Dr. Sebastian Buschjäger, Scientific Coordinator and Researcher on Resource-aware Machine Learning at the Lamarr Institute, recently participated in the 32nd Annual Meeting of the German Sleep Society (DGSM) in Essen. Showcasing advancements in AI in sleep medicine, Buschjäger presented cutting-edge applications aimed at transforming patient care and clinical workflows.
During the pre-conference workshop, Buschjäger delivered a compelling talk highlighting the critical importance of high-quality data in developing robust AI models for clinical applications. His insights underscored how AI in sleep medicine relies on precise and reliable data to effectively diagnose and treat sleep disorders.
At the main conference, Buschjäger and Dr. rer. medic. Sarah Dietz-Terjung presented their innovative Sleepwalker tool. This project, originally designed to predict optimal therapies for patients with sleep apnea, revealed an initial need to enhance the (semi-)automatic annotation processes widely used in clinical practice. By extending the U-sleep model’s capabilities to include tasks such as apnea and arousal detection, the team has taken a significant step toward developing AI models that effectively support clinical workflows in sleep medicine.
Building on this progress, the team is now transitioning their analysis from a controlled laboratory setting to in-home environments. This approach enables patients to sleep naturally at home with significantly fewer sensors while maintaining accurate data collection – a leap forward for AI in sleep medicine. The methodology exemplifies Lamarr’s Triangular AI paradigm, which integrates domain expertise in medicine with state-of-the-art AI and sensor systems to deliver impactful solutions for sleep analysis.
The conference also underscored the vital role of clinical evaluation in deploying AI tools for real-world impact. Reflecting on this, Buschjäger emphasized: “Indeed, just building the thing and showing that it works in principle is not enough. We have to test it in the field, with real people to make an impact”.
As the team looks to the future, their focus will shift to rigorous testing and further development to ensure that Sleepwalker not only supports clinicians in their daily work but also enhances patient care.
Lamarr continues to pioneer research at the intersection of AI and healthcare, demonstrating how AI in sleep medicine can address real-world challenges and deliver tangible societal benefits. By integrating advanced machine learning with clinical expertise, the institute is leading the way in transforming how sleep disorders are understood and treated.
