

Your face is controlled by two different neural pathways. One runs through the motor cortex and handles deliberate expressions: the smile you produce for a photograph, the concerned frown you arrange when a colleague describes a problem. The other runs through older subcortical structures and fires automatically in response to emotion, before you decide anything.

When you feel something you would rather not display, both systems engage at once. The automatic pathway produces the genuine expression, and the deliberate pathway moves to suppress or replace it. Micro-expressions are the brief window where the first system wins before the second one arrives. It is a race, and for a fraction of a second the honest signal is ahead.

This is why they are considered difficult to fake. You can hold a pleasant expression indefinitely, but you cannot easily prevent the flash that precedes it, because the suppression begins after the expression has already started.

Paul Ekman's research identified a set of facial expressions that appear consistently across cultures, including in communities with minimal exposure to outside media. These are the patterns most micro-expression training is built on.

Happiness. The key is the orbicularis oculi, the muscle ring around the eye. A genuine smile crinkles the outer corners of the eyes and lifts the cheeks. A social smile moves the mouth alone. This is the single most useful distinction on the list, because polite smiles are everywhere and the difference is visible once you know where to look.

Sadness. The inner corners of the eyebrows pull up and together, creating a small triangle of tension in the middle of the forehead. The corners of the mouth turn down. This inner-brow movement is notoriously hard to produce voluntarily, which makes it one of the more reliable signals when it appears.

Anger. Eyebrows lower and draw together, the eyelids tighten, and the lips press into a thin line or square open. The forehead often shows vertical lines between the brows rather than horizontal ones.

Fear. Eyebrows raise and pull together, upper eyelids lift so more white shows above the iris, and the lips stretch horizontally toward the ears. Fear and surprise share the raised brows, but surprise has no horizontal lip stretch and no brow convergence.

Surprise. Brows arch upward, eyes widen, jaw drops. It is the shortest of all the expressions, usually resolving within a second into whatever the person actually feels about the news.

Disgust. The nose wrinkles, the upper lip rises, and the cheeks push up. It appears in response to physical revulsion and also to moral or social distaste, which is why it turns up in conversations that have nothing to do with smell or taste.

Contempt. The only asymmetric expression on the list. One corner of the mouth tightens and lifts. It signals a sense of superiority, and in relationship research it is treated as an unusually strong warning sign when it appears regularly between partners.

The obstacle is speed. Micro-expressions occur faster than untrained attention reliably registers, so most people simply miss them in live conversation. There are a few practical ways around that.

Use still images first. A photograph freezes the moment. You lose the movement, but you gain unlimited time to examine the eye corners, the brow position and the mouth. Working through photos is how most people build the initial vocabulary before attempting anything live.

Watch video with the sound off. Interviews and unscripted footage work well. Removing audio forces you to read the face instead of the words, and pausing on a reaction lets you examine what you would otherwise miss. Political interviews and press conferences are useful precisely because the speaker is managing their expression.

Look at transitions rather than states. The informative moment is not the settled expression, it is the shift into it. When someone receives news, watch the half second before their face arranges itself. That gap is where the unmanaged reaction lives.

Focus on the eyes and brows. The lower face is easier to control consciously, which is why fake smiles are so common and so recognisable. The upper face leaks more, particularly the inner brow and the muscles around the eyes.

Treating a flicker as proof. A micro-expression tells you an emotion occurred. It does not tell you why. Fear during a question might mean the person is lying, or it might mean they are afraid of being disbelieved while telling the truth. These are different situations with an identical facial signal.

Ignoring context. A flash of anger during a conversation about a delayed project could be aimed at you, at a supplier, at themselves, or at a memory the topic triggered. The face reports the emotion, not its target.

Assuming this detects lies. It does not. Research on deception detection consistently shows that people perform close to chance, and training produces modest improvements at best. Micro-expressions reveal concealed emotion, and concealed emotion is not the same thing as dishonesty.

Reading strangers with confidence. Baselines differ. Some people have naturally furrowed brows or asymmetric mouths. Without knowing someone's neutral face, you are comparing their expression to an average that may not apply to them.

It is worth knowing that this field is not settled. The universality of facial expressions has been challenged by researchers who argue that emotional expression is more culturally variable and context dependent than the classic model suggests. Some studies find weaker links between specific facial configurations and specific internal states than the popular accounts imply.

The practical takeaway is not that reading faces is useless, but that confident interpretation of a single frame is unjustified. Facial signals are evidence, and evidence is something you weigh alongside tone of voice, word choice, the situation and what you know about the person.

Used carefully, this skill improves ordinary attention rather than granting special insight. You notice that a friend's agreement did not reach their eyes and you ask a follow-up question. You catch a flash of confusion during an explanation and you slow down. You register discomfort in a photograph of a family gathering and you understand something about that day that nobody wrote down.

The goal is not to catch people out. It is to notice the moments when what someone shows and what someone feels have come apart, and to respond to the second one with a little more care. That is the difference between reading faces and interrogating them.

If you want to practise on real images, upload a photo to our face emotion scanner and see which expressions the AI identifies. Comparing your own reading against the analysis is a fast way to find out which signals you are catching and which ones you are still missing.