A new open-source tool harnesses the power of Artificial Intelligence & simplifies the analysis of Animal Behavior

Artificial Intelligence 

A team at the University of Michigan has developed a new software tool to help life scientists analyze animal behavior more effectively.

   Photo credit : The US sun | AI development that robots can now talk to animals 

The open-source software LabGym uses artificial intelligence to identify, classify, and count specific behaviors in various animal model systems.

Scientists need to measure animal behavior for a variety of reasons, from understanding all the ways a given drug can affect the body to mapping how circuits in the brain communicate to produce a particular behavior.

Scientists in the lab are analyzing the movements and behaviors of Drosophila melanogaster - or fruit flies - as a model for studying the development and function of the nervous system. Because fruit flies and humans share many genes, these fruit fly studies often provide insight into human health and disease.

“Behavior is a function of the brain. Therefore, analyzing animal behavior provides important information about brain function and how it changes in response to disease," said Yujia Hu, a neuroscientist in Yes's lab at the Institute of Science of U-M Life and lead author of the Feb. 24 Cell Reports Methods study, which published the describes new software.

However, identifying and manually counting animal behavior is time-consuming and highly subjective to the researcher analyzing the behavior. And while there are several programs to automatically quantify animal behavior, they present challenges.

    Photo Credit: The guardian

To address these challenges, Hu and his colleagues set out to create a new program that better mimics the human cognition process one that "thinks" more like a scientist and that's easier for biologists. Researchers can capture examples of the behavior they want to analyze and teach the software what to count. The program then uses deep learning to improve its ability to recognize and quantify behavior.

One of the new LabGym developments that help apply this more flexible knowledge is the use of both video data and the so-called model image to improve program reliability. Scientists use videos of animals to analyze their behavior, but the videos contain time series that are difficult for AI programs to analyze.

To help the program recognize behaviors more easily, Hu created a still image showing the animal's movement pattern by combining outlines of the animal's position at different times. The team found that combining video data with model images increased the program's accuracy in recognizing behavioral patterns.

The program can also track multiple animals at the same time. The team therefore plans to further refine the program to improve its performance in even more complex contexts such as observing animals in the wild. These studies were supported by the National Institutes of Health.