Did you know that mice have a variety of facial expressions, similar to those of humans? Mice will express differently when they are eating bitter or sweet foods, or when they are anxious. Measuring these expressions has been made possible due to machine learning, and researchers believe that by identifying these expressions they will be able to better understand how emotions are created and processed by the brain.
The emotions of mice
Human beings experience a wide spectrum of emotion, from pleasure and pain to happiness and disgust, and they are all reflected in our facial expressions. For example, in the case of disgust, our upper lip becomes distorted, our nostrils flare and we may frown. The patterns for these expressions are inborn, and they can be noticed in babies. We might also recognize emotions in the facial expressions of our pets. However, the rest of the animal kingdom seems expressionless. We now know that this is not true.
Using machine vision and AI, researchers at the Max Plank Institute of Neurobiology have been able to identify five emotions and their facial expressions in mice: fear, pain, nausea, disgust and pleasure. Furthermore, the algorithms were not only able to reliably identify these expressions, they could also measure their intensity.
The research has also shown that the emotional reactions were not merely triggered by the environment. Instead, they represented the value of the object that was triggering the emotion. For example, when the mice were thirsty and they drank a sugar solution, their facial expression was more joyful than the expression of mice who had the solution but were fairly satiated. In a different experiment, mice that had a slightly saline solution expressed satisfaction, but when they tasted a solution that was heavily saline they expressed disgust. Scientists wanted to take the experiment a step further and find out which areas of the brain affect which emotions. They achieved this by using light activation on various areas of the brain responsible for emotional processing.
How emotions work
Through this experiment, scientists identified the mechanism behind emotions. This was not possible to achieve without new AI technology because the emotions could never be measured accurately. Even if you could recognize subtle changes in the facial expression of the mouse with the naked eye, it is still impossible to tell which emotion it is expressing and the intensity of the emotion. With the facial recognition technology however, researchers are able to identify the type and intensity of the emotion expressed in a matter of milliseconds and then map it to the corresponding neurons. One hotbed of neuronal activity during emotional activation was the insular cortex.
During the experiments, researchers found that whenever they triggered a facial expression on a mouse, corresponding neurons in the insular cortex fired at the same time. This research is important because it has opened the door to further experimentation that can shed light on the emotional mechanisms behind mental illnesses such as depression in humans.