LabGym: The AI Software Accurately Recognizes & Classifies Animal behavior Across Species

The scientists at the University of Michigan have created a user-friendly, artificial intelligence-driven software called LabGym. The open-source program that automatizes animal behavior analysis in various model systems could transform basic science and the process of drug development.

The findings from the software were recently published in the article “LabGym: quantification of user-defined animal behaviors using learning-based holistic assessment,” in the journal Cell Reports Methods.

Recommended: Virtual Cards for Your Business – How They’re Beneficial For Security

To understand the fundamental neural processes and assess the adverse effects of drugs, classifying animal behavior is extremely vital.

Bing Ye, Ph.D., professor of life sciences, University of Michigan, along with his team monitors movements in the model organism Drosophila melanogaster (fruit flies) to recognize mechanisms that are involved in the nervous system development in humans.

Explaining the concept further, Yujia Hu, a neuroscientist in Ye’s lab and lead author of the study, said that as behavior comes under the function of the brain, so analyzing animal behavior with the same logic helps to gather information about how the brain works and changes when the response to disease.

Understanding the different characteristics of animal behavior manually can be exhaustive, time-consuming, and prone to errors. Also, the existing programs to automate the quantitative assessment of animal behaviors come with challenges.

Recommended: Quantexa Acquires Aylien an Award-Winning Provider of AI-Based Risk & Market Intelligence Solutions 

Ye explained that most of the existing analysis programs are based on ‘pre-set definitions of a behavior.’

“If a Drosophila larva rolls 360 degrees, for example, some programs will count a roll. But why isn’t 270 degrees also a roll? Many programs don’t necessarily have the flexibility to count that, without the user knowing how to recode the program.”

LabGym – The New Program That Thinks Like a Scientist

To overcome these challenges, Hu and his colleagues designed a new program that closely replicated the human cognitive process. They decided to build a program that thinks like a scientist and is user-friendly for biologists who may not be that well-versed in coding.

• With LabGym, researchers can enter the details of the behavior they wish to analyze and instruct the software on what to count. With the help of deep learning, the program improves its ability to recognize and quantify behavior.
• LabGym works on a combination of video and ‘pattern image’ data for cognitive flexibility as well as reliability.
• AI programs may not be very successful to extract time series data from video recordings. And so, to train LabGym to identify behaviors better, Hu created different images that depicted the pattern of the animal’s movement. He merged the outlines of the animal’s position at different time points. Using a combination of video data with the pattern images drastically enhanced the program’s accuracy.
• Another advantage of using LabGym is that it not only tracks multiple animals simultaneously but can also disregard irrelevant background information. LabGym also enables us to consider both the animal’s overall movement and the changes in position over space and time.

LabGym offers species flexibility which means you can study different species at the same time. Though it is designed using Drosophila, species are not restricted. Ye, while highlighting how scientists don’t require any technical expertise, says the program is “written for biologists, so they can adapt it to the species and the behavior they want to study without needing any programming skills or high-powered computing.”

Ye and his team plan to enhance the program’s performance under more complex conditions, like assessing animal behavior in natural environments.