Search Results (589)

Background/Objectives: Repetitive transcranial magnetic stimulation (rTMS) is a non-invasive neuromodulatory approach with potential therapeutic value for cognitive impairment in Alzheimer’s disease (AD). Ferroptosis-related oxidative injury has been implicated in AD-associated neuronal dysfunction, but whether rTMS-induced functional improvement is accompanied by changes in ferroptosis-related oxidative injury remains unclear. This study evaluated the effects of high-frequency rTMS on cognitive function, hippocampal neuronal excitability, and ferroptosis-related oxidative injury in amyloid precursor protein/presenilin-1 (APP/PS1) mice, using Ferrostatin-1 (Fer-1) as a pharmacological comparator. Methods: Six-month-old female mice were used, including age-matched C57BL/6J controls and APP/PS1 mice assigned to the AD + Sham, AD + rTMS, and AD + Fer-1 groups (n = 6 per group). After 14 days of intervention, cognitive performance was assessed using behavioral tests. Whole-cell patch-clamp recordings were performed in hippocampal dentate gyrus granule neurons to evaluate neuronal excitability and voltage-gated sodium (Na⁺) and potassium (K⁺) channel properties. Biochemical assays and transmission electron microscopy were used to assess oxidative, iron-related, and mitochondrial changes, and mitochondrial ultrastructure was examined in an independent cohort (n = 3 per group) using transmission electron microscopy. Results: Compared with AD + Sham mice, high-frequency rTMS improved cognitive performance, increased evoked action potential firing, lowered the elevated action potential threshold, partially restored voltage-gated Na⁺ and K⁺ current amplitudes, and accelerated recovery of Na⁺ currents from inactivation. Fer-1 produced partially overlapping, but not identical, effects across behavioral, electrophysiological, biochemical, and ultrastructural outcomes. Both interventions increased hippocampal glutathione (GSH) levels, reduced malondialdehyde (MDA) and total iron levels, partially restored superoxide dismutase (SOD) activity, and improved mitochondrial ultrastructure and reduced the prevalence of mitochondrial profiles with small cross-sectional areas. Conclusions: High-frequency rTMS improved cognitive and hippocampal neuronal outcomes in female APP/PS1 mice. These improvements were accompanied by biochemical and mitochondrial changes compatible with attenuation of ferroptosis-related injury. However, the findings do not establish ferroptosis inhibition as either necessary or sufficient for the effects of rTMS. Full article

Biomimetic perception systems integrate heterogeneous cues selectively rather than treating all available information as equally reliable. Inspired by this principle, this study proposes QGFace-LLaVA, a multimodal large language model (MLLM)-centered framework for robust face analysis under imperfect metadata. A pretrained MLLM serves as the shared prompt-conditioned visual–language reasoning backbone, while structured side information, including age, gender, confidence cues, and availability indicators, is regulated through task-aware quality estimation, reliability-guided metadata calibration, gated residual correction, and counterfactual metadata reliability regularization (CMRR). Experiments on FER2013, CelebA-40, and UTKFace cover facial expression recognition, facial attribute recognition, and age estimation under clean, noisy, missing, shuffled, naturally erroneous, and counterfactual metadata conditions. The results show that metadata utility depends jointly on task relevance, metadata reliability, and fusion strategy, and that clean-setting gains do not necessarily imply robustness. QGFace-LLaVA reduces harmful dependence on unreliable metadata, while CMRR provides additional stability under corruption and mismatch. Overall, the framework transfers biomimetic selective cue integration into MLLM-based face analysis by treating metadata as reliability-controlled auxiliary evidence rather than a uniformly beneficial input. Full article

This study evaluates a fixed Kronecker-sequence interleaver under controlled synthetic three-dimensional correlated-fault models. Spatial-cluster, column-correlated, and bit-plane-dependent probability fields are used as coding-level abstractions and are not calibrated device measurements. The K-IPA mapping is compared with random, structured 3D block, modular-stride, and length-adapted quadratic-permutation-polynomial (QPP-style) mappings using BCH(63,45) and RS(63,45) backends. At p = 0.015 and ρ = 0.85, K-IPA BCH has lower FER than random, 3D block, and QPP-style BCH, but its difference from stride BCH is small, and the paired confidence interval includes zero. Within the RS backend, the paired comparisons among K-IPA, stride, and QPP-style mappings do not resolve a difference at this operating point. Because the BCH and RS tensor partitions contain different numbers and types of decoder units, their FER values are not used to rank the two code families. The topology study further shows that no fixed mapping is uniformly best. A separate address-remapping implementation check verifies the fixed lookup table only; device-calibrated fault validation and complete codec hardware evaluation are outside the evidence provided here. Full article

Facial expression recognition in the wild is challenged by both noisy supervision and degraded visual evidence: subtle expression cues must be interpreted under blur, contrast changes, image noise, and annotator disagreement. Existing noisy-label FER methods mainly regulate samples, labels, or attention, while frequency information is rarely used to adapt the semantic representation itself. We propose Frequency-Guided Expert Modulation (FARM-FER), which treats local and global frequency descriptors as a control signal rather than an additional classifier input. A joint Haar-DWT and radial-FFT context guides soft routing among nonlinear experts and channel-wise affine recalibration of the semantic feature, while a learned gate combines the two corrections before a lightweight classifier predicts the expression from the refined representation. Across RAF-DB, FER+, and AffectNet under symmetric label noise, with additional evaluations under class-dependent label noise on RAF-DB and native crowd-label ambiguity on FER+, FARM-FER consistently improves matched baselines. At 30% symmetric noise, FARM-FER reaches 89.18% accuracy on RAF-DB, with a 1.6% performance gain over the matched Swin-Tiny baseline; the gains also hold in a controlled ResNet18 reimplementation and in class-sensitive AffectNet evaluation. Measured cost analyses show only modest parameter and FLOP overhead, supporting a lightweight yet effective design in terms of model size and arithmetic cost for noisy-label FER. Full article

Fusarium ear rot (FER), caused by Fusarium verticillioides (F. verticillioides), is a major disease of maize that reduces grain yield and quality globally. However, few major loci for FER have been verified and cloned. Resistance to FER is a quantitative trait influenced by environmental conditions, and maize genotypes completely resistant to the pathogen remain unknown. To gain a comprehensive understanding of the genetic basis of natural variation in FER resistance, a recombinant inbred line (RIL) population consisting of 257 progenies was developed by crossing the resistant line BT with the susceptible line Xi502. This population was genotyped using a set of 6807 high-density single nucleotide polymorphism (SNP) markers developed in this study. As a result, a total of five QTLs were identified by linkage mapping across three years, located on five chromosomes, and explaining 4.38–13.13% of the phenotypic variation. Among these was a major QTL, qFER5-2. Located on chromosome 5 within the interval 185568562–185574073, qFER5-2 explained 13.13% of the total phenotypic variance. The two candidate genes within qFER5-2 exhibited distinct expression profiles between the BT and Xi502 inbred lines, suggesting their potential association with FER resistance. Collectively, these findings provide candidate genetic resources for further investigation and offer potentially useful materials for maize disease resistance breeding. Full article

Recognizing facial emotions automatically from images/videos (FER) still represents a difficult problem for emotion computing, mainly due to variations in the face pose, lighting, occlusion, facial features, and expression intensity in the wild. Recent CNN–Transformer-based hybrid models like POSTER have leveraged local feature learning, landmark guidance, and global dependency modeling to achieve strong performance. Yet these methods give the main focus to spatial and contextual representations while not really going deep into adaptive channel-wise feature importance over multi-scale representations. As different feature channels represent emotions in varying degrees, it is likely that by treating all feature channels equally, one would limit the ability of the learned features to discriminate effectively. To overcome this weakness, this article presents a ResNet-18–Transformer landmark-guided module called SE-POSTER that fuses lightweight Squeeze-and-Excitation (SE) attention modules into the multi-scale feature pyramid of the baseline POSTER architecture. The proposed method carries out feature channel recalibration adaptively at the level of features before Transformer-based global attention modeling, thus allowing the network to focus on emotionally informative feature channels and suppress less relevant responses. The inclusion of SE attention in the network enhances fine, mid, and global levels of feature representations at a very low cost in terms of computation. On the basis of the RAF-DB, FERPlus, and AffectNet datasets, enormous experiments prove that the SE-POSTER framework proposed is capable of steadily boosting recognition accuracy relative to the baseline POSTER and several state-of-the-art FER methods. Especially, the proposed model delivers 92.78% accuracy on RAF-DB while it also shows better robustness and generalization capability under difficult real-world conditions. Moreover, additional ablation studies reveal that multi-level channel recalibration is effective in improving discriminative emotional feature learning. Full article

Gardenia fruit, a medicinal and edible resource rich in bioactive compounds, faces several limitations, including low extraction efficiency, poor stability, and intense bitterness. In this study, gardenia fruit powder at varying concentrations (2.5%, 5.0%, 7.5%, 10.0%, and 12.5%) was mixed with a kombucha inoculum and fermented for 8 days to develop functional beverages. The results demonstrated that fermentation with gardenia fruit powder effectively supported SCOBY growth. The 5.0% addition of gardenia fruit (G5Fer) produced thicker bacterial cellulose, higher total titratable acidity (10.43 g/L), and increased total polyphenols (37.12 mg GAE/100 mL) after 8 days of fermentation. G5Fer also achieved the highest sensory scores, especially in aroma, flavor, and overall acceptance. Higher doses of gardenia fruit (≥7.5%) caused turbidity, dark color, and strong sourness, leading to lower sensory scores. Non-targeted metabolomics revealed that kombucha fermentation reshaped the metabolic profile, enriching phenolic acids (e.g., gallic acid and caffeic acid) and flavonol glycosides (e.g., kaempferol-3-O-rutinoside) via the biosynthesis of phenylpropanoid pathways while antioxidant activity (DPPH and ABTS⁺ scavenging rate) was well maintained. Overall, G5Fer achieved the best balance among acidity, bioactive compounds, and consumer preference. In conclusion, gardenia fruit represents a promising substrate for developing a functional, palatable kombucha-like beverage, aligning with the growing demand for value-added fermented products. Full article

Salt stress limits tomato productivity, yet how translational regulation contributes to root salt adaptation remains poorly understood. We integrated RNA-seq and ribosome profiling in wild-type (WT) and FERONIA (FER) mutant (fer) tomato roots under control and 150 mM NaCl conditions. In WT roots, the salt response was predominantly transcript-driven, but a 29-gene ribosome-associated module showed reduced RNA abundance alongside increased translational efficiency, indicating selective translational buffering. FER loss-of-function disrupted this balance, constitutively elevating ribosome occupancy of ribosome-associated genes while reducing basal expression of stress- and ion-transport-related genes; under salt treatment, fer also showed stronger ion-transport transcriptional responses but weaker translational efficiency responses of this module. WT salt stress further shifted ribosome allocation from the 5′ untranslated region (UTR) toward the coding sequence (CDS), an effect attenuated in fer, alongside positive coupling between uORF and CDS translational efficiency. Feature modeling identified sequence and structural predictors of uORF translation, including weaker local RNA folding near the start codon and specific amino acid and stop codon preferences. Together, these results reveal FER-associated changes in ribosome-associated translational buffering during tomato root salt responses. Full article

Facial expression recognition (FER) plays an important role in human–computer interaction and affective computing. Although combining multiple FER datasets in joint training can potentially improve model generalization, it is hindered by inconsistent emotion labeling criteria across datasets. To address this issue, we propose a Prototype-Guided Multimodal Alignment Joint Training framework for multi-dataset FER. The core idea is to leverage the image–text alignment knowledge learned by Vision–Language Models from the target dataset as a unified emotion labeling criterion across datasets, while deriving prototypes from visual features to serve as emotion anchors for reliable feature alignment in the latent space. Based on the consistency among semantic predictions, prototype-distance predictions, and auxiliary labels, semantically consistent auxiliary samples are selected for joint training under a unified labeling criterion. Extensive experiments show that the proposed framework, with the Amending Representation Module as the backbone network, achieves 93.74% accuracy on the RAF-DB dataset and 98.31% on the CAER-S dataset, attaining state-of-the-art performance. The experimental results demonstrate that establishing semantically consistent labeling criteria across datasets is an effective strategy for multi-dataset FER learning. Full article

Emotion recognition remains a challenging task despite substantial progress in machine learning and affective computing. This study examines challenges in emotion recognition through a comparative analysis of two widely used modalities: facial images and speech signals. The analysis was conducted using FER-2013 for facial emotion recognition and the TESS and RAVDESS datasets for speech emotion recognition. A MobileNetV2-based approach was applied to visual data, while speech analysis employed MFCC-based representations and both classical and deep learning models. The study combines quantitative performance evaluation with qualitative analysis of classification behavior, focusing on emotion-specific recognition difficulties and recurring error patterns across modalities. Model performance was assessed using accuracy, precision, recall, F1-score, and confusion matrices. Across the analysed datasets, overall classification accuracy ranged from approximately 73% to 96%, while class-level F1-scores ranged from 0.48 to 0.89 depending on the emotion and modality. Happiness and surprise consistently achieved the highest recognition performance, whereas neutral emotion, fear, and disgust exhibited the lowest class-level F1-scores and generated the highest numbers of misclassifications. The experimental results confirmed that happiness and surprise achieved the highest classification performance across modalities, while neutral emotion, fear, and disgust showed reduced recognition accuracy due to weak expressive cues and overlapping feature representations. These difficulties are associated with weak or ambiguous expressive signals, overlap between emotional categories, and variability in emotional expression. The comparative findings suggest that recognition challenges arise from both modality-specific limitations and the inherent properties of emotional expression. The results highlight the importance of multimodal approaches and more flexible representations for improving emotion recognition systems. Full article

Diabetes mellitus (DM) is increasing rapidly worldwide, and diabetic cardiomyopathy (DCM) has become a leading cause of death in diabetic patients. Therefore, effective strategies for the prevention and treatment of DCM are urgently needed. Ferroptosis, a form of regulated cell death, has been implicated in the pathogenesis of DCM. This study integrated network pharmacology, data mining, molecular docking, molecular dynamics simulations, and in vitro experiments to investigate whether baicalein inhibits high glucose and sodium palmitate (HG + PA)-induced ferroptosis in cardiomyocytes and to elucidate the underlying mechanisms. Baicalein significantly improved the viability of H9c2 and AC16 cells, reduced cell death, and decreased LDH release under HG + PA conditions. Network pharmacology predicted that the protective effects of baicalein against HG + PA-induced cardiomyocyte injury were associated with ferroptosis regulation. Transcriptomic data mining further identified ferroptosis-related pathway enrichment in complementary in vitro and diabetic rat cardiac datasets. Molecular docking predicted favorable binding poses of baicalein with Nrf2, SLC7A11, and GPX4, while molecular dynamics simulations suggested general stability of the modeled complexes. In vitro experiments further confirmed ferroptosis involvement, as the ferroptosis inhibitor Ferrostatin-1 (Fer-1) reversed the HG + PA-induced decline in cell viability. Conversely, the ferroptosis inducer Erastin diminished cell survival and antagonized the protection conferred by baicalein, indicating that baicalein acts by inhibiting ferroptosis. Baicalein reduced lipid peroxidation, MDA and Fe²⁺ levels, and the mRNA expression of ACSL4 and PTGS2, while restoring the GSH/GSSG ratio and the protein expression of Nrf2, SLC7A11, and GPX4. These protective effects were partially reversed by the Nrf2 inhibitor ML385. In conclusion, baicalein protects cardiomyocytes from HG + PA-induced injury by activating the Nrf2/SLC7A11/GPX4 signaling pathway and inhibiting ferroptosis. Full article

Reliable facial expression recognition (FER) is essential for human–computer interaction across scenarios. However, existing methods often overlook spatial structure when seeking region-specific features due to fixed or limited local regions, limiting representation capability under real-world variations. Inspired by the multi-location attention of the human visual system, we propose a Landmark-guided Diverse Regional Representation Network (LDR-Net), using dynamic landmarks as patch centers to preserve structural details while locating expression-critical areas, facilitating effective regional representations for FER. First, a novel Diverse Regional Feature Extraction (DRFE) module operates via complementary operations: landmark-guided cropping for local details and cross-level integration with feature reorganization for holistic aggregation. Second, a novel Diverse Representation Learning (DRL) module is proposed with a collaborative dual-stream mechanism that captures fine-grained local dependencies via Transformers while reinforcing global features through attention-based enhancement, enabling comprehensive feature learning. Finally, a new Hybrid Feature Fusion (HFF) module is proposed for joint decision optimization via a hierarchical hybrid strategy, which aggregates intra-branch predictions followed by weighted branch-level fusion. Experiments on three FER benchmarks (RAF-DB, AffectNet, SFEW) and five occlusion/pose test sets demonstrate that LDR-Net outperforms state-of-the-art methods, while cross-scene validation on KMU-FED confirms its effectiveness in real-world driving. Full article

This paper presents a preliminary technical feasibility study of a landmark-based motion analysis system designed for potential future application in home-based rehabilitation monitoring for frozen shoulder (adhesive capsulitis), developed using computer vision (CV) and human–computer interaction (HCI) principles. The proposed system utilizes real-time body landmark detection to quantify shoulder joint kinematics and provide rule-based automated feedback on exercise execution. The system combines automated and manual components: while shoulder angle assessment, cosine similarity analysis, and keyframe matching are automated, manual researcher input is required to define keyframes corresponding to movement states (start, midpoint, peak) for each therapeutic pose. The CV-driven perception is translated into HCI output, including quantitative movement scores and rule-based feedback indicators, demonstrating the technical potential for objective evaluation of rehabilitation exercise execution without specialized wearable sensors. Technical validation was conducted with 14 healthy volunteers (not frozen shoulder patients) executing standardized shoulder rehabilitation activities, demonstrating shoulder angle measurement with an overall mean absolute error (MAE) of 7.03° against general goniometry and 6.61° against clinical goniometry (RMSE: 8.50° and 8.79°, respectively). Movement similarity classification achieved F1-scores ranging from 0.870 (flexion) to 1.0 (internal rotation) when compared against expert evaluation, though these results are based on a controlled and largely imbalanced dataset with limited incorrect movement examples. The system additionally incorporates a facial expression recognition (FER) module, previously developed and validated in the authors’ prior work, as a supplementary component to support future integration of pain monitoring; this module was not independently validated in the present study. This preliminary technical feasibility study contributes to rehabilitation engineering by demonstrating the potential of semi-automated CV-based motion analysis and rule-based HCI feedback for shoulder movement assessment. The findings indicate technical feasibility for future investigation in home-based exercise monitoring; however, clinical utility cannot be claimed at this stage, as validation with actual frozen shoulder patient cohorts is required. Full article

Reperfusion after acute ischemic stroke (AIS) triggers a series of ferroptosis-related molecular events, including iron dyshomeostasis, oxidative/nitrative stress, antioxidant depletion, and membrane lipid peroxidation. Conventional ferroptosis assays mainly rely on ex vivo or endpoint measurements, limiting their ability to dynamically monitor the spatiotemporal evolution of these events during ischemia–reperfusion. Recent advances in chemical sensing and optical imaging have enabled in situ detection of key ferroptosis-related nodes, such as Fe²⁺/labile iron pool, ROS/ONOO⁻, GSH/Cys/GPX4, H₂S/Cys–Met metabolism, and lipid peroxidation. In this review, we summarize sensing targets, reaction-based probe design, near-infrared and two-photon imaging, photoacoustic imaging, and multimodal validation strategies for AIS-related ferroptosis. Representative probes for H₂O₂, ONOO⁻, H₂S, Fe²⁺, and lipid peroxidation are discussed in the context of cellular models, oxygen-glucose deprivation/reoxygenation, middle cerebral artery occlusion/reperfusion, and in vivo brain imaging. We emphasize that a single probe signal cannot independently confirm ferroptosis and should be interpreted together with GPX4/ACSL4 alterations, MDA/4-HNE levels, tissue injury, neurological outcomes, and Fer-1/Lip-1 rescue experiments. Finally, we discuss current challenges, including limited tissue penetration, blood–brain barrier delivery, quantitative stability, probe safety, and clinical translation, and highlight future directions involving ratiometric, NIR/NIR-II, two-photon, multitarget, and imaging-guided validation strategies. Full article

The present study investigated the effects of different soil substrates (mixtures, based on a mica schist-derived soil), on the plant growth, soil chemical properties, nutritional status and photosynthetic parameters of chestnut (Castanea sativa) seedlings. The aim was to produce robust seedlings, suitable for plantations or reforestation of forest ecosystems, while also provide the necessary nutrient input during the early stages of development. A pot experiment was conducted comprising four treatments: (i) CONTROL, consisting of soil derived from mica schist weathering (MS), (ii) MS with inorganic fertilization (MS-FER), (iii) MS amended with forest floor, derived from evergreen broad-leaved trees (MS-EFF); and (iv) MS amended with forest floor, derived from chestnut trees (MS-CFF). Regarding seedlings’ growth, both types of forest floor exerted a positive effect similar to that observed after inorganic fertilization (MS-FER), while all treatments maintained a satisfactory and stable photosynthetic performance; however, an enhancement in leaf chlorophyll content (CCI) was specifically observed under inorganic fertilization. Except Olsen P and exchangeable Mg, which were significantly higher in the MS-FER treatment, all the other soil nutrients were higher either in the MS-EFF, or in the MS-CFF treatments. Overall, both types of forest floor proved effective as organic amendments, suggesting that they could serve as an alternative or complement to inorganic fertilization, potentially reducing fertilizer inputs in chestnut seedling production. Full article