Scientists at the University of Michigan have developed an open-source, user-friendly, artificial intelligence driven software called LabGym that automatizes animal behavior analysis in various model systems and could be a boon to life scientists across the spectrum of basic science and drug development.
The findings were published in the article”LabGym: quantification of user-defined animal behaviors using learning-based holistic assessment,” in the journal Cell Reports Methods on February 24.
Measuring animal behavior is instrumental in understanding fundamental neural processes as well as assessing therapeutic and adverse effects of drugs. Bing Ye, PhD, professor of life sciences at the University of Michigan, and his team analyze movements and behaviors in the model organism Drosophila melanogaster (fruit flies) to understand mechanisms involved in the development and function of the nervous system in humans.
“Behavior is a function of the brain. So, analyzing animal behavior provides essential information about how the brain works and how it changes in response to disease,” said Yujia Hu, a neuroscientist in Ye’s lab and lead author of the study.
Recognizing and scoring facets of animal behavior manually is tedious, time-consuming and prone to human error. A few programs exist that automate the quantitative assessment of animal behaviors, but they present challenges.
“Many of these behavior analysis programs are based on pre-set definitions of a behavior,” said Ye. “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.”
Getting the program to think like a scientist
To overcome these challenges, Hu and his colleagues decided to design a new program that more closely replicates the human cognitive process—that “thinks” more like a scientist would—and is more user-friendly for biologists who may not have expertise in coding. Using LabGym, researchers can input examples of the behavior they want to analyze and teach the software what it should count. The program then uses deep learning to improve its ability to recognize and quantify the behavior.
LabGym exploits a combination of video and ‘pattern image’ data to gain cognitive flexibility and reliability. Time series data alone, as obtained through video recordings can be challenging for AI programs to analyze. To train LabGym to identify behaviors better, Hu generated images that depict the pattern of the animal’s movement by merging outlines of the animal’s position at different timepoints. Combining video data with the pattern images increased the program’s accuracy in recognizing different behaviors.
LabGym not only tracks multiple animals simultaneously, it is also designed to disregard irrelevant background information and consider both the animal’s overall movement and the changes in position over space and time.
Another advantage of LabGym is its species flexibility. Despite being designed using Drosophila, it is not limited to any one species. “That’s actually rare,” said Ye. “It’s 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.”
Carrie Ferrario, PhD, an associate professor of pharmacology, who studies neural mechanisms that contribute to addiction and obesity in rat models, helped Ye and his team test and refine the program in the rodent model system. “I’ve been trying to solve this problem since graduate school, and the technology just wasn’t there, in terms of artificial intelligence, deep learning and computation,” said Ferrario. “This program solved an existing problem for me, but it also has really broad utility. I see the potential for it to be useful in almost limitless conditions to analyze animal behavior.”
In future studies, Ye’s team plans to refine the program further to improve its performance under more complex conditions, such as assessing animal behavior in natural environments.