In a groundbreaking study published in Nature Scientific Reports (Computational investigation of the social function of domestic cat facial signals), researchers from the University of Haifa’s Tech4Animals Laboratory have discovered that cats use subtle facial mimicry to strengthen social bonds with one another. Using artificial intelligence to analyze facial expressions, the team found that felines rapidly imitate each other’s facial movements, a behavior previously associated primarily with humans and other social mammals.
“Our study reveals that cats are much more social than we initially thought,” said Prof. Anna Zamansky, head of the Tech4Animals Laboratory in the Department of Information Systems at the University of Haifa and a co-author of the study. “Their facial expression language is a critical tool in how they bond with each other.”
While cats have long been stereotyped as independent and aloof creatures, the study highlights their nuanced communication systems and social flexibility. Prior research has mostly focused on feline interactions with humans or their responses to pain, leaving their intraspecies communication largely unexamined.
The research team, including doctoral student George Martvel, Prof. Ilan Shimshoni, Dr. Lauren Scott from the University of Kansas Medical Center, Dr. Brittany Florkiewicz from Lyon College, and Dr. Teddy Lazebnik from Ariel University, collected data at the CatCafe Lounge in Los Angeles, a non-profit rescue center. Observations spanned from August 2021 to June 2022, encompassing 186 interactions among 53 adult domestic shorthaired cats.
The results demonstrated that cats exhibit Rapid Facial Mimicry (RFM), linked to emotion recognition and empathy in other mammals. Positive interactions, such as playful exchanges, often involved quick imitation of facial movements—a cat turning its ears upward, for example, would prompt a similar motion from its companion in under a second. Negative encounters, however, lacked this mirroring behavior and instead featured deliberate, independent gestures like defensive head-turning or fixed gazes.
To identify these patterns, the researchers developed two AI models. One employed machine learning to analyze interaction outcomes, while the other tracked 48 facial points, focusing on ear angles, eye shapes, and mouth movements. Combined, the models achieved a 77% accuracy rate in predicting whether an interaction was positive or negative.
Beyond advancing our understanding of feline behavior, the findings carry practical implications for cat owners, shelters, and veterinarians. Recognizing signs of rapid facial mimicry could help identify strong bonds between cats, aiding adoption decisions and improving multi-cat household dynamics.
“Understanding cats’ facial language can help cat owners better understand their pets’ emotions and improve interactions with them,” said Prof. Zamansky. “This knowledge could also enhance welfare in shelter settings, ensuring cats are paired in socially compatible ways.”
The study marks a significant step in decoding the complex social world of one of humanity’s most beloved companions.
