5 minNews
MRI Study Finds Dogs’ Brains Separate Fear From Anger and Sadness
Researchers scanning awake pet dogs found neural distinctions among some negative human facial expressions, while anger and sadness remained difficult to separate.
A functional MRI study of awake pet dogs has found that canine brains respond differently to fearful human faces than to angry or sad faces, providing new evidence that dogs can discriminate among some expressions with the same negative emotional valence.
The research, reported in iScience, used two experiments in which trained dogs remained awake and still inside an MRI scanner while viewing photographs of unfamiliar people. The design allowed researchers to measure neural responses to facial expressions without anesthesia.
In the first experiment, eight dogs viewed happy and neutral faces. Happy expressions were associated with greater activity in a right temporal cluster extending into the caudate nucleus. The caudate is involved in reward processing, and previous canine imaging research has linked positive social cues with responses in reward-related circuitry.
The second experiment included twelve dogs. They viewed faces showing happiness, anger, fear and sadness. Researchers first tested whether the brain region identified in the first experiment contained enough information to distinguish among the expressions.
A classifier could separate happy faces from each of the three negative expressions. It could not, however, reliably separate the negative expressions from one another within that region.
Researchers then performed a whole-brain representational analysis. That comparison found distinct patterns for angry versus fearful faces and for sad versus fearful faces. It did not find a reliable distinction between angry and sad faces.
The authors say the finding provides the first neural evidence that dogs can distinguish between at least some human facial expressions that carry the same negative valence. Earlier research frequently focused on broader contrasts between positive and negative emotional signals.
The researchers suggest the difference may reflect the type of threat conveyed by an expression. A fearful face can signal that danger exists somewhere in the environment, while an angry face can indicate a more direct social threat. Sadness may not signal immediate danger. Those interpretations remain hypotheses about why the neural patterns differ.
The results should not be read as proof that dogs understand human emotions in the same conceptual way people do. fMRI detects patterns of brain activity associated with stimuli. It does not reveal whether a dog has a human-like concept of fear, anger or sadness, and it cannot determine the animal’s subjective experience.
The study also used small samples, with eight dogs in the first experiment and twelve in the second. Awake-animal neuroimaging requires extensive training and limits the number of participants that can complete a protocol, but larger studies will be needed to test how broadly the findings apply across breeds, ages and backgrounds.
The use of unfamiliar human faces is significant because it reduces the possibility that the result is simply a memory response to a particular owner. At the same time, everyday interactions are richer than the experiment. Dogs normally combine facial information with voice, posture, movement, odor and prior experience.
The analysis itself offers another point of interest. A predefined temporal region was sufficient to identify the broad difference between happiness and negative expressions, while distinctions within the negative category emerged only in distributed whole-brain patterns. The finding argues against a simplistic idea that one brain area alone “reads” each emotion.
Researchers can now ask whether the neural distinctions predict observable behavior. A dog that sees a fearful person might search the environment for danger, while an angry face might prompt avoidance or a change in distance. Those possibilities were not established by the imaging study.
For now, the study narrows a longstanding question about dog-human communication. Dogs are not merely reacting to whether a face is positive or negative. At least for fear compared with anger and sadness, the canine brain preserves more specific information about what a human face looks like and may be signaling.
