Funny soul! Artificial intelligence cracks "whispering" in mice in social behavior

Learn about the latest progress in neuroscience ● click the blue word to pay attention to us ● the normal social communication between animal individuals is conducive to the formation of stable social networks, which is of great significance to the survival of individuals and the reproduction of races.social behaviors include dynamic and complex social interactions among animals, which can promote cooperation or competition within groups.animals complete social behaviors through vision, smell, hearing and sound, among which sound plays an important role in warning other animal members to avoid predators, and can also be used as a signal for courtship.using mice as a model system, researchers have made important progress in neurobiology of social behavior and behavior in complex social groups.mice are gregarious animals and participate in a variety of behaviors accompanied by ultrasonic vocalization (USV).studies have shown that mice are more likely to approach their vocal counterparts, and they don't like playing with dumb people.although USV plays a communicative role in social interaction, there is no direct evidence that voice patterns are directly related to specific social behaviors.in March 2020, Joshua P. neuunebel
research team from the Institute of psychology and brain science, University of Delaware, published an article in nature neuroscience in March 2020, revealing the above-mentioned connection by using the method of deep learning: the voice patterns of mice under different social behaviors, revealing the "whispering" of mice in different social behaviors, and translating the mouse language.different species have different hearing ranges. The hearing frequency range of mice is 4k-100khz, and that of humans is 20-20000hz. Therefore, it is necessary to pay attention to animal experiments, otherwise your behavioral results will be affected.there are still some difficulties in the study of the relationship between voice production and social behavior. Researchers need to solve these technical problems in order to carry out their work better.first of all, it is necessary to accurately identify which mouse is vocal in social behavior and track social behavior.secondly, we need a systematic, objective and unbiased method to classify voice patterns of different social behaviors.finally, it should also be noted that in the animal community, not one, but many animals together. Therefore, it is necessary to develop high-throughput detection methods to detect the social behavior and voice patterns of multiple mice at the same time.the researchers observed a total of 11 groups of mice, each containing two estrus female mice and two male mice, living together.in order to observe the voice of mice in these social behavior states, the researchers used artificial intelligence algorithm to detect the specific behavior of mice, and used the sound source localization system to track the voice status of each mouse.according to different behavior characteristics, the social intercourse of male mice can be divided into six categories: walking, circling around another male mouse, attacking, escaping, male chasing female, male chasing male (yes, you are right, male). through experiments, they found that the social behavior characteristics of these male mice were different, mainly manifested in the frequency and duration of behavior completion, which was similar to that of human beings. through the sound detection system, a total of 135078 signals were detected in mice, among which the proportion of ultrasonic signals produced by male mice was 82.5%. These sound signals were disordered, how to distinguish them effectively? According to the characteristics of its acoustic structure, researchers have developed a clustering method for automatic grouping of speech signals, which can be divided into 22 subcategories. the classification of sound signals enables us to describe the voice patterns of male mice more objectively. The duration, frequency range and amplitude of these signals vary among different subtypes. surprisingly, it is not the duration of these voice signals that affects social behavior, but that social behavior affects vocal patterns. the above results are based on the whole, and no specific social behavior analysis is conducted. researchers at higher social levels are more likely to learn voice signals at different social levels. is this the rhythm of the task that the rat king is giving each soldier mouse? What's more interesting is that under the social behaviors of attacking, escaping or chasing other males, male mice prefer to send out the sound signal of falling tone, and the sound tends to be heavy and thick, which is the feeling of roaring and roaring; while in the social behavior of walking and courtship, male mice prefer to send out the signal of rising tone, and the sound tends to be light and short, giving the same kind of gentle feeling. in addition, the researchers found that the vocal patterns of male mice in the same social network affected the behavior of social partners. in conclusion, this paper analyzes the voice communication signals between mice in social interaction based on artificial intelligence algorithm, and finds that specific social behavior will produce specific voice patterns. social communication is the transmission of information between two or more animals, which usually leads to behavioral changes. in humans, social disorder is the characteristic of neurodevelopmental disorders such as autism. the discovery of these vocal patterns has laid a foundation for the study of social disorder related diseases. Ultrasonic signals associated with different types of social behavior of mice: 1