Researchers have unveiled a groundbreaking method that significantly improves the accuracy of brain-state classification using functional near-infrared spectroscopy (fNIRS). This innovative approach allows for enhanced measurement of neural activity by tracking changes in blood flow and oxygen levels in the brain, which are indicative of active brain regions. fNIRS is a non-invasive brain-imaging technique that provides a portable and cost-effective means of monitoring brain function, even when patients are in motion.
Revolutionizing Brain Imaging with Riemannian Geometry
Despite its advantages, traditional methods for analyzing fNIRS data lag behind those used with other brain-imaging technologies. To address this gap, an international research team has tailored a new classification method that leverages the unique characteristics of fNIRS signals. Unlike other imaging techniques, fNIRS captures both oxygenated and deoxygenated blood, offering complementary insights into brain activity.
Tim Näher, the first author of the study based at the Max Planck Institute for Biological Cybernetics in Tübingen, Germany, stated, “These two signals naturally show opposite patterns, but that doesn’t mean they are redundant.” The team’s research harnesses advanced mathematical tools from Riemannian geometry to interpret these dual signals effectively.
In their study, participants were tasked with performing eight different mental activities, including tasks like imagining playing tennis or rotating an object mentally. The new computational framework demonstrated exceptional accuracy in identifying which task participants were engaged in, outperforming traditional fNIRS analysis methods.
Implications for Disorders of Consciousness
The implications of this research extend beyond academic interest; it holds potential for transforming the diagnosis and treatment of disorders of consciousness. Patients with such conditions often lack the ability to move or communicate, complicating efforts to assess their level of awareness.
Näher collaborated with Lisa Bastian from the University of Tübingen on a separate study conducted at Maastricht University. This research aimed to determine whether a non-responsive patient could still demonstrate signs of consciousness. In this study, healthy participants either engaged in a mental task or remained inactive.
The novel fNIRS paradigm, combined with the Riemannian geometry-based analysis, proved highly effective. It accurately identified responsiveness in all cases tested and recognized unresponsiveness in nine out of ten instances. Näher remarked, “So far, we provided a proof of concept that the new fNIRS framework can serve as a fast, objective, and accessible tool to support more reliable diagnoses and improve treatment decisions for disorders of consciousness.”
Both studies were published in the journal Neurophotonics in 2025, showcasing the potential of fNIRS technology in enhancing our understanding of brain function and addressing critical health challenges. The next step for the research team is to implement their method in real patient scenarios, which could pave the way for significant advancements in medical diagnostics.
