Porcine-CaMo multimodal wearable ECG, kinematics and heat- production data in respirometry chambers
The Porcine-CaMo dataset is a dual-cohort multimodal resource from China Agricultural University that integrates continuously recorded wearable ECG and kinematic signals with 5-minute heat-production data from growing pigs in respirometry chambers to support research in porcine metabolic modeling, heart rate variability analysis, and cross-cohort method evaluation.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Measuring how much energy an animal burns is a fundamental question in biology, one that helps scientists understand everything from how efficiently a pig converts feed into meat to how a large animal manages its body temperature. For decades, the most reliable way to get this answer has been to place an animal inside a sealed, climate-controlled room and carefully measure the oxygen it breathes in and the carbon dioxide it breathes out. This method, known as indirect calorimetry, acts as a gold standard for calculating heat production. However, these rooms are expensive, difficult to maintain, and often too small or restrictive to let animals move naturally, which means the data collected might not reflect how the animal behaves in the real world. Scientists have long wanted a simpler way to estimate this energy use by looking at signals the animal gives off naturally, such as its heartbeat and its movements, but they have lacked a large, high-quality set of data to teach computers how to make that connection accurately.
A team of researchers at China Agricultural University has now filled this gap by creating a massive, detailed record of growing pigs moving freely inside these special rooms. They equipped the animals with small, wearable devices that acted like a continuous diary of their lives, recording their heartbeats, their body movements, and their heat output simultaneously. The result is a dataset called Porcine-CaMo, which captures over 150 days of continuous data from two different groups of pigs. One group consisted of larger pigs observed over ten separate four-day periods, while the other group was made up of smaller pigs monitored over a twenty-two-day window. By combining these wearable records with the precise measurements from the rooms, the researchers have provided a new reference library that allows others to build better models for predicting energy use without needing to lock animals in a chamber.
The wearable technology used in this study was designed to be unobtrusive and robust. Strapped securely to the pigs' backs, these devices recorded the electrical activity of the heart five hundred and twelve times every second, capturing the precise moment each beat occurred. At the same time, they tracked the animal's motion ten times a second, noting how the pig rotated and accelerated in three dimensions. This high-speed recording is crucial because it allows scientists to see not just that the animal moved, but exactly how the movement correlated with the heart's rhythm. The researchers also ensured the devices were safe for the animals, using detachable wires to prevent damage if a pig chewed on them, and the animals were free to eat, drink, and rest normally throughout the experiment.
To make sense of this flood of information, the team developed a sophisticated computer system to clean and organize the data. The raw heartbeat signals often contain noise caused by the animal's movement, so the researchers used a specialized artificial intelligence model to identify the true heartbeats and remove the false ones. They then grouped the data into five-minute blocks, matching the heart rate and movement patterns to the heat production measurements taken from the room's air sensors. This process created a synchronized timeline where every change in the pig's activity or heart rate could be directly compared to a change in the energy it was burning. The team released two versions of the data: a "corrected" version that is suitable for most research, and a "perfect" version that has been manually checked by experts to ensure every single heartbeat is labeled correctly, providing a gold standard for training new computer programs.
The study revealed that the data is consistent and reliable enough to be used for complex analysis. The researchers checked that the heart rate patterns followed natural daily cycles, rising and falling with the pigs' activity levels, and confirmed that the heat production numbers made biological sense, such as showing lower energy use on days when the pigs were not fed. They also noted that while the two groups of pigs were different in size and the equipment used for them had slight variations, the core signals remained clear and comparable. This consistency is vital because it means the data can be used to test whether a model trained on one group of pigs can work on another, helping to build more universal tools for animal science.
This work does not claim to have solved the problem of measuring animal energy use on its own, but it provides the essential raw material needed to do so. By making this dataset and the software used to process it freely available, the researchers have lowered the barrier for other scientists to develop better prediction tools. Instead of needing to build expensive chambers for every new study, future researchers can use these wearable signals to estimate heat production with greater accuracy. The dataset includes not just the numbers, but also the code and instructions needed to replicate the analysis, ensuring that the scientific community can build upon this foundation to improve animal nutrition, welfare, and breeding practices.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.