AntScan: High-Resolution 3D Ant Anatomy from Synchrotron CT Scans

Science Friday profiles AntScan, a collaboration led by Flora Lichtman and researcher Dr. Julian Katzka, who use a synchrotron-based micro-CT scanner to generate high-resolution 3D images of real ants from 700 species. The data reveal inner anatomy beyond what the naked eye can see, from exoskeleton to muscles and nerves, with voxel sizes down to 1.22 micrometers. The project, which scanned ants in batches of 50 and required night shifts to complete a week of imaging, makes 3D data freely accessible for evolutionary biology, biodiversity, and even 3D animation and video game design. The researchers envision AI-assisted annotation to distinguish tissues and feed into phylogenetic analyses, enabling new insights into ant evolution and biomechanics.

Overview of AntScan and the Tech Behind It

AntScan is a pilot study that uses a synchrotron light source to perform micro CT scans on ants, producing high-resolution 3D reconstructions from grayscale tomography slices. The team, including Flora Lichtman and Dr. Julian Katzka, leverages fast imaging and automated sample handling to scan thousands of individuals across hundreds of species. The data capture not just the external form but the squishy interiors—muscles, guts, nerves—and even tiny hairs on the exoskeleton, giving researchers a detailed look at ant anatomy that is usually beyond ordinary microscopy.

Quote: "The data that we end up with are like slices of images that are like grayscale. And so they contain the anatomy of the ants." - Dr. Julian Katzka

From X-ray Slices to 3D Volumes

Images originate from X-ray tomography, where multiple scans of an ant are mathematically composed into a 3D volume. The resolution is extraordinary: a voxel of 1.22 micrometers allows visualization of delicate hairs and, on the interior, single muscle fibers. This capability opens up possibilities to study brain structure, musculature, and other internal features that underpin ant behavior and strength.

Quote: "the absolute resolution of like a voxel, which is a 3D pixel, is 1.22 micrometers." - Dr. Julian Katzka

Workflow, Speed, and Scale

Compared with traditional lab micro CT, the AntScan setup is dramatically faster. One ant takes about 30 seconds to image, plus another 30 seconds to transfer data. In practice, batches of about 50 ants were scanned, requiring some night shifts to complete within a week. The system combines a robot for automatic sample exchange and a high-speed camera to keep pace with throughput while ensuring safety from high-energy X-rays.

Quote: "Having 3D data to look inside of them, that makes it a lot easier." - Flora Lichtman

Why Mouthparts and Ant Diversity Matter

The project focuses on ant mouthparts because these structures are key tools for daily life and vary dramatically across species. The richness of ant forms is well known, and the detailed 3D data enable researchers to examine evolutionary patterns across thousands of individuals, shedding light on how these traits evolved in relation to ecology and behavior.

Public Access, AI, and the Next Steps

One of AntScan’s goals is to make the data freely accessible to both scientists and the public. In addition to large-scale evolutionary questions, the data are poised to support AI developments such as tissue segmentation and anatomy tagging. Annotating the dataset to train models could enable automated recognition of exoskeleton, muscles, and nervous tissue, accelerating analysis and feeding into phylogenetic studies. The team envisions applying this workflow to other small invertebrates and expanding to a broader range of taxa while keeping scan-time bottlenecks in check.

Quote: "Yes, for sure. I think as soon as we're done here, I will have to sit down and do a bunch of annotations to train these models." - Dr. Julian Katzka

Broader Implications and Future Vision

Beyond ants, the AntScan approach demonstrates how high-resolution 3D data can support biology, education, and digital media. The architectural efficiency of scanning and processing paves the way for scalable projects across diverse small organisms. The collected 3D datasets also promise to fuel the development of AI-assisted tools for anatomy recognition and animation, potentially transforming both scientific workflows and public engagement with micro-scale life.