See rats dominated by fear? Scientists have identified their emotions through AI.

According tomedia reports, humans have a variety of emotions, and can be perfectly expressed through the movement of facial muscles and organs. In fact, this ability to express emotions is not unique to humans, and U.S. researchers have found that mice also have joyand. Researchers at the California Institute of Technology used machine learning algorithms to successfully decipher seemingly incomprehensible facial expressions in mice in the lab, a work that could have a huge impact on pinpointing neurons in the human brain that encode specific expressions.

See rats dominated by fear? Scientists have identified their emotions through AI.

Rats express emotions by moving their ears, cheeks and eyes.

David Anderson, a neuroscientist at the California Institute of Technology, said their study “took a very important first step” to help demystify emotions and how they manifest them in the brain.

About 150 years ago, charles Darwin, a renowned biologist, suggested that animal facial expressions may be like humans’, providing a window into their emotions. But researchers didn’t get tools such as powerful microscopes, cameras and genetic technology in recent years to accurately capture and analyze facial movements and investigate how emotions are generated in the brain.

Nadine Gogolla, a neuroscientist at the Max Planck Institute for Neurobiology in Germany who led the three-year study, said: “I am fascinated by the abundance of emotional states in our human son. I want to see if we can understand from animal studies how these conditions occur in the brain. “

See rats dominated by fear? Scientists have identified their emotions through AI.

The study found that mice also had similar emotions to humans, including joy and anger, and could express them through their faces

Gergoula was inspired by a cell paper co-authored by Anderson and Ralph Adolphs, a professor at the California Institute of Technology. In the study, they suggested that “brain states” such as emotions may exhibit particular characteristics. For example, they should be durable, and they should last for a while after the stimuli that evoke them disappear. In addition, emotions may become more intense as stimuli intensify.

Gegula’s team first fixed the mice’s heads to keep them stationary, then provided different sensory stimuli to trigger specific emotions and photograph their faces. For example, the researchers placed sweet or bitter liquidons on the lips of mice to evoke feelings of pleasure or disgust. They also gave the mice an appropriate degree of electric shock on their tails, or injected them with lithium chloride to induce discomfort.

Scientists know that mice can change their expressions by moving their ears, cheeks, nose and upper eyes, but they can’t reliably match them to specific emotions. So when the mice responded to different stimuli, the researchers broke down videos of the mice’s facial muscle movements into ultra-short snapshots.

The AI algorithm recognizes the different expressions produced by the movement of specific facial muscle groups. These expressions are associated with evoked emotional states, such as pleasure, disgust, or fear. For example, a happy mouse pulls his nose down toward the mouth, ears and chin forward. In contrast, when it feels pain, it pulls the ear back, bulgs his cheeks, and sometimes squints.

These facial expressions in mice had characteristics suggested by Anderson and Adolphus, such as that they could last for a long time and that the intensity was closely related to the intensity of the stimuli. Camilla Bellone, of the University of Geneva in Switzerland who studies neuropsychiatric disorders, said: “This way of recognizing facial expressions has a big advantage because it avoids any bias that researchers use. “

The scientists then looked for brain cells that might encode these emotions in the brain. Using a technique called optogenetics, they targeted individual neuronal circuits in mice that have been shown to trigger specific emotions in humans and other animals. When the researchers directly stimulated these circuits, the mice showed corresponding facial expressions.

Finally, the researchers used a technique called double photonic calcium imaging to identify individual neurons in the mouse brain that only activate when specific emotions and specific facial expressions are aroused. “They may represent part of the emotional coding in the brain, ” Gegula speculates. We think that emotional coding may be evolutionaryly conservative, so human and mouse coding may have many common characteristics. “