Browse Articles

High-intensity interval training differentially modulates acute BDNF and cognitive responses in young adult males: a randomized crossover trial

Abstract Exercise is widely recognized for its beneficial effects on brain health, yet the extent to which exercise intensity modulates acute neurochemical and cognitive responses remains unclear. Particularly, the role of exercise intensity in shaping brain-derived neurotrophic factor (BDNF), lactate responses, and executive function requires further investigation. This study compared the acute effects of low-intensity continuous training (LICT), moderate-intensity continuous training (MICT), high-intensity interval training (HIIT), and a resting control condition (CTRL) on BDNF levels, blood lactate concentration, and cognitive responses in healthy young adult males. Twelve healthy young adult males completed LICT, MICT, HIIT, and the control condition using a randomized crossover design with a 7-day washout period. Serum BDNF, blood lactate concentration, and executive function assessed by the Stroop Test were measured before and immediately after each experimental condition. HIIT induced significantly greater post-exercise increases in BDNF and lactate compared with all other conditions, while MICT elicited moderate elevations relative to LICT and rest. Lactate responses increased progressively with exercise intensity. Improvements in executive function were observed exclusively following HIIT, reflected by significantly faster Stroop Test completion times. HIIT produced concurrent elevations in lactate and serum BDNF together with improved executive function performance. HIIT may represent an effective acute stimulus for cognitive benefits, with potential relevance for exercise approaches aimed at supporting brain health via neurotrophic signaling. Trial registration: The study was retrospectively registered on ClinicalTrials.gov (identifier: NCT07137611; https://clinicaltrials.gov/study/NCT07137611 ) on 22 August 2025.

High-throughput Raman-activated cell sorting of microalgal genome-wide edited library revealed a regulatory pathway for carotenoid synthesis

Assessment of gabapentin efficacy in patients with KCNQ2-developmental epileptic encephalopathy

Compositionality of social gaze in the prefrontal-amygdala circuits

Joint association of triglyceride-glucose index and atherogenic lipid markers with incident stroke risk

Abstract The triglyceride-glucose (TyG) index and atherogenic cholesterol markers, including non-HDL cholesterol and remnant cholesterol, are significant predictors of atherosclerotic cardiovascular disease (ASCVD). However, the joint effects and predictive value of TyG and atherogenic cholesterol markers for incident stroke remain insufficiently understood. We included participants from the China Health and Retirement Longitudinal Study (CHARLS) enrolled at baseline in 2011 and followed them through 2020. Participants were categorized into four groups according to the median values of the TyG index and each cholesterol marker, with those having both values below the median serving as the reference group. Cox proportional hazards models were used to evaluate the independent and joint associations of TyG and atherogenic cholesterol markers with incident stroke. Restricted cubic spline models were applied to assess dose–response relationships, and receiver operating characteristic (ROC) curve analyses were used to examine predictive performance. A total of 8,544 participants were included (mean [SD] age 59.0 [9.5] years; 52.1% female). 522 incident stroke events occurred during a maximum follow-up of 9.0 years. TyG, non-HDL cholesterol and remnant cholesterol were all independently associated with stroke risk. Compared with participants with both TyG and cholesterol markers below median, those with both markers above median had the greatest stroke risk in fully adjusted models (HR for high TyG and high non-HDL cholesterol: 1.76 [95% CI 1.41–2.20]; HR for high TyG and high remnant cholesterol: 1.45 [95% CI 1.18–1.77]). Adding TyG and each cholesterol marker to traditional risk factors modestly improved risk discrimination, with the model combining TyG and non-HDL cholesterol yielding the highest AUC. The TyG index and atherogenic cholesterol markers were independently and jointly associated with increased stroke risk among middle-aged and older adults. Adding these routinely available metabolic and lipid markers modestly improved discrimination for incident stroke.

Dual-phase eutectic ceramics with improved hardness and toughness via nano-coherent high-entropy oxides

Performance of GPT-based large language models in hepatocellular carcinoma stratification: liver function assessment, BCLC staging, and treatment recommendations

Abstract Large language models (LLMs) like GPT have been proposed to support complex clinical decision-making. This study evaluated the performance of GPT-based LLM in analyzing clinical, radiological, and laboratory data from patients with hepatocellular carcinoma (HCC) to assess liver function, assign BCLC stage, and recommend treatment. Data from 106 HCC patients (82% male, median age 65 [22–86]) were compiled into anonymized integrated reports. Four GPT-versions (4, o1, o3, 5.4) were prompted—using both short and long instructions—to calculate MELD, ALBI, and Child–Pugh scores, assign BCLC stage, and generate treatment recommendations based on current guidelines. Outputs were compared to expert consensus and tumor board decisions. Errors were categorized by type and source. Time and cost analyses compared GPT to clinical staff. All GPT versions achieved high accuracy (> 85%) in liver function assessment, with MELD calculation being the most error-prone. BCLC staging accuracy ranged from 46.2% (version 4) to 84.0% (o3), with misclassification of radiological reports as the main error source. Reasoning-optimized models (o1, o3) performed best for treatment recommendations, achieving an overall accuracy (correct suggestions and acceptable alternatives) of up to 90.6%. In 9–14% of cases, GPT suggestions were retrospectively more guideline-concordant than tumor board decisions. GPT processing was significantly faster and reduced costs by approximately 300- to 1300-fold compared to clinical staff. GPT-based LLMs show potential as decision-support tools for liver function assessment, BCLC staging, and treatment guidance in HCC. Particularly with reasoning-optimized models and detailed prompting, LLMs may serve as valuable adjuncts in multidisciplinary HCC workflows. However, a non-negligible error rate requires expert oversight and further model refinement.

On-chip quadratically nonlinear photodetector

On the value of radiomics in addition to clinical measures in emotional conflict fMRI for predicting sertraline response in major depressive disorder

Unraveling the mechanistic origins of efficiency and stereoselectivity in EDA-mediated photocatalysis

A multi-task learning framework for diagnosing partial discharge types and assessing severity

Hot exciton dissociation in graphene nanoribbons

Abstract Exciton dissociation in semiconducting nanostructures is crucial for optoelectronic applications, especially when free-carrier generation is required. Despite considerable research, the question of whether and how such generation occurs in strongly excitonic systems remains elusive. Here, we use one-dimensional precision graphene nanoribbons (GNRs) as a model system to investigate exciton dissociation. We systematically explore the interplay between ribbon length ( l ), excitation energy, and band dispersion in various precision GNRs. Ultrafast Terahertz conductivity measurements reveal that hot exciton dissociation dominates carrier generation, with ribbon length significantly influencing free carrier lifetimes. We identify a critical Bjerrum length ( R B ) of approximately 20 nm that determines whether photoexcited hot carriers in GNRs can dissociate before forming tightly bound excitons. For shorter ribbons ( l  < 2 R B ), rapid ~ps exciton formation prevails. Furthermore, the charge-carrier band dispersion in GNRs plays a critical role in determining dissociation efficiency. Long GNRs with strongly dispersed bands, and consequently low effective carrier masses, exhibit higher mobilities that promote efficient hot-exciton dissociation. These results advance fundamental understanding of dimensionality, energetics, and electronic structure in excitonic materials, providing design principles for optoelectronic devices based on excitonic materials.

HOA-OBL: hybrid opposition-based hippopotamus optimization framework for efficient UAV task allocation

Integrated theranostic nanoplatform empowers precision cancer care via radionuclide-labeled NIR-II aggregation-induced emission luminogens

Hybrid RSM–SF-AHP–fuzzy MARCOS approach for multi-response optimization of WAAM-fabricated Ni–SS bimetallics in face milling

Influence of protein aggregates, extracellular vesicles, and lipoprotein fusion on ionizable lipid nanoparticles protein corona analysis

Critical re-evaluation of experimental settings with fluorescent probes for bioorthogonal chemistry with palladium antimony inactivation-resistant catalyst

Ultra-rapid cryo-EM data acquisition method enabled by continuous recording based beam image shift

Abstract Cryo-electron microscopy (cryo-EM) data acquisition is time-intensive given that a large amount of data is required to obtain a high-resolution reconstruction. Here, we eliminate camera-induced delay time by continuously recording during beam-image shift acquisition using a method called Continuous Recording Beam-Image Shift (CR-BIS). The utilization of CR-BIS with K3 and Falcon 4 direct electron detectors and conventional data acquisition conditions enables the acquisition of ~34,000 micrographs and ~1,000 tilt series per 24 h in single-particle analysis mode and cryo-electron tomography mode, respectively. Three-dimensional reconstructions of single-particle and tomographic datasets show that CR-BIS accelerates data collection and maintains data quality. CR-BIS is broadly applicable for efficient high-resolution cryo-EM since it can be implemented into existing acquisition software through scripting and it does not require hardware modification.

Clinical and genomic surveillance of rotavirus causing acute gastroenteritis in the community of Madrid, Spain, 2021–2023: predominance of equine-like G3 strains

Higher visual areas act like domain-general filters with strong selectivity and functional specialization

Abstract Neuroscientific studies rely heavily on a-priori hypotheses, which can bias results toward existing theories. Here, we use a hypothesis-neutral approach to study category selectivity in higher visual cortex. Using only stimulus images and their associated fMRI activity, we constrain randomly initialized neural networks to predict voxel activity. Despite no category-level supervision, units in the trained networks act as detectors for semantic concepts like ‘faces’ or ‘words’, providing solid empirical support for categorical selectivity. Importantly, this selectivity is mostly maintained when training the networks without images that contain the preferred category, strongly suggesting that selectivity is not domain-specific machinery, but sensitivity to generic patterns that characterize preferred categories. The ability of the models’ representations to transfer to perceptual tasks further reveals the functional role of their selective responses. Finally, our models show selectivity only for a limited number of categories, all previously identified, suggesting that the essential categories are already known.