In a major leap for neuroscience, biomedical engineers at Duke University have unveiled a revolutionary 3D imaging technique designed to effectively map and categorize the intricate social behaviors of animal models. This cutting-edge method enables researchers to quantitatively analyze the movements, interactions, and physical contacts among rodents, providing unprecedented insights into how various genetic forms of autism influence social behavior in rats.

Published in the esteemed journal Cell, this research promises to pave the way for a deeper understanding of a broad spectrum of neuropsychiatric disorders through enhanced observational capabilities in laboratory animals.

Historically, neuroscience has been equipped with a growing array of high-resolution imaging tools capable of tracking neuronal activity, such as calcium imaging and CRISPR technology, which allow for precise manipulation of brain functions. However, until now, the tools for effectively quantifying movement and behavior—critical outputs of brain activity—have lagged behind.

Timothy Dunn, an assistant professor of biomedical engineering, emphasizes this gap: “Despite movement and behavior being the principal outputs of the brain, tools to quantitatively measure and track that output were almost an afterthought. Without precise measurement, we cannot accurately gauge how diseases or treatments affect behavior.”

Traditionally, researchers relied on rudimentary methods to observe animal behavior, including manual scoring of actions or simplistic imaging techniques that tracked an animal's position over time. This limited the overflow of data capable of driving forward crucial scientific discoveries.

In response, Dunn and his team introduced an innovative system known as DANNCE (3-Dimensional Aligned Neural Network for Computational Ethology) in 2021. By utilizing video footage of freely moving rats, they trained machine-learning algorithms to identify and map the intricate 3D locations of the animals’ joints. By correlating this data with brain recordings, they made significant strides in linking neuronal activity to behavior.

Now, with their latest development—social-DANNCE (s-DANNCE)—they have broadened their innovative approach to capture the nuances of social interactions between animals. Dunn explains the complexity: "Tracking social movements is a significant challenge as computer vision struggles to differentiate and follow animals that often overlap and resemble each other. Furthermore, separating individual actions from collective social behaviors is far from straightforward, given the subtleties involved.”

The researchers captured videos of groups of two to three rats engaging freely in a controlled environment and then employed their neural network to analyze these interactions. By constructing 3D models of the animals' joints from their movements, they successfully classified a range of behaviors such as grooming, chasing, sniffing, and even aggressive interactions like fighting.

Results revealed an astonishing breadth of communication among rats—Dunn's team identified hundreds of distinct social behaviors, providing new quantitative units to assess how social interactions are altered in various disease models or under the influence of medications.

In one instance, the researchers tested their model against the effects of amphetamines, which commonly induce hyperactive behaviors in humans and animals alike. The findings indicated not only an increase in the overall activity of the rats but also significant disruptions in their social dynamics and physical communications.

Additionally, the team examined different genetic models of autism, successfully detecting alterations in social behaviors and touch patterns within these subjects.

As part of their commitment to advancing the field, the researchers have made the s-DANNCE platform and a comprehensive dataset encompassing over 150 million 3D behavioral samples freely accessible to other researchers.

“Many areas of neuroscience have been hindered by the absence of precise, objective, and reproducible behavioral descriptions,” Dunn emphasizes. “Our tool addresses this persistent challenge. We believe this technology, paired with our extensive library of social interactions, will catalyze new research linking social behavior to brain functions and neuropsychiatric mechanisms.”

This pioneering work not only transforms our understanding of animal behavior but also opens up new avenues for exploring the mysterious connections between the brain and social dynamics. Stay tuned as this technology continues to redefine the landscape of behavioral neuroscience!