Machine Learning Spots Hidden Balance Loss at Moderate Altitudes Before

Researchers say the body’s balance system may start failing at altitudes much lower than the point when common hypoxia symptoms typically show up. A study in Frontiers in Bioengineering and Biotechnology used tree-based machine learning alongside classical statistics to detect measurable postural instability at simulated 3,000 meters—below the roughly 5,000-meter level linked to symptoms such as color vision loss. The team, led by Yasaman Baradaran of the BioSIS Lab at the University of Canberra with colleagues at Embry-Riddle Aeronautical University, tested 31 healthy adults aged 19–23 in a normobaric hypoxia chamber at sea level, 3,000, 4,000 and 5,000 meters. Center-of-pressure data from standing balance tasks were recorded for analysis.