Add BrightSurf on Google Email

Science News Archive December 2025 - Page 32


Page 32 of 46

Slow turning could be an indicator for Parkinson's disease

A decade-long study of 1,051 participants found that slower peak angular velocity was linked to a higher risk of developing Parkinson's disease, with estimated turning speeds declining 8.8 years before clinical diagnosis. The study used wearable sensors and machine learning models to predict Parkinson's cases with high accuracy.

SourceMurdoch University·JournalAnnals of Neurology·TypeObservational study·DateDec 8, 2025

Tuberculosis: Scientists develop novel drug candidate for combating resistant pathogens

Researchers at Martin Luther University Halle-Wittenberg have developed a promising new substance that inhibits the ability of tuberculosis bacteria to produce energy and causes them to die. The compound, PRP020, targets the pathogen's ATP synthase but attacks a different site than existing drugs like bedaquiline.

SourceMartin-Luther-Universität Halle-Wittenberg·JournalJournal of Medicinal Chemistry·TypeExperimental study·DateDec 8, 2025

Open-source sub-nanometer stabilization system for super-resolution fluorescence microscopy

A new open-source stabilization system enables sub-nanometer precision in super-resolution fluorescence microscopy, overcoming technical complexity. The system achieves high stability, allowing for long-duration single-molecule localization experiments with controlled drift-induced errors.

Foshan University pioneers nanoparticle method to curb cadmium in rice crops

A research team at Foshan University has developed a method to reduce cadmium accumulation in rice crops using foliar-applied molybdenum nanoparticles. The study shows that MoNPs down-regulate key genes involved in cadmium uptake and cell wall modification, reducing Cd accumulation in root tissues and limiting its entry into grains.

SourceKeAi Communications Co., Ltd.·JournalThe Crop Journal·TypeExperimental study·DateDec 8, 2025

New artificial intelligence tool shows promise in accurately diagnosing brain tumors without surgery

Researchers developed an automated machine learning model to differentiate between pituitary macroadenomas and parasellar meningiomas using preoperative MRI scans. The model achieved over 97% accuracy in distinguishing the two brain tumors, improving surgical planning and potential patient outcomes.

Positive and poolished: Student writing has evolved in the AI era

A University of Warwick-led analysis of 4,820 student reports found that student writing has become more sophisticated, formal, and positive since the introduction of ChatGPT in 2022. However, grades have remained stable despite stylistic shifts, suggesting core academic skills remain essential.

SourceUniversity of Warwick·JournalComputers and Education Artificial Intelligence·TypeData/statistical analysis·DateDec 8, 2025

Novel AI technique able to distinguish between progressive brain tumours and radiation necrosis, York University study finds

A new AI-based method can differentiate between tumour progression and radiation necrosis with high accuracy, enabling more accurate identification and treatment of late-stage cancer patients. This discovery could lead to better treatments for brain metastasis patients treated with stereotactic radiosurgery.

SourceYork University·JournalInternational Journal of Radiation Oncology*Biology*Physics·DateDec 8, 2025

AI supports home-based balance training

A new machine learning model developed at the University of Michigan uses data from wearable sensors to predict how physical therapists would assess balance training performance. The model was trained with sensor data combined with analysis from physical therapists and predicted patients' balance ratings with nearly 90% accuracy.

SourceUniversity of Michigan·JournalJournal of NeuroEngineering and Rehabilitation·DateDec 8, 2025

E2CNN: Entity-type-enriched cascaded neural network for Chinese financial relation extraction

The E2CNN model resolves overlapping relations by employing a cascaded neural network and incorporates entity type information. It achieves this by first identifying entities and their types, then predicting subject positions and extracting objects corresponding to the subjects under predefined relations.

SourceHigher Education Press·JournalFrontiers of Computer Science·TypeExperimental study·DateDec 8, 2025

Cutting-edge 3D AI technology transforming dental care for children

A new study reveals that a hand-held, wireless intraoral scanner paired with AI-assisted software can identify early childhood tooth decay as accurately as visual exams. The technology creates a detailed 3D model of a child's mouth within minutes, providing a comprehensive picture of their dental health.

SourceMurdoch Childrens Research Institute·JournalJMIR Public Health and Surveillance·TypeExperimental study·DateDec 8, 2025

Mount Sinai study finds childhood leukemia aggressiveness depends on timing of genetic mutation

A Mount Sinai study found that the timing of a genetic mutation in children with leukemia can significantly impact its aggressiveness. The researchers discovered that leukemia caused by mutations occurring before birth is often more aggressive and harder to treat than those occurring later in life.

SourceThe Mount Sinai Hospital / Mount Sinai School of Medicine·JournalCancer Discovery·TypeExperimental study·DateDec 8, 2025

An approach for learning NeRFs from uncalibrated few-view images by CAD model retrieval

The researchers proposed CAD-NeRF, a method that reconstructs 3D scenes from fewer than 10 images without known camera poses. It uses a mini library of CAD models to retrieve a similar shape and guide density supervision and pose initialization. Comprehensive evaluations demonstrate strong generalization capabilities.

SourceHigher Education Press·JournalFrontiers of Computer Science·TypeExperimental study·DateDec 8, 2025

Study measures behavioral markers of alliance in support groups, viability of machine learning in mental health

Researchers measured dyadic alliance in support groups using computational algorithms and found verbal and nonverbal communication features, such as head poses and facial expressions, to be strong predictors of connection. The study suggests that AI can potentially identify behavioral markers of alliance and may have a role in mental h...

SourceUniversity of Kansas·TypeData/statistical analysis·DateDec 8, 2025

NEJM Group launches NEJM Clinician

NEJM Clinician curates highly relevant clinical insights from over 150 peer-reviewed journals to give physicians clear guidance without the noise. The weekly publication delivers summaries and commentary in a concise, 30-minute read, helping healthcare providers stay informed and confident in their decisions.

SourceNEJM Group·JournalNew England Journal of Medicine·DateDec 8, 2025

ASH 2025: Safety and efficacy data from a phase 1 study show promise of next-generation CAR-T cell therapy in patients with refractory B-cell acute lymphoblastic leukemia

Researchers present phase 1 study results of rapcabtagene autoleucel, a rapid CAR-T cell manufacturing platform that reduced production time to under two days. The treatment showed a manageable safety profile and promising antitumor activity, with complete remission rates ranging from 70% to 100% in patients.

Enabling a single pixel to "see' complex environments

Researchers developed a physically consistent model for single-pixel imaging, capturing multiple degradations and improving image resolution and fidelity. The method outperforms competing approaches under real-world complex degradations, offering robustness and superiority in practical environments.