Researchers used machine learning to analyse brain scans of 12 border collies ​as the animals viewed images of human ‌faces displaying different emotions. When shown images of angry, fearful and sad emotions, those regions did not produce a response. “To the best of our knowledge,” the team said in its study, “this is the first evidence that dogs can distinguish between two human faces showing negative expressions. “This suggests that the dog brain separates human facial expressions by valence, and shows separability among some negative human facial contrasts, particularly those involving fear.” Researchers added that the 12 participants ​in the study were all pets from loving ​families, and cautioned that dogs outside such environments may respond to human emotional cues differently.