In a new study, Northwestern University scientists developed a 90s-style video game to help chronic stroke survivors regain lost arm function. While wearing a small device on their impaired arm and using a laptop computer, players use their arm muscles to complete tasks such as flying a helicopter around the screen to hit a moving target. The muscle retraining helps separate the brain’s uncoordinated movement signals, enabling muscles to work independently again.

After six weeks of the game-based therapy, chronic stroke survivors improved arm function by as much as 7.8 times as much as those in the control group. 

As artificial intelligence systems grow larger and more powerful, their energy demands are rising dramatically. But recent research from the University of Massachusetts Amherst published Monday in Nature Communications suggests that advanced AI capabilities may be achievable with dramatically lower energy consumption.

Machine learning (ML) is reshaping pipeline integrity management (PIM) from physics-based to data-driven paradigms. This systematic review comprehensively surveys ML applications across pipeline reliability design, integrity assessment, inspection, and maintenance; identifies key bottlenecks including data scarcity, poor interpretability, and limited field validation; and proposes a research roadmap toward tractable, trustworthy ML for full-lifecycle pipeline safety.

A team at the Texas A&M Health Institute of Biosciences and Technology has developed an artificial intelligence-designed molecular switch that uses caffeine to rapidly separate engineered proteins inside living cells and trigger cellular responses on demand. The platform, called CODS, short for caffeine-operated dissociation system, could help scientists build safer and more controllable gene and cell therapies.

As health systems explore AI tools for vaccine communication, a new Penn-led randomized controlled trial offers a cautionary finding: Brief chatbot conversations can make vaccine-hesitant parents more willing to vaccinate their children against HPV, but they did not outperform standard public health materials. The effects of the written materials also persisted, while those of the chatbots faded. The researchers say the next step is to test AI as part of broader support systems that help families move from intention to action.