New York Tuesday, August 11, 2026

Boldest Voice

Search

News

Dog Brains Distinguish Fear From Anger and Sadness in Human Faces

An awake-dog fMRI study found that fearful human faces produced brain patterns distinguishable from anger and sadness, refining what scientists know about canine social perception.

San906 / Wikimedia Commons

A dog looking at a frightened human face is not simply registering “negative emotion.” New fMRI research suggests the canine brain represents fear differently from anger and sadness, adding a layer of detail to what scientists know about dogs’ sensitivity to human expressions.

The study, published in iScience, used functional MRI with awake pet dogs trained to remain still in the scanner. That matters because anesthesia would interfere with the very social responses researchers were trying to measure. Instead, the dogs could look at photographs of unfamiliar people while researchers tracked patterns of activity across the brain.

The project began with a simpler comparison. In the first experiment, eight dogs viewed happy and neutral faces. Happy expressions produced stronger activity in a right temporal region that extended into the caudate nucleus, a structure involved in reward processing. The finding fits a broader body of work showing that dogs are particularly responsive to positive human social signals.

The second experiment asked a harder question. Twelve dogs saw four expressions: happiness, anger, fear and sadness. A classifier trained on the region identified in the first experiment could distinguish happiness from each of the three negative expressions. But within that region, it could not reliably tell the negative expressions apart.

The researchers then widened the analysis to patterns across the whole brain. That is where the crucial distinction appeared. Brain responses to fearful faces differed from responses to angry faces, and fearful faces also differed from sad faces. Anger and sadness, however, were not reliably separable in the same analysis.

That makes the result more specific than the familiar claim that dogs can tell when people are “happy” or “upset.” The authors describe it as the first evidence that the dog brain can distinguish between human facial expressions that share the same negative valence.

One possible explanation is that the expressions carry different kinds of threat information. A fearful person may signal danger somewhere in the environment. An angry person may be the source of a direct social threat. A sad face does not necessarily signal an immediate danger at all. For an animal that has lived alongside humans for thousands of years, those differences could be useful.

The study does not show that dogs understand anger, fear or sadness as human concepts. fMRI measures differences in brain activity associated with what the animal sees; it cannot tell researchers whether a dog has an internal label for an emotion or experiences it the way a person does. In everyday life, dogs also have far more information than a still photograph provides, including tone of voice, posture, movement, smell and familiarity.

The sample sizes were also small: eight dogs in the first experiment and twelve in the second. Animal neuroimaging is demanding, especially when participation depends on a dog remaining awake and voluntarily still, but the findings will need replication in larger and more varied groups.

Methodologically, the study is notable because the more detailed result emerged only when researchers moved beyond a single predefined brain region. The local analysis captured a broad positive-versus-negative distinction. A whole-brain representational analysis found structure within the negative expressions. That suggests social information may be encoded across networks rather than in one simple “emotion center.”

The next question is behavioral. If a dog’s brain treats a fearful face differently from an angry one, does the dog also make a different choice—look around for danger, increase distance, approach its person or seek reassurance? Combining brain imaging with real-world behavior could show how these neural distinctions become part of the remarkable everyday coordination between dogs and humans.

Blake Kendall

Author

Science Correspondent

Blake Kendall covers public affairs, politics, business, culture and daily news for Boldest Voice. The role focuses on verification, context, and clear explanations for readers.

Read on