AI Helps Spot Exercise Limits Without Guesswork

Deep learning can read the breath and heart data that people use to find their exercise limits. In a test called a cardiopulmonary exercise test, doctors look at how the body uses oxygen and releases carbon dioxide. They mark two key points: one where lactate starts to build up, and another where breathing changes sharply. Traditionally, experts draw these points by eye, which can vary from person to person. A new approach trains a neural network on more than 1, 200 of these tests. The model learns the patterns that signal each threshold and can then point them out automatically. Because it works the same way every time, it removes the subjectivity that comes from human judgment.Researchers compared the machine’s calls with those made by experienced clinicians. The results showed that the AI matched or even outperformed the human experts in locating both thresholds. This suggests that automated analysis could become a reliable tool for sports scientists and medical teams. Beyond speed, the system offers consistency. Every test is evaluated with the same algorithm, reducing errors that arise from fatigue or bias. It also frees specialists to focus on interpreting results and planning training or treatment plans. Future work may integrate this technology into wearable devices, giving athletes real‑time feedback on their limits. If proven safe and accurate, AI could help people push harder while staying within healthy boundaries.