The study, published in the Cell Press journal iScience, used functional magnetic resonance imaging, or fMRI, to measure changes in brain activity as trained dogs viewed photographs of unfamiliar human faces. Functional MRI does not directly record thoughts or emotions. Instead, it detects changes in blood oxygenation and blood flow that occur when particular brain regions become more active. The participating dogs remained awake and still during the scans, allowing the researchers to observe their neural responses without anesthesia, which would have interfered with the brain processes under investigation.

The researchers first examined how dogs responded to happiness. Eight dogs were shown images of people with either happy or neutral expressions. Compared with neutral faces, smiling faces produced stronger activity in the temporal cortex and the caudate nucleus. The temporal cortex is involved in processing complex visual information, including faces and socially meaningful stimuli. The caudate nucleus, a structure deep within the brain, plays an important role in reward processing, motivation, learning, and decision-making. The combined response suggests that a smiling human face may be more than visually recognizable to dogs; it may also carry positive emotional and motivational significance.

The researchers then expanded the experiment to include 12 dogs and four human facial expressions: happiness, anger, fear, and sadness. They applied machine-learning methods to patterns of activity across the dogs’ brains, searching for consistent neural signatures associated with each expression. The analysis showed that the temporal cortex–caudate nucleus network responded in a distinctive way to happiness. The pattern was sufficiently consistent for researchers to distinguish the dogs’ responses to smiling faces from their responses to all three negative expressions.

The most surprising result emerged when the team examined the negative emotions separately. A whole-brain analysis identified activity patterns that differentiated fear from anger and sadness. This is important because it indicates that the canine brain was not merely dividing faces into two broad categories, such as positive and negative. Instead, different negative expressions generated partially distinct neural configurations. The researchers describe this as the first evidence that dogs can distinguish among specific negative human facial expressions at the level of whole-brain activity.

The findings do not prove that dogs experience happiness, fear, anger, or sadness in exactly the same way humans do. A facial photograph provides only one type of social information, and dogs normally evaluate people through a much richer combination of cues. Body posture, movement, voice pitch, speech rhythm, smell, and the surrounding context all influence how a dog interprets a human interaction. A frightened face paired with a calm voice may communicate something different from the same face accompanied by shouting or rapid movement. The neural responses observed in the experiment therefore reflect the dogs’ processing of visual facial signals, not a complete map of canine emotional experience.

The study also highlights the unusual evolutionary history of dogs. Humans and other primates may share certain social abilities because they inherited them from a common ancestor. Dogs followed a different evolutionary route, yet thousands of years of domestication placed intense pressure on them to attend to human behavior. Animals that were better at recognizing human intentions, responding to expressions, and cooperating with people may have had advantages in living alongside us. Over generations, this relationship could have encouraged the development of specialized neural systems for extracting emotional information from human faces.

Because all of the animals in the study were pets living with families, the results may not apply equally to every dog. Companion dogs receive extensive exposure to human faces, voices, gestures, and routines from an early age. Free-ranging dogs, including animals living on streets or in loosely managed communities, encounter different social pressures and may rely more heavily on posture, movement, scent, or direct behavioral cues. Their brains could process human expressions differently. Future studies comparing pet, shelter, working, and free-ranging dogs may reveal how experience and environment influence canine social perception.

The researchers say the next stage of their work will examine how dogs combine facial information with other signals to understand human behavior. They also hope to compare human and canine brain responses when both species view the same emotional cues. For dog owners, the message is immediate and striking: dogs may not understand human emotions in a human way, but their brains appear finely tuned to the emotional information written across our faces. Recognizing dogs as emotionally responsive social partners could change how people communicate with them, train them, and care for them.

Subject of Research: Animals

Article Title: Dog Brain Representations of Human Facial Expressions: Encoding Happiness and Differentiating Specific Negative Expressions

News Publication Date: 10 August 2026

Web References: https://doi.org/10.1016/j.isci.2026.116900; http://www.cell.com/iscience

References: Hernández-Pérez et al., “Dog Brain Representations of Human Facial Expressions: Encoding Happiness and Differentiating Specific Negative Expressions,” iScience, DOI: 10.1016/j.isci.2026.116900

Image Credits: Laura V. Cuaya

Keywords: dogs, canine neuroscience, fMRI, human facial expressions, animal emotions, happiness, fear, anger, sadness, brain activity, domestication, pet animals, social cognition