Researchers have developed a microscopic 3D-printed optical device that can efficiently combine light from dozens of small semiconductor lasers into a single multimode optical fiber with very low loss. The team demonstrated photonic lanterns that multiplex 7, 19, and 37 multimode VCSEL lasers directly into a fiber while preserving brightness and easing alignment constraints. By enabling scalable incoherent beam combining of many multimode lasers, the technology could simplify and improve high-power laser systems, optical communications, and other photonic applications where efficiently delivering large optical power through fibers is critical.

A research paper by scientists from Guangzhou University of Chinese Medicine described a framework that leverages bionic, wearable electrocardiogram (ECG) sensor technologies along with multimodal large language models using a coherent temporal modeling effort to address the intertwining of fine-grained temporal dependencies, heterogeneous biomedical modalities, and interpretable risk stratification.

The new research paper, published on Mar. 02 in the journal Cyborg and Bionic Systems, unveiled a first-of-its-kind intelligent cardiovascular monitoring framework that merges bionic wearable ECG technology with multimodal large language models, achieving unprecedented accuracy in early myocardial ischemia detection and post-reperfusion risk stratification.

A research paper by scientists from East China University of Science and Technology, University of Applied Sciences Campus Vienna, and other institutions proposed a domain generalization model (DGIFE) for electroencephalography (EEG) signals, featuring structured feature decoupling and fine-grained data augmentation to address the domain bias challenge in cross-subject brain-computer interface (BCI) applications.

The new research paper, published on Feb. 24 in the journal Cyborg and Bionic Systems, presented the development, validation, and optimization of the DGIFE model, demonstrating its superior generalization performance and noise robustness across multiple public datasets, providing an effective solution for practical BCI deployment.