Inactive H5N1 influenza virus in pasteurized milk poses minimal health risks
Peer-Reviewed Publication
Updates every hour. Last Updated: 3-Oct-2025 22:11 ET (4-Oct-2025 02:11 GMT/UTC)
H5N1 influenza virus continues to infect dairy cows in the United States, causing concern about potential spillover into humans. Though pasteurization effectively kills the virus, a significant portion of commercial milk still contains viral components, leading scientists to worry that regularly drinking these inactivated viral components could teach the immune system that these molecules were safe, resulting in greater susceptibility to later influenza infections. St. Jude researchers found that consuming pasteurized H5N1-infected milk had no impact on flu immunity in model systems.
In Vietnam, like many countries around the world, modern food retail is expanding, food is becoming more processed and packaged, and unhealthy food marketing is pervasive across physical and digital spaces. These shifts deliver convenience and (in some cases) safer handling, but they also make less healthy options easier to find and harder to avoid. A new study by the Alliance of Bioversity International and CIAT, along with their colleagues, maps these changes and the policy responses and proposes a more robust framework for a faster response.
A study from the Texas A&M University School of Public Health showed that smokers had higher rates of moderate to severe depressive symptoms and sleep disorders than non-smokers. Notably, smokers aged 40 to 59 who did not meet physical activity guidelines were especially likely to report moderate to severe depressive symptoms and sleep disorders.
Researchers at the University of Michigan are using artificial intelligence to predict the health consequences that sport-related concussions might have on student athletes over the course of their college athletic careers.
MIT researchers developed an interactive, AI-based system that enables users to rapidly annotate areas of interest in new biomedical imaging datasets, without training a machine-learning model in advance. As the user uploads new images, the number of interactions needed to accurately segment the image drops, eventually to zero, enabling rapid annotation of the entire dataset.