This study aimed to investigate the mechanisms underlying impaired oral wound healing in aged individuals, with a focus on the roles of neutrophil extracellular traps (NETs) and NLRP3 inflammasome activation. Palatal wound healing was compared between young and aged mice. We assessed the change of neutrophil infiltration, NETs formation, and NLRP3 inflammasome activation in mucosa using immunohistochemistry, RNA sequencing, and flow cytometry. In vitro co-culture assays evaluated NETs' effects on fibroblast function and NLRP3 inflammasome activation. The therapeutic potential of NETs-NLRP3 clearance was tested in vivo with DNase1 and MCC950. To validate clinical relevance, human single-cell RNA sequencing data were further analyzed. Aged mice exhibited delayed oral wound healing, characterized by increased neutrophil recruitment and a heightened propensity for fibrin-induced NETs formation. NETs activated the NLRP3 inflammasome in mucosal fibroblasts, suppressing their proliferation and migration. Notably, targeting NETs-NLRP3 inflammasome reduced inflammation and accelerated wound closure in the aged group. Corresponding human data further revealed increased NETs-associated neutrophil subsets and enriched NLRP3 pathways in aged oral mucosa. Our findings demonstrate that enhanced NETs formation and subsequent NLRP3 inflammasome impair wound healing in aged oral mucosa. Targeting NETs-NLRP3 thus represents a promising therapeutic strategy to improve mucosal repair.
To provide kinetic data for a key radical-induced subnetwork in the oxidation of oxygenated polyether fuels, the potential energy surfaces for H, CH3, OH, and HO2-mediated H-abstraction from monoglyme (MON, CH3OCH2CH2OCH3), together with the subsequent isomerization and β-scission decomposition reactions of the resulting fuel radicals, were constructed at the CCSD(T)/CBS//M06-2X/6-311++G (d, p) level. High-pressure-limit and pressure-dependent rate constants were predicted using master-equation calculations over 300-2000 K. The calculated H-abstraction barriers for MON range from -1.6 to 18.5 kcal mol-1, corresponding to the reactivity order of OH > H > CH3 > HO2. The site-dependent barriers are consistent with the calculated C-H bond dissociation energies (BDEs) and the stabilization of the resulting radicals. The subsequent isomerization is competitive at low temperatures. With increasing temperature, the C-O bond cleavage pathways become kinetically favored within the investigated post-H-abstraction unimolecular subnetwork, particularly the channel leading to methoxy acetaldehyde (CH3OCH2CHO). The H-abstraction of methoxy acetaldehyde by the four radicals mentioned above and the subsequent reactions were further investigated using the same methodology. The results indicate that the reactivity sequence of H-abstraction in methoxy acetaldehyde is consistent with that of MON. The methoxy acetaldehyde-derived radicals preferentially evolve through lower-barrier decomposition pathways rather than extensive isomerization. Comparisons with related ether systems and uncertainty analyses support the reliability of the calculated rate constants within the expected theoretical uncertainty. Implementation of the updated H-abstraction kinetics into a detailed MON oxidation mechanism affects the predicted ignition delay times. These results characterize the H-abstraction and subsequent unimolecular-reaction subnetwork from MON to methoxy acetaldehyde and smaller products, providing mechanism kinetic data for future refinement of oxygenated polyether combustion models.
Patients diagnosed with cancer are more prone to depressive symptoms and cognitive dysfunction. This study aimed to explore the dynamic relationship between depressive symptoms and cognitive function in middle-aged and older adults before and after a cancer diagnosis. The data were derived from Waves 1 to 5 of the China Health and Retirement Longitudinal Study (CHARLS). The cognitive functioning and depressive symptoms were measured using the validated Chinese version of the Mini-Mental State Examination (MMSE) and the Center for Epidemiologic Studies Depression Scale-10 (CESD-10), respectively. Cross-sectional network analysis was utilized for constructing the contemporaneous network, and cross-lagged panel network analysis was subsequently employed for longitudinal analysis. In the temporal network, greater "Hope" before cancer diagnosis was associated with improved "Recall" after diagnosis (β = 0.112) between cognitive function and depressive symptoms. "Attention" (predictability = 0.797) exhibited the highest in-prediction values among all nodes. In contrast, "Drawing" (influence = 1.333) exerted the strongest out-prediction on other symptoms in the cross-lagged network. This study utilized the CHARLS database and employed cross-lagged network analysis to elucidate the dynamic mechanisms of influence between depressive symptoms and cognitive function before and after a cancer diagnosis. This study identified the strongest predictive edges in the temporal network, providing new targets for clinical interventions regarding depressive symptoms and cognitive function before and after a cancer diagnosis. Furthermore, we identified the node with the strongest out-prediction in the temporal network spanning from pre-diagnosis to post-diagnosis, which is critical for developing targeted intervention strategies.
Knowledge of protein acid sensitivity remains sparse and is largely derived from low-throughput, enzyme-specific assays. We used a scalable framework to map acid stability across the Escherichia coli proteome to assess the acid stability of 1 675 unique proteins, estimating pH50 values for over 90% of them. The parameter pH50 was defined as the pH value at which only 50% of the initial protein remains in solution following acid treatment. Proteome-wide pH50 values ranged from 2.28 to 6.33 (median 5.11). Approximately 9% of detected proteins remained stable across all tested pH conditions. Our results align with published data and the assay of citrate synthase (GltA) performed here. Protein acid stability differed significantly by subcellular localization: periplasmic proteins were relatively more abundant in the acid-stable group, cytoplasmic proteins were abundant at pH50 values 4.5-5.5 and inner membrane proteins at higher pH50 between 5.5 and 6.0. Outer membrane proteins were too few to draw strong conclusions regarding enrichment within specific pH₅₀ groups. Notably, the periplasmic binding protein of the molybdate ABC transporter (ModA), was enriched after incubation at low pH. Estimated pH50 values showed no correlation with protein isoelectric point and molecular weight. Together, this work provides the first proteome-wide map of protein acid stability and establishes a general framework for studying different chemical stressors.
Ferroptosis-based cancer therapy shows promise in tumor suppression but is often limited by insufficient activation of antitumor immunity. Here, a hepatocellular carcinoma (HCC) targeted nanoplatform was constructed by coating FePDA nanozymes with Hepa1-6 cell membranes overexpressing ovalbumin (OVA) and CD40 ligand (CD40L), generating tumor membrane-derived OVA/CD40L-functionalized FePDA nanoparticles (FePDA-TMOC). This hybrid architecture integrates homologous tumor targeting with iron homeostasis disruption and co-stimulatory immune activation within a single nanosystem. Mechanistically, the FePDA core induces sustained depletion of intracellular glutathione (GSH), leading to enhanced lipid peroxidation and promoting ferroptosis by disrupting the GPX4-SLC7A11 axis. Simultaneously, the OVA/CD40L-modified membrane promotes efficient uptake by antigen-presenting cells and enhances pro-inflammatory immune activation through co-stimulatory signaling. In vitro experiments demonstrated that FePDA-TMOC selectively induces ferroptosis in Hepa1-6 cells and drives macrophage polarization toward a pro-inflammatory phenotype. Following systemic administration, FePDA-TMOC preferentially accumulates in tumors, markedly increases CD8+ T cell infiltration, elevates pro-inflammatory cytokine levels, and suppresses tumor growth and metastasis without observable systemic toxicity. By integrating ferroptosis induction with antitumor immune activation, this work highlights a rational strategy for transforming immune "cold" tumors into "hot" tumors, providing a potential nanotherapeutic approach for HCC treatment.
The β-xylosidase Cbxyl1 of GH43-1 subfamily from Cellulomonas bogoriensis 69B4T was expressed in Escherichia coli. It showed optimal activity at 30 °C and pH 7.0, retaining 40% of its activity at 10 °C. It displayed Ca2+-dependent activity and remarkable salt tolerance, maintaining 87% activity in the presence of 4 M KCl. Cbxyl1 hydrolyzed p-nitrophenyl-β-D-xylopyranoside, p-nitrophenyl-α-L-arabinofuranoside, and xylooligosaccharides, but could not cleave arabinose from arabinoxylan oligosaccharides. Structural analysis reveals that Cbxyl1 adopts a five-bladed β-propeller fold characteristic of the GH43 family, with Asp5, Asp125, and Glu212 as the catalytic residues, along with a high proportion of random coils and a low number of salt bridges and hydrogen bonds. Phylogenetic analysis and comparison with other GH43-1 enzymes suggest functional diversity within this subfamily, including both strict β-xylosidases and bifunctional enzymes. Collectively, these results establish Cbxyl1 as a cold- and salt-tolerant β-xylosidase with strict substrate selectivity, thereby providing a candidate enzyme together with experimental information for further investigation and potential application, and contributing to a refined understanding of the substrate specificity landscape within the GH43-1 subfamily.
Flexible and transparent triboelectric nanogenerators (TENGs) have exhibited tremendous application potential in the fields of human-machine interfaces (HMIs), invisible anticounterfeiting, and environmental monitoring. However, developing electrode materials that simultaneously achieve high stability, excellent electrical conductivity, and good optical transparency remains a key challenge. Existing electrodes such as aluminum, indium tin oxide (ITO), and silver nanowires (Ag NWs) are electropositive, which greatly limits their applications in flexible wearable electronics and human-machine interfaces because human skin, clothing, and gloves are usually electropositive. Here, in this work, we propose flexible and transparent triboelectric electrodes through rationally designing a MXene-Ag NWs-MXene (MAM) 3D stacked interconnect structure. The electrodes exhibit high transparency, flexibility, stability, and electronegativity. The top layer MXene acts as both an electronegative layer and protective layer, which can help generate electrical output after contact and separation with electropositive materials (clothes, gloves, and human skins) and protect the Ag NWs from being oxidated even under circumstance of high temperature (85 °C) and humidity (99% RH). The Ag NW layer performs as a conductive layer to provide a cross-interface electron transport channel even though the top MXene layer is oxidated. The bottom MXene layer serves as both an adhesive and a conductive component. The MAM electrodes demonstrate a high optical transmittance of ∼84% at 550 nm and maintain stable conductivity and output performance after 1,000 bending and twisting cycles with different angles. To further demonstrate the practical potential, the MAM electrode-based TENG array is designed as invisible HMIs to control the LED and as a security overlay for keyboards to realize the function of identity recognition through machine learning. This work provides a strategy for developing advanced flexible and transparent electrodes in HMI systems.
Triple-negative breast cancer (TNBC) is a particularly aggressive subtype of breast cancer, known for its high malignancy, elevated risk of recurrence and metastasis, and limited therapeutic options, resulting in the poorest prognosis among breast cancer types. This study explores the anticancer effects of VALD-3, a Schiff base ligand derivative, on breast cancer cells. While VALD-3 exhibited cytotoxic effects on both triple-negative breast cancer (TNBC) and estrogen receptor-positive (ER+) MCF-7 cells, it more potently inhibited TNBC cell viability. More importantly, VALD-3 induced characteristic pyroptotic features selectively in TNBC cells, including cell swelling, balloon-like protrusions, and the release of inflammatory cytokines due to pore formation in the plasma membrane, ultimately inhibiting tumor growth. Mechanistically, VALD-3 increased reactive oxygen species (ROS) levels and JNK phosphorylation, leading to the recruitment of Bax to the mitochondria and the formation of a Bax-Bcl-2 heterodimer, which facilitated cytochrome c release into the cytoplasm. This cascade activated caspase-3 and triggered gasdermin E(GSDME)- dependent pyroptosis in TNBC cells. Thus, VALD-3 treatment initiated the ROS/JNK/Bax-mitochondrial apoptosis pathway, leading to caspase-3 activation and GSDME cleavage, thereby executing pyroptosis. These findings suggest that GSDME-dependent pyroptosis is a novel mechanism by which VALD-3 eradicates cancer cells and offer new insights into potential clinical applications for anticancer therapies.
Allele-specific expression (ASE) is a key regulatory mechanism linking genetic variation to phenotypic diversity. This study conducted a genome-wide ASE analysis in embryonic (brain, kidney, liver) and extraembryonic (chorion) tissues of Landrace × Meishan crossbred pigs. By integrating DNA sequencing of parental and F1 hybrids and RNA-sequencing of F1 tissues, a high-resolution ASE landscape was built via strict allele-specific SNP identification. Genomically, ASE genes (ASEG) were enriched on chromosomes 6, 7, 12, X and sparse on chromosome 11, and can be classified into extreme/moderate types by expression pattern and parental bias. Paternal ASEGs had higher expression levels, while maternal ones had higher abundance. Functionally, tissue-concordant ASEGs supported embryo/placenta basal development, tissue-specific ones matched organ core functions; maternal ASEGs were enriched in mitochondrial pathways, paternal ones in cell motility and RNA transcription. Besides, chorion had more allele-specific methylated regions than embryonic tissues, with opposite ASE directions. We proposed "orthogonal dual regulatory dimensions" for the relationship between ASE and tissue-specific expression. We also proposed regulatory models of extreme and moderate ASEGs, functional division modes between embryo and extraembryonic tissues, and between paternal and maternal alleles. This study clarifies tissue-specific ASE patterns and mechanisms, providing a framework for future ASE research.
Nitrofurazone, a commonly used nitrofuran antibiotic, is widely applied in antimicrobial therapy as well as in aquaculture and livestock farming. However, its residues can pose risks to human health and cause environmental pollution. Therefore, the development of a highly sensitive detection method for nitrofurazone is of great significance. In this study, a novel core-shell composite material, UiO@TATF COF, was prepared by in situ growing TATF COF on the UiO-66-NH2 surface through a Schiff-base reaction using 4,4',4″-(1,3,5-triazine-2,4,6-triyl) tris(benzaldehyde) (TFPT) and 1,3,5-tris(4-aminophenyl) benzene (TAPB) as monomers. Subsequently, by further introducing the conductive material carbon black (CB), a novel electrochemical sensor UiO@TATF COF/CB/GCE was successfully constructed. Covalently linked UiO@TATF COF achieves a robust core-shell interface and possesses a large specific surface area and abundant active functional groups. The resultant sensor system exhibits clear synergistic effects, enabling electrocatalysis and sensitive detection toward nitrofurazone in food, environmental, and ointment samples with a detection limit of 0.05 µM (S/N = 3).
Herein, by integrating benzo[1,2-c:4,5-c']bis([1,2,5]thiadiazole) with strong electron-withdrawing capacity and 4,4'-azanediyldibenzoic acid with strong electron-donating capacity into one donor-acceptor-donor type organic linker, a near-infrared-II (NIR-II) emissive zirconium-tetracarboxylate framework, HIAM-4030, was reported. HIAM-4030 possesses an ftw topology and shows a maximum emission peak at 1052 nm. The HIAM-4030-based nanocomposite exhibits targeted delivery, anti-inflammatory efficacy, and real-time, noninvasive visualization of inflammatory bowel disease via NIR-II fluorescence imaging. This work sheds light on the rational design and construction of luminescent metal-organic frameworks with NIR-II emission behaviors for biomedical applications.
This study investigates the performance of a sequencing batch electro-membrane bioreactor (SB-EMBR) operated under low electric charge loading (39.9 mAh L⁻1), focusing on the effects of aeration intensity on treatment performance, biomass activity, and membrane fouling. The reactor was operated at a current density of 10 A m-2 and three specific aeration demand levels (SADₘ = 0.48, 0.24, and 0.12 m3 m⁻2 h⁻1). Organic matter and phosphorus removals remained consistently high (> 90% COD removal; TP < 1.0 mg L⁻1) regardless of aeration intensity. In contrast, ammonium removal efficiency declined from 99.5 to 74.7% as the SADₘ decreased from 0.48 to 0.12 m3 m⁻2 h⁻1. Batch assays revealed reduced activity of polyphosphate-accumulating organisms under oxygen-limited conditions. The pronounced decrease in the P-release/COD-uptake ratio from 0.172 to 0.0164 mol P mol⁻1 C indicates that TP removal at low dissolved oxygen became predominantly governed by chemical coagulation rather than the biological phosphorus removal process. The calculated Al/P molar ratio of 2.32 mol Al mol⁻1 P was sufficient to sustain phosphorus removal through both precipitation and adsorption onto aluminum hydroxides. Reduced aeration favored anoxic phosphorus uptake, increasing the denitrifying phosphate assimilation potential from 18 to 41%. The membrane fouling rate increased from 1.02 to 4.81 kPa d⁻1 as aeration decreased, mainly due to diminished shear forces, soluble microbial products accumulation (+ 205%), floc size reduction (- 50.1%), and higher capillary suction time (+ 123%). Owing to the short current application (1.6 h d⁻1), the additional electrocoagulation cost was only 0.07 USD m⁻3, lower than values reported for continuous-flow EMBRs. Overall, operation of the SB-EMBR under reduced electric charge loading demonstrated promising energy efficiency while maintaining stable and satisfactory pollutant removal even at low aeration intensities.
Symptomatic lumbar spinal stenosis (sLSS) is associated with sagittal imbalance and degeneration of paraspinal muscles. While morphological changes are well documented, the functional relationships between electromyographic paraspinal fatigability, endurance, and global imbalance remain unclear. This study aimed to examine these relationships in preoperative patients with sLSS. In this cross-sectional study, 109 preoperative patients with sLSS underwent magnetic resonance imaging, EOS radiography, and a modified Biering-Sørensen test with bilateral surface electromyography (EMG) of the longissimus, iliocostalis and multifidus muscles. Paraspinal muscle fatigability was defined by three EMG-based indices: slope of median frequency (MDF) decline (Hz/s), normalized slope (%/s), and absolute MDF decrease (Hz). Endurance was defined as the duration of the modified Biering-Sørensen test. Global spinal imbalance was quantified using the Full Balance Integrated (FBI) score. Associations between fatigability, endurance, and FBI were assessed using Pearson correlation and stepwise multiple regression analyses. Valid EMG data were available for 90 of the 109 patients who performed the modified Biering-Sørensen test. Greater absolute MDF decreases (but not MDF slopes) were significantly associated with lower FBI scores (r = -0.38, p < 0.001). Longer endurance times correlated with both lower FBI scores (r = -0.41, p < 0.001) and greater absolute MDF decreases (r = 0.40, p < 0.001). Regression analysis revealed that absolute MDF decrease (β = -0.3, p = 0.004), endurance time (β = -0.1, p = 0.005), and age (β = 0.31, p < 0.001) independently predicted FBI, explaining 35% of its variance. Mediation analysis confirmed that the association between absolute MDF decrease and FBI was predominantly direct (71%) and only partially mediated through endurance time. In preoperative patients with sLSS, both reduced paraspinal endurance and lower absolute MDF decrease were independently associated with greater global spinal imbalance. The association between absolute MDF decrease and imbalance was only partially mediated through endurance time, indicating that these two measures capture complementary aspects of paraspinal muscle function. Since lower absolute MDF decrease coincided with shorter endurance times, reduced electromyographic fatigability in this population likely reflects diminished physiological reserve rather than enhanced neuromuscular efficiency.
We examined the effects of fucoidan, a sulfated polysaccharide, administered concurrently with the onset of a high-fat diet (HFD) for a short period in C57BL/6 N mice to evaluate its immunological effects under obesity-inducing conditions. Numerous genes that were downregulated in the HFD group relative to the control-diet group were restored in the HFD + fucoidan group; these genes were mainly associated with inflammatory responses and cellular functions, including migration and viability, as identified by DNA microarray analysis. Upstream regulator analysis in ingenuity pathway analysis identified numerous cytokines and inflammatory mediators. Quantitative PCR confirmed that the expression of selected inflammatory cytokine genes-particularly Ifn-γ and Il-5-was decreased in the HFD group but restored in the HFD + fucoidan group. Our study suggests that fucoidan administered concurrently with an HFD attenuates HFD-induced downregulation of inflammatory and immune-related functions, including T-cell and B-cell functions, and may improve HFD-induced immune dysfunction and imbalance.
Hydrogels are cross-linked polymeric networks with wide applications in drug delivery, tissue engineering, biosensing, and environmental remediation. These hydrogels additionally host living cells, small molecules, and biological propagules, which further expand the applications of these materials. However, most, if not all, fabrication methods require covalent modifications. In this work, by deliberately selecting polymers with a known propensity to phase separate and formulating compositions far from the binodal boundary, we demonstrate the propensity of the system to transition directly into viscoelastic liquids or gels. This behavior is demonstrated using a model system of poly(ethylene glycol) (PEG) and dextran (DEX). We carried out rheological studies to provide insights into the viscoelastic behavior of these gels. We systematically characterized the gels through colorimetric assays, FTIR, MALDI-TOF, and thermogravimetric analysis (TGA) to discern the molecular compositions and solvent content of the gels. These experimental findings are supplemented with coarse-grained (CG) simulation insights to investigate the mechanistic origins of phase separation propensity with varying molecular weights of DEX. We utilized coexisting densities in the two phases using CG simulations to predict the role of DEX molecular weight in the partitioning of PEG and DEX in the two phases. Finally, we exploit the fabricated gel's ability to encapsulate live cells, antibiotics, and plant seeds. We anticipate that this ATPS-based fabrication technique will provide a scalable, cross-linker-free route to multifunctional gels, enabling advanced applications in drug delivery and responsive materials.
Elevated intracranial pressure (ICP) remains a challenge and immediate danger in patients with acute brain injury. Current literature supports the use of a single hyperosmolar agent to reduce ICP. However, the efficacy and safety of overlapping hyperosmolar agents have not been evaluated. The objectives of this study are to analyze the ICP lowering effects and evaluate the safety profile of overlapping hyperosmolar therapy. This single-center, retrospective study included adults (≥18 y) admitted to the Neurosciences Intensive Care Unit between May 2018 and September 2024 with acute brain injury leading to intracranial hypertension refractory to a single hyperosmolar agent and an ICP monitor. They received overlapping doses of mannitol and hypertonic sodium approximately every 3 hours. We evaluated the impact of overlapping hyperosmolar therapy on ICP. Safety outcomes included acute kidney injury incidence, electrolyte disturbances, and volume status changes. Forty-six patients were included in this study with a median (IQR) age of 46.5 (35.9, 62.4) years. Median (IQR) dose of mannitol was 40.1 (34.6, 50.0) g, while median (IQR) dose of hypertonic sodium was 120 (90, 125) mEq. About 86% of patients experienced a decrease in ICP after initiation of overlapping hyperosmolar therapy. Six acute kidney injuries occurred, and 7 patients experienced pulmonary edema. Median (IQR) peak sodium, chloride, and osmolar gap values were 149 (144, 155) mEq/L, 115 (111, 119) mEq/L, and 5.7 (0.7, 9.0) mOsm/kg. Overlapping hyperosmolar therapy may provide a transient ICP reduction without excessive adverse effects in patients with refractory ICP to a single hyperosmolar agent.
Fetal fat accretion follows a spatiotemporal pattern. Fetuses who are small-for-gestational-age (SGA) demonstrate reduced fat accumulation, but whether specific body regions are disproportionately affected remains unclear. To characterize regional fat differences between SGA and appropriate-for-gestational-age (AGA), assess the modifying effects of SGA-onset timing and cerebroplacental ratio (CPR), and evaluate associations with neonatal morbidity. SGA pregnancies were prospectively recruited, and AGA controls retrospectively identified. SGA was defined as estimated fetal weight <10th centile and classified as early-onset (< 32 weeks) or late-onset (≥ 32 weeks), with CPR categorized as normal (≥ 5th centile) or abnormal (< 5th centile). Fetal Magnetic resonance imaging was performed at 3-T using T1-weighted two-point Dixon. Subcutaneous fat was segmented and subdivided into cheeks, trunk, upper and lower limbs. Fat signal fraction and fat mass were computed, with regional fat mass adjusted to global fat mass. Linear mixed models compared groups, and univariate logistic regression assessed associations with adverse outcomes. Sixty-four participants (35 SGA, 29 AGA) were included. Fat signal fraction was significantly lower in SGA across all regions (P<0.001). Upper limb adjusted fat mass was reduced in SGA (P=0.043), while other regions showed no differences (P≥0.08). Fat signal fraction and adjusted fat mass did not differ by SGA onset or CPR status. Lower fat signal fraction in all regions was associated with higher morbidity rates, whereas adjusted fat mass was not. SGA fetuses exhibit globally reduced lipid content with a disproportionate upper limb fat mass deficit, suggesting selective vulnerability of peripheral fat depots.
Upstream sample preparation often limits the sensitivity and reproducibility of bioanalysis, particularly for rare or weakly magnetic targets. Here, we present a tunable high-gradient magnetic separation system based on an ordered Fe-Ni wire-array column, which generates stable and quantifiable local high-gradient regions under a uniform external magnetic field. The magnetic flux density (B) and volumetric flow rate (Q) can be independently adjusted, allowing separation conditions to be systematically regulated and interpreted using the Mason number (Mn). Using magnetically homogeneous metHb-RBCs and heterogeneous SPION-labeled RAW264.7 macrophages as model systems, we demonstrate tunable capture and enrichment behavior under different magnetic-field and flow-rate conditions. For metHb-RBCs, near-equal Mn conditions produced comparable captured-cell fractions across different B-Q combinations, supporting the use of Mn as a force-flow descriptor for weak and relatively homogeneous magnetic cells. For SPION-labeled RAW264.7 macrophages, OTMS-derived single-cell magnetic moment distributions showed preferential enrichment of higher-moment cells, while near-matched apparent Mn conditions gave similar enrichment trends despite pronounced magnetic heterogeneity. These results show that the ordered wire-array column provides a geometry-defined and quantitatively interpretable HGMS microenvironment for magnetic cell enrichment. The system may also provide a useful basis for future studies on other magnetic targets.
Parkinson's disease (PD) diagnosis remains challenging because subtle neural alterations may be difficult to capture using conventional clinical assessment alone. This study proposes an attention-based deep learning framework for classifying PD from resting-state EEG with minimal preprocessing and leakage-safe evaluation. Raw EEG recordings were first partitioned at the subject level. Within each fold, the selected motor-related EEG channel was decomposed into canonical sub-bands using discrete wavelet transform, and the resulting sub-band signals were then segmented into overlapping temporal windows. Each sub-band window was transformed into a time-frequency spectrogram using the short-time Fourier transform and classified using a ResNet-101 backbone enhanced with dual channel-spatial attention. Hyperparameters were optimized using an Enhanced Adaptive Hybrid Covariance Matrix Adaptation Evolution Strategy (AH-CMA-ES), applied only within training/internal-validation subjects in each fold. Model performance was evaluated on two independent public EEG datasets, UC San Diego and University of Iowa, using subject-wise nested leave-one-subject-out cross-validation. In each fold, the held-out subject was excluded from training, augmentation, hyperparameter optimization, early stopping, and model selection. The proposed framework achieved 95.2% segment-level and 96.77% subject-level accuracy on UCSD, and 92.1% segment-level and 92.86% subject-level accuracy on Iowa, with subject-level decisions obtained by majority voting over non-augmented test segments. In addition to classification, sub-band topographical analysis provided exploratory neurophysiological interpretation across canonical EEG rhythms, revealing patterns consistent with reported PD-related oscillatory alterations. These findings suggest that resting-state EEG combined with attention-based deep learning can support robust, interpretable PD classification, while larger heterogeneous cohorts are needed to further validate clinical generalizability.