Researchers are increasingly interested in how everyday sounds, such as coughing, can offer clues about our health. As part of the NIH Common Fund’s Bridge2AI Program, which focuses on the use of complex biomedical data for future artificial intelligence (AI) tools, Anaïs Rameau, MD, the Bridge2AI Voice AI study team, and their colleagues, examined whether cough sounds contain patterns that can be linked to gender. This question matters because coughs are natural, hard to control, and are widely recorded in health settings; understanding what information they carry is important for both scientific discovery but also patient privacy.

A cough happens in several quick steps: a breath in, pressure building in the chest, a sudden burst of air called the explosive phase, and a fading tail as the sound trails off. The researchers analyzed 327 publicly available cough recordings and used an AI model to examine which parts of the cough provided the clearest cues about gender. The study team found that the explosive burst—the brief, forceful moment when air is rapidly released—had the clearest acoustic cues the model could use to distinguish gender, more so than traditional audio measures. 

The findings do not suggest that gender should be predicted from coughs or that such tools are ready for real-world use. Instead, the study shows that different parts of a cough carry different types of information, and that AI methods can help identify where these acoustic signals appear.  This matters for privacy because, unlike speech, coughs are hard to control on purpose, so they might reveal gender-related cues even when someone does not want them to. It is also important to note that the study used a relatively small number of recordings from one dataset, so more research with larger groups is needed before drawing firm conclusions. Despite the limitations, by mapping what information may naturally exist in common health sounds like coughs, this research contributes to Bridge2AI’s broader effort to prepare for responsible, transparent AI approaches in future biomedical research.

Reference: 

He L, Li H, Wang S, Baik E, Kervin S, Zhao R, Ramos JM; Bridge2AI‐Voice Consortium; Colonel J, Rameau A. Decoding Gender in Cough Sounds: A Transformer-Based Analysis. Laryngoscope. 2026 Jun;136(6):2519-2527. doi: 10.1002/lary.70393. Epub 2026 Jan 26. PMID: 41588706; PMCID: PMC13032795.