Brainana: an end-to-end preprocessing framework for macaque neuroimaging
Brainana is an automated, containerized, and cloud-accessible framework designed to standardize and streamline the end-to-end preprocessing of macaque neuroimaging data by integrating deep learning models, anatomical optimizations, and quality control to ensure reproducible, scalable, and cross-study comparable results.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
The brain of a macaque monkey is a powerful bridge in neuroscience. It connects the broad, non-invasive maps we can make of the human brain with the tiny, cellular details we can only see by looking inside a living animal. Researchers use these monkeys to test ideas about how the brain is wired, hoping those ideas will explain how human minds work. To do this, they take detailed pictures of the monkey's brain using magnetic resonance imaging, a technology that creates clear images of soft tissue without surgery. But turning those raw pictures into useful scientific knowledge is incredibly difficult. Unlike human scans, which often follow a standard format, monkey scans come from many different labs, using different machines, different positioning, and different settings. This variety creates a chaotic mix of data that is hard to compare. Without a way to clean up and standardize these images, the valuable information inside them often stays locked away, usable only by the specific team that collected it.
A team of researchers has now built a solution to this problem called Brainana. It is a complete, automated system designed to take raw brain scans from macaques and turn them into clean, standardized, and easy-to-use data. Before this tool existed, scientists had to piece together many different software programs, often writing their own custom code to make them work together. This process was slow, prone to errors, and difficult to repeat. Brainana replaces that patchwork with a single, unified pipeline that handles every step of the job. It takes the messy, raw images, fixes common errors like tilted orientations or missing parts of the brain, and then creates a detailed 3D map of the brain's surface. Crucially, it does all of this while keeping track of exactly how the data was changed, allowing researchers to compare results from one monkey to another, or from one lab to another, with confidence.
The system was tested on data from 130 monkeys collected across 23 different imaging sites. These sites used a wide variety of scanners and protocols, representing the full range of conditions found in real-world research. The tool successfully processed this diverse data, producing consistent results that matched known anatomical patterns. For example, it correctly identified that rhesus macaques have larger brains, averaging 84 milliliters, while cynomolgus macaques have smaller brains, averaging 62 milliliters. It also mapped the thickness of the brain's outer layer, finding it thicker in the front of the brain and thinner in the back, a pattern that matches previous studies. When the researchers looked at brain activity during tasks, the system preserved the precise locations where the brain responded to faces, bodies, and objects. It also maintained the natural patterns of connection between different brain regions when the animals were at rest. These results show that the tool can handle the messy reality of scientific data without losing the important details.
What makes Brainana particularly useful is that it does not just produce numbers; it produces data that humans can actually see and understand. The researchers built a companion viewer that automatically organizes all the output files. A scientist can open this viewer and immediately see the brain in three dimensions, with the surface of the brain linked to the inside slices. They can click on a spot on the surface and instantly see the corresponding location in the volume, along with labels that identify the brain region. This removes the need for specialized training in complex file formats. A researcher planning an experiment can use the tool to see exactly where an electrode should go, or a clinician can look at a scan to understand how a specific brain area relates to a behavior, without needing to be an expert in image processing software.
The system also solves a specific headache common in monkey research: the way the animals are positioned during scans. Because monkeys are often held in place with head posts or other hardware, their brains can appear tilted or cut off in the images. Standard computer programs often fail when faced with these odd angles. Brainana includes a special step that gently rotates and aligns the image into a standard position without distorting the shape of the brain. This simple fix allows the rest of the automated process to work correctly, even on difficult scans. The tool also uses advanced computer models trained specifically on monkey brains to separate different types of tissue, such as gray matter and white matter, with high accuracy. In tests, these models correctly labeled more than 90 percent of the brain's tiny building blocks, even in areas where the tissue is thin or hard to see.
By making the process of cleaning and analyzing monkey brain scans automatic and accessible, this work opens the door for much larger collaborations. Scientists can now share their data more easily, knowing that their results will be processed in the same way as everyone else's. This allows for the pooling of data from many different studies to answer questions that are too big for a single lab to tackle alone. The tool is available to the public as a package that can run on standard computers or in the cloud, meaning that even researchers without powerful supercomputers can use it. It represents a shift from isolated, custom-made solutions to a shared, reliable infrastructure for understanding the primate brain.
The researchers emphasize that this tool is designed to work alongside other methods, not to replace them. It handles the heavy lifting of preparation, freeing scientists to focus on the biological questions they want to answer. Whether the goal is to link a brain scan to a recording from a single neuron, or to compare brain structure across different species, having a clean, standardized starting point is essential. The work demonstrates that the diversity of data in macaque neuroscience, once seen as a barrier, can be managed effectively. With this new framework, the field can move forward with a clearer, more unified view of how the brain is organized, bringing us closer to understanding the mechanisms that underlie both monkey and human behavior.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.