Machine-learning immune-system analysis study may hold clues to personalized medicine
Peer-Reviewed Publication
Updates every hour. Last Updated: 8-Jun-2026 11:16 ET (8-Jun-2026 15:16 GMT/UTC)
Fried foods are popular with consumers, but their high fat content can contribute to health challenges like obesity and hypertension. If the food industry can offer lower-fat options of similar quality, people can more easily make health-conscious food choices. Researchers at the University of Illinois Urbana-Champaign have explored microwave frying of French fries, providing insights that can help food manufacturers modify their production methods.
When a cocaine addict relapses, it isn’t a matter of personal failure — it’s the biological result of their brain’s rewiring, new research finds. Michigan State University scientists found that cocaine changes how the hippocampus functions, contributing to the ongoing compulsion to seek out the drug. Their National Institutes of Health-supported research, published in Science Advances, not only explains why cocaine addiction is notoriously difficult to treat, but it could also help scientists develop new pharmaceutical therapies.
AI algorithms are increasingly developed to monitor vector populations based on either photos or sounds. However, the real-life accuracy of the models is highly dependent on the training data. Unfortunately, training data is still limited, both in matter of the species that are actually included, and the natural variation represented within a given species. In this study, researchers from HUN-REN Centre for Ecological Research, ELTE University, Budapest and University of Szeged show that mosquito sound is fairly consistent among species, especially if environmental parameters such as temperature are taken into account. This means that, to develop field-accurate AI identification models based on mosquito sounds, we either need to properly represent the natural variations caused by environmental and biological factors into the training data (especially temperature), or adjust the identification algorithm based on those factors. These types of models could help us monitor vector populations and implement efficient public health strategies.
For the first time, research shows that a key social trait in honey bees is linked to measurable physiological advantages that can improve colony survival.
A new study published in PLOS ONE uncovers how hygienic honey bee colonies mount stronger individual immune defenses against Nosema ceranae, a widespread pathogen that threatens honey bee health worldwide.