Neuroimaging analysis in brain disorders faces a persistent challenge: brain signals are complex and high-dimensional, while high-quality labeled datasets remain limited. This review article systematically examines how self-supervised learning can help address that gap by learning meaningful representations directly from unlabeled neuroimaging data. It covers major methodological families, including contrastive, generative, and hybrid generative-contrastive approaches, and discusses their applications in functional MRI, EEG, and multimodal brain network analysis.

The review argues that self-supervised learning offers more than annotation efficiency. It may enable more transferable and clinically useful representations for disease screening, diagnosis, and prognosis across heterogeneous datasets and disorders. At the same time, interpretability, data heterogeneity, missing modalities, and clinical validation remain major barriers. Future work will likely focus on stronger multimodal fusion, better cross-site generalization, and more clinically adaptable model design.

Dietary management in patients with end-stage renal disease (ESRD) has long lacked robust evidence directly linking nutrient intake to clinical outcomes, particularly in peritoneal dialysis populations. This study analyzed 12 years of longitudinal data from 656 patients, combining frequent follow-up assessments with detailed dietary records to evaluate associations between nutrient intake and mortality risk using both multivariable and nonlinear modeling approaches.

The findings demonstrate that most diet–outcome relationships are nonlinear, highlighting the importance of identifying optimal intake ranges rather than relying on single thresholds. In addition, the effects of nutrient intake vary across different nutritional states, as indicated by serum albumin levels, emphasizing the need for personalized dietary strategies. This study provides evidence-based reference ranges for nutritional management and introduces a methodological framework for future research on diet and long-term outcomes in chronic disease populations.

This study investigates the causal relationship between epilepsy subtypes and brain connectivity within resting-state networks using a bidirectional Mendelian randomization framework. By leveraging large-scale genetic and neuroimaging datasets, the researchers provide a more reliable assessment of causality beyond traditional observational studies.

The results demonstrate that genetic generalized epilepsy is associated with functional alterations in attention and motor-related brain networks, while brain connectivity changes do not appear to causally drive epilepsy. These findings deepen our understanding of epilepsy as a network-level disorder and suggest new avenues for targeted intervention and precision medicine.

A new review published in the journal Magnetic Medicine indicates that static magnetic fields (SMF) and alternating magnetic fields (AMF) are expected to provide a non-invasive new method for organ preservation, addressing many limitations of current organ preservation technologies and bringing new hope to the field of organ transplantation.

 

Researchers at Peking University has developed a "two-step" signaling switch—transitioning from Hippo-YAP/Notch to TGFβ1 pathways—to capture and stably maintain a novel stem cell state termed Trophectoderm-Like Stem Cells (TELSCs). These cells, derived from totipotent blastomere-like cells (TBLCs) or 8-cell embryos, precisely mirror the transcriptomic and epigenetic landscape of the E4.5 pre-implantation trophectoderm. Functionally, TELSCs exhibit extraordinary developmental plasticity; in chimeric assays, they contribute to all eight known trophoblast lineages at single-cell resolution. Furthermore, TELSCs efficiently assemble into 3D trophoblast organoids (TELSC-TOs) that recapitulate the coupled self-renewal and multi-lineage differentiation characteristic of the native placenta.

 

The carbonation of K₂CO₃ to KHCO₃ is an interesting CO₂ capture process due to its low material cost, high selectivity, and substantial CO₂ capacity. In this study, KHCO₃ particles were packed in a dielectric barrier discharge reactor and exposed to plasma. It was found that the decomposition of KHCO₃ is mainly driven by a thermal mechanism. The energy consumption is more than one order of magnitude higher compared to the thermal approach. However, production of CO and H₂ was achieved, with the best CO₂ conversion of 9.0% ± 0.2% and energy efficiency of 3.0% ± 0.1% at a syngas ratio of 0.35 ± 0.01.

In an era where electric vehicles (EVs) are accelerating toward mainstream adoption, the global push for sustainable transportation is undeniable. With fossil fuels dwindling and climate concerns mounting, EVs promise cleaner roads and reduced emissions. However, this surge in EV popularity is straining our existing power grids, especially at charging stations where unpredictable fleets of vehicles plug in and out randomly. This creates imbalances in power demand, leading to issues like voltage drops, harmonic distortions, and overall poor power quality that could hinder widespread EV integration. Enter the innovative solution explored in this research: using a device called D-STATCOM (Distribution Static Compensator) to dynamically balance loads and supply reactive power right at the charging station. By addressing these local challenges, the study paves the way for more reliable, efficient EV infrastructure, making electric mobility not just viable but truly attractive for everyday users.