With millions of tuberculosis (TB) cases left undiagnosed each year due to limited healthcare access or incorrect diagnoses, researchers are turning to artificial intelligence (AI) to help bridge the gap. A new AI tool from Google, called Health Acoustic Representations (HeAR), holds promise for detecting TB and other respiratory diseases early. Using the sounds of coughing and breathing, HeAR could provide a fast, low-cost alternative to traditional diagnostics.
HeAR’s Potential as an Acoustic Health Tool
HeAR has been trained on more than 300 million audio clips, from coughs to throat clearings, to analyze the human body’s unique sounds. By converting these sounds into spectrograms, Google researchers created a robust database for AI training. They trained HeAR to predict missing snippets of these audio patterns. The technology is similar to how large language models are trained on vast amounts of human text. This technology holds the potential to monitor respiratory health, detect diseases like TB, and evaluate lung function in a cost-effective way.
HeAR’s purpose has been clear from the start: create an accessible diagnostic tool that can work where medical imaging tools are unavailable. Its development stemmed from research during the pandemic when scientists found that COVID-19 could be identified by the sound of a cough.
Improving Diagnosis with Data
Google’s library of 100 million cough sounds has allowed HeAR to achieve high accuracy across diverse tasks. HeAR is particularly effective in areas with limited data, a challenge that often hinders healthcare research. The technology’s ability to function with less data and minimal computational resources is critical for communities where healthcare facilities are scarce.
HeAR’s goal is to be a flexible foundation for further bioacoustic studies, supporting the development of custom models for specific conditions, regions, and demographics. Google’s partnership with Salcit Technologies, an AI healthcare startup in India, reflects this mission. Salcit’s Swaasa® technology already uses AI to analyze cough sounds and other information for lung health. The company is now incorporating HeAR to improve early TB detection across India. HeAR could support a scalable, location-independent method for respiratory health assessments, potentially offering underserved populations a means of diagnosis without the need for expensive equipment.
A Step Toward Accessible, AI-Based Health Monitoring
The potential of HeAR to detect TB in a simple, non-invasive way could have a global impact. Stop TB Partnership, an organization hosted by the United Nations with the mission of ending TB by 2040, has reviewed HeAR’s performance, comparing it with other cough-based AI algorithms. Among the group’s conclusions is that “AI cough classification holds potential as an accessible, low-impact means of respiratory monitoring, suitable for resource-limited settings.”
Unlike conventional diagnostics, which may rely on costly X-rays or biopsies, HeAR represents a new era of “acoustic biomarkers” capable of providing valuable health insights at minimal cost. AI allows HeAR to process cough sounds from a basic smartphone, making it both accessible and affordable. This innovation could particularly benefit populations in remote areas, where a diagnosis might otherwise require extensive travel or resources.
The Expanding Field of Bioacoustic Research
While Google’s HeAR is still in its early stages, its potential is promising. The model is now available for researchers to build on its foundation, driving forward bioacoustic research and expanding the possibilities for future diagnostics in respiratory health.
The project signals a shift in global healthcare research. Instead of only investing in physical diagnostic equipment, Google’s research team envisions a future where HeAR’s technology is integrated into various healthcare systems, offering early, affordable diagnoses. “Every missed case of tuberculosis is a tragedy; every late diagnosis, a heartbreak,” says Sujay Kakarmath, a product manager at Google Research working on HeAR. “Acoustic biomarkers offer the potential to rewrite this narrative.”
Did you enjoy this blog post? Check out our other blog posts as well as related topics on our Webinar page.
QPS is a GLP- and GCP-compliant contract research organization (CRO) delivering the highest grade of discovery, preclinical and clinical drug research development services. Since 1995, it has grown from a tiny bioanalysis shop to a full-service CRO with 1,100+ employees in the US, Europe and Asia. Today, QPS offers expanded pharmaceutical contract R&D services with special expertise in pharmacology, DMPK, toxicology, bioanalysis, translational medicine, cell therapy (including PBMCs, leukopaks and cell therapy products) and clinical development. An award-winning leader focused on bioanalytics and clinical trials, QPS is known for proven quality standards, technical expertise, a flexible approach to research, client satisfaction and turnkey laboratories and facilities. Through continual enhancements in capacities and resources, QPS stands tall in its commitment to delivering superior quality, skilled performance and trusted service to its valued customers. For more information, visit www.qps.com or email info@qps.com.
