Ever wondered how scientists study something as complex as social behavior in animals? It’s not easy! Observing animals interacting is one thing, but truly understanding the nuances of their communication, their subtle movements, and the patterns in their social lives requires specialized tools. Until recently, we’ve lacked the technology to capture these details in a way that’s both precise and measurable.

This is where a groundbreaking new approach comes in, offering a fascinating window into the social world of animals like rats and mice. Imagine being able to track every twitch of a whisker, every shift in posture, every instance of touch between individuals – all in three dimensions and with incredible accuracy. That’s precisely what this new technique allows.

So, how does it work? It’s a clever combination of cutting-edge technologies:

• 3D Geometric Reasoning: This helps the system understand the spatial relationships between the animals and their environment, even when they’re partially hidden from view (a common problem when observing groups).
• Graph Neural Networks: These powerful algorithms learn the complex patterns of movement and interaction, allowing researchers to identify meaningful social behaviors.
• Semi-Supervised Learning: This approach uses a small amount of labeled data (e.g., identifying specific behaviors like grooming or fighting) to train the system, which can then analyze vast amounts of unlabeled data on its own.

Researchers used this technique to collect a massive dataset of over 110 million 3D pose samples from interacting rats and mice, including several strains of rats that model autism. This enormous library of information provides a treasure trove for understanding the subtleties of social interaction. By applying a “multi-scale embedding approach,” they were able to uncover a rich tapestry of behaviors, from individual actions to complex social dynamics.

Think of it like zooming in and out on a map. At one level, you see individual movements and postures. Zooming out reveals patterns of interaction between individuals, like synchronized movements or specific types of physical contact. Zooming out even further allows scientists to see the overall structure of social interactions within the group.

This level of detail is crucial for understanding how social behavior is altered in neuropsychiatric disorders like autism. The study found a range of differences in the social behavior of the autism model rats, differences that conventional methods would likely have missed. For example, they might show less social interaction overall, different patterns of touch, or changes in the timing and coordination of their movements. The research also explored how amphetamine, a drug known to affect social behavior, influences these intricate patterns of interaction.

This new technology and the vast dataset it has generated offer an unprecedented opportunity for scientists. It opens doors to a deeper understanding of:

• The basic principles of social behavior: How do animals communicate? How do they form social bonds? How do their social interactions adapt to changing environments?
• The neurological basis of social behavior: Which brain regions and circuits are involved in social interactions? How do genes influence these circuits and, ultimately, behavior?
• Neuropsychiatric disorders: How are social behaviors disrupted in conditions like autism? Can these detailed observations lead to new diagnostic tools or therapeutic strategies?

By providing such a high-resolution view of social interaction, this research promises to revolutionize our understanding of the social lives of animals and shed light on the complexities of social behavior in both health and disease. It’s a powerful example of how innovative technologies can unlock new avenues of scientific discovery.