Machine Learning for Recognizing Mice Emotional Expression

Unless referring to cartoon mice like Jerry from Tom and Jerry, Mice aren’t generally considered expressive creatures; generally, they are more of a nuisance. But neuroscientist Nadine Gogolla and her colleagues at the Maz Planck Institute of Neurobiology in Martinsried, Germany are part of a growing wave of researchers who are interested in discovering the emotional nuances of animals. In a series of experiments, Gogolla and her researchers provided the mice with stimuli and used high-speed video cameras to capture the subtle physiological reactions each stimulus elicited. What they discovered with each experiment is that different stimuli caused distinct emotions, shown in the ears, nose, whiskers, and other parts of the face. Sugar water evoked pleasure, with the ears moving forward and folding, while water tainted with quinine sent its ears straight back with a slightly curled up nose. When researchers would probe a place where a shock had previously been administered, it caused fear. All of these distinct emotional responses were confirmed with neural scans of the mice’s neural cortexes, a part of the brain in humans that is also a part of human emotional responses. Many animals are more complex than we had previously allowed for, which opens up new questions about the value and rights of animals who feel and express themselves. There are ethical questions that come up when conducting experiments like those performed by Gogolla: is it right to intentionally inflict pain and emotionally manipulate animals in order to prove some level of sentience and emotional acuity? It might be better to approach studies on animals and plants with the assumption that all living things in some way experience pain and awareness, and need to be treated with due respect.

Sanders, Laura. “Mice’s facial expressions can reveal a wide range of emotions.” Science News, April 2, 2020.

Dolensek, Nejc, Daniel A. Gehrlach, Alexandra S. Klein, and Nadine Gogolla. “Facial expressions of emotion states and their neuronal correlates in mice.” Science 368, no. 6486 (2020): 89-94.


Artificial Intelligence, Data, Ecological Modelling, Lifestyle, Visual Technologies