Research progress on Computational Ethology

Animal behavior, mostly controlled by the central brain, has been studied in natural environments and laboratory settings. In early 20th century, researchers behavior to reveal how it is built from components organized over time response stimuli. settings, study ability of brain generate behaviors rewards punishments. However, there are limitations quantifying animal these two approaches. Recent advances breakthroughs computer science provide an important opportunity for overcoming behavioral studies. this review, we focus on emerging new discipline called “Computational Ethology”, which uses a wide variety techniques, including vision machine learning, measure analyze patterns behaviors. Computational Ethology allows quantitative analyses with high efficiency, made significant progress recent years towards better understanding as well their underlying neural mechanism. Over last decade, application artificial intelligence study, numerous methods have developed automatically quantify automatic tracking movements. Classical estimate centroids ellipses animals, reflect locomotion orientation, respectively. This estimation was then extended multiple animals. Nevertheless, non-locomotion information pose cannot be reliably captured. To solve problem, deep learning-based setting various species. present pipeline collecting analyzing data. High-dimensional raw data first subjected dimension reduction. The low-dimensional segmented analyzed supervised classification or unsupervised clustering order produce modules. After this, transition probabilities between modules calculated elucidate pattern We also review diagnosing evaluating neuropsychiatric diseases discuss opportunities challenges Ethology. Ethology, result substantial interdisciplinary research neuroscience, psychology, physics, science, ethology great potential deeper nature still its infancy many questions remain explored. provides summary reference resource yet rapidly advancing discipline. hope that will informative useful scientific community studying intelligence.