Scans revealed heightened activity in two key brain regions when the animals observed smiling faces
Scientists have presented the first neural evidence dogs' brains can differentiate between specific negative human emotions, rather than simply categorising facial expressions as positive or negative.
The study by researchers at the University of Vienna, published in the journal iScience, employed functional magnetic resonance imaging scans to examine canine brain activity whilst the animals viewed photographs of human faces.
The findings demonstrate happy expressions activate distinct reward-processing regions in dogs' brains.
More significantly, whole-brain analysis revealed separate neural patterns for fear, anger and sadness, suggesting domestication has equipped dogs with sophisticated abilities to interpret human emotional states.
The research challenges the previous assumption dogs merely sort human faces into "good" and "bad" categories, instead indicating a more nuanced processing system for negative emotions.
In the initial phase of the research, eight companion dogs — comprising six border collies, one Labrador, and one golden retriever — underwent fMRI scanning whilst being presented with images of unfamiliar human faces displaying either happy or neutral expressions.
The dogs viewed each photograph for seven seconds, with strangers' faces used to eliminate any familiarity effects.
The scans revealed heightened activity in two key brain regions when the animals observed smiling faces.
The temporal cortex, which processes complex visual and social information, showed increased activation alongside the caudate nucleus, a structure associated with reward processing and motivation.
"The activity in the caudate nucleus is particularly interesting," said co-author Dr Laura Cuaya. "It suggests that happy human faces may have emotional significance for dogs."
A subsequent experiment expanded the research to include 12 dogs — eight border collies, three golden retrievers and one mixed breed — who were shown photographs depicting four distinct human emotions: happiness, fear, anger and sadness.
Machine learning techniques were applied to analyse patterns of brain activity across the animals' entire brains.
The temporal cortex-caudate nucleus network demonstrated a unique response to happiness, allowing researchers to reliably distinguish neural reactions to smiling faces from all three negative expressions.
However, the most striking discovery emerged from whole-brain analysis, which identified separate activity patterns that differentiated fear from both anger and sadness.
"Humans are really good at picking up small details on the face, and so are dogs," said co-author Raúl Hernández-Pérez. "This makes dogs very interesting."
The research represents the first evidence that dogs can distinguish among specific negative human facial expressions at the neural level.
The researchers emphasised several important limitations to their findings.
Whilst the neural evidence demonstrates dogs can process different facial expressions distinctly, this does not establish that canines experience or interpret emotions in the same manner as humans.
Still photographs represent only one element of the information dogs typically access during real-world interactions with people.
In natural settings, the animals integrate facial cues with body language, vocal tone, speech patterns and olfactory signals to form their assessments of human emotional states.
The study exclusively involved companion dogs from families, raising questions about whether the findings can be generalised to other canine populations.
Free-ranging dogs and street animals navigate markedly different social environments and may process human emotional signals differently, potentially relying more heavily on movement, posture, or scent-based information.
The researchers contextualised their findings within dogs' unique evolutionary history.
Unlike comparisons between humans and primates, which examine abilities potentially inherited from common ancestors, dogs followed an entirely separate evolutionary trajectory, developing their social capabilities through tens of thousands of years of domestication alongside people.
Dr Cuaya noted this co-evolutionary process shaped dogs' capacity to decode human emotional signals.
"They followed a very different evolutionary path, yet through domestication, they developed the social abilities needed to understand and cooperate with us," she explained.
The team intends to pursue further research examining how dogs integrate facial information with other sensory inputs to interpret human behaviour.
They also plan to compare how human and canine brains process identical emotional cues.
"I would like people to see dogs as emotional beings," Dr Cuaya concluded. "They have evolved with us to understand our expressions and cooperate with us, and recognising that can change how we communicate with and care for them."
