Most of what is written about facial micro-expressions comes from two sources: television series, and courses that promise to teach you to "read people like an open book." Both distort the picture, one upward and the other downward.

This guide tries to do something different: to explain what a micro-expression really is, what the research has shown, what it has not shown, and why a window opened in recent years that did not exist before.

What is a micro-expression

A micro-expression is a movement of a facial muscle that lasts between about a fifth of a second and a quarter of a second. That is all. There is nothing mystical about it.

What makes it interesting is when it appears. When a person experiences an emotion, the brain sends a command to the facial muscles automatically, before it has time to decide whether it is appropriate to show that emotion. That command goes out first. Only after it, a second or two later in thousandths of a second, does the social layer come in, the one that says "don't show that it bothered you," and covers the movement with a more acceptable expression.

The result is a sequence of two things: a brief flash of the real emotion, and immediately after it, the cover. Whoever sees only the cover is seeing half of the story.

Why it happens: two pathways in the brain

The physiological explanation is relatively simple. Facial expressions are driven through two separate neural pathways:

The spontaneous pathway comes from deep areas of the brain that are responsible for emotional processing. It is fast, automatic, and does not pass through judgment.

The voluntary pathway comes from the motor cortex, the same area that moves your hand when you lift a cup. It is the one that produces the smile you smile when you are being photographed.

The two do not produce the same result. The most famous example is the smile: a spontaneous smile also activates the muscle around the eye, and that is what creates the little wrinkles at the corners. A voluntary smile mostly activates the muscles of the mouth. The muscle around the eye is very hard to activate on purpose; most people simply cannot.

This is why one can tell in a photograph who really smiled and who smiled for the camera, even without knowing anything about the person.

What micro-expressions do not do

Here we need to stop, because this is the point where most of what is written on the subject starts to slide.

They do not detect lies. There is no muscle in the face that is activated when a person lies. What does happen is that people whose lie might be discovered sometimes experience fear, and fear creates an expression. But that very same fear also appears in an innocent person who worries that they will not be believed. The expression is identical. It is the interpretation that fails.

They do not read minds. A micro-expression points to the presence of an emotion, not to its content. It can say "something unpleasant is going on here." It cannot say about what.

A single expression says nothing. This is perhaps the most important point. A person can raise an eyebrow because they are surprised, because the sun is in their eyes, or because that is how they are built. Meaning is created only when a pattern repeats itself again and again, in similar contexts, over time.

The research foundation: coding facial movements

In the 1970s, Paul Ekman and Wallace Friesen developed a method called the Facial Action Coding System. The idea was to map systematically every possible movement of the facial muscles, and to give each one a number, without interpreting it.

This is a subtle point that is worth understanding: the method itself does not say "this movement means anger." It says "this movement is the tightening of the muscle that pulls the eyebrows inward and down." Interpretation is a separate layer, built on top.

This separation is what made the method an accepted research infrastructure. Researchers in different fields, such as development, pain, and psychiatry, use it to describe expressions objectively, even if they completely disagree about what the expressions say.

The real debate in the field

It is important to say: there is no full agreement in the scientific community. The central debate concerns the question of how universal the connection is between a particular expression and a particular emotion.

One side holds that there is a core of basic expressions that appear similarly in all cultures. The other side argues that context, culture, and bodily state affect far more than was assumed, and that the same expression can carry different meanings in different situations.

Both sides agree on one thing: the movements themselves exist, can be measured, and can be documented repeatedly. The disagreement is about interpretation, not about the data.

In practical terms, this means that anyone who presents the interpretation of expressions as an exact science is exaggerating. And anyone who dismisses the whole field is ignoring a large and well-established body of research.

What has changed in recent years

The theory has existed for decades. What has changed is the ability to work with it in practice.

Until not long ago, coding expressions was terribly slow manual work. A trained coder needed many hours to analyze one minute of video, frame by frame. That made every study expensive, and every practical application impossible.

What has opened up in recent years is the computing ability to do it quickly and at scale. A video of a few minutes contains thousands of frames. A system can go through all of them, identify in each one which action units are active, and cross-reference the results, producing hundreds and thousands of data points from what looks like an ordinary conversation.

This is not a change in theory. It is a change in quantity. But in that quantity there is a qualitative difference: a pattern that repeats itself forty times is a finding. The same pattern, seen twice, is noise.

A review published in recent years in the journal BMC Psychiatry examined the use of algorithms for analyzing facial expressions in psychiatric contexts, and reported high levels of accuracy in the studies reviewed. It is important to qualify: these figures describe the performance of the algorithms in those studies and under those conditions, not a promise of a result in every context. We present this as an indication of the direction in which the field is moving, not as proof.

Where this fits into personality assessment

The reason this field is relevant to personality assessment is not that it reveals secrets. It is that it bypasses a built-in barrier.

Every accepted tool for understanding personality, whether a questionnaire, a test, or an interview, is based in the end on what the person reports about themselves. And every self-report is blocked at two points: what the person knows about themselves, and what they are willing to say.

A person cannot report a pattern they are not aware exists. And a person who is aware of it will not always want to write it down in a questionnaire that someone will read.

Micro-expression analysis does not pass through that channel at all. It does not ask the person who they are; it measures reactions that happen too fast to be edited. That does not make it more accurate in every respect. It makes it an independent source of information, and that is exactly its value: not in place of the other tools, but as another angle that does not overlap with them.

What to take from here

If you remember three things from this guide, let them be these:

A micro-expression is a fast muscle movement that appears before social censorship; that is the phenomenon, and that is all it is. It indicates the presence of an emotion, not its content, and certainly not truth or lies. And meaning is created only from recurring patterns in large quantity, not from a single flash that someone caught.

Anyone who promises you more than that is selling you something.

Keep reading