Search Results (5)

Urban pluvial flooding in highly developed basins is challenging to forecast in real time because detailed 1D–2D hydraulic models are computationally expensive, while purely data-driven approaches often lack physical consistency. This study aims to enable operational urban flood nowcasting by proposing a model-informed AI framework for short-term stream-stage and urban inundation prediction in the Bisan-dong district of Anyang, South Korea, where the Anyang and Hagui Streams frequently overflow. A gated recurrent unit (GRU) network was trained on 10 min rainfall and stream-stage observations from 2011 to 2018 and independently validated on 2019–2022 data at four gauges to forecast stream stage at lead times of 10–60 min. In parallel, an ANN–CNN inundation surrogate was trained on 864 XP-SWMM 1D–2D simulation scenarios, forced by design storms and downstream water-level boundary conditions, to produce 256 × 256 maps of maximum inundation depth. The GRU model achieved ${\mathrm{R}}^2$ and Nash–Sutcliffe efficiency values generally above 0.95, with a mean absolute percentage error (MAPE) below approximately 5% for 10–30-min lead times; performance decreased but remained useful at 60 min. The inundation surrogate reproduced XP-SWMM results with an MAPE of 8.89% for inundation area and 19.49% for grid-based depth. Together, the ANN–CNN system enables rapid generation of high-resolution flood maps and provides a practical basis for AI-assisted urban flood nowcasting and risk management. Full article

Recent advancements in decoding electroencephalogram (EEG) signals for motor imagery tasks have shown significant potential. However, the intricate time–frequency dynamics and inter-channel redundancy of EEG signals remain key challenges, often limiting the effectiveness of single-scale feature extraction methods. To address this issue, we propose the Dual-Branch Blocked-Integration Self-Attention Network (DB-BISAN), a novel deep learning framework for EEG motor imagery classification. The proposed method includes a Dual-Branch Feature Extraction Module designed to capture both temporal features and spatial patterns across different scales. Additionally, a novel Blocked-Integration Self-Attention Mechanism is employed to selectively highlight important features while minimizing the impact of redundant information. The experimental results show that DB-BISAN achieves state-of-the-art performance. Also, ablation studies confirm that the Dual-Branch Feature Extraction and Blocked-Integration Self-Attention Mechanism are critical to the model’s performance. Our approach offers an effective solution for motor imagery decoding, with significant potential for the development of efficient and accurate brain–computer interfaces. Full article

Tau is an intrinsically disordered protein involved in several neurodegenerative diseases where a common hallmark is the appearance of tau aggregates in the brain. One common approach to elucidate the mechanisms behind the aggregation of tau has been to recapitulate in vitro the self-assembly process in a fast and reproducible manner. While the seeding of tau aggregation is prompted by negatively charged cofactors, the obtained fibrils are morphologically distinct from those found in vivo. The Tau AD core fragment (TADC, tau 306–378) has emerged as a new model and potential solution for the cofactor-free in vitro aggregation of tau. Here, we use TADC to further study this process combining multiple amyloid-detecting fluorophores and fibril bioimaging. We confirmed by transmission electron microscopy that this fragment forms fibrils after quiescent incubation at 37 °C. We then employed a panel of eight amyloid-binding fluorophores to query the formed species by acquiring their emission spectra. The results obtained showed that nearly all dyes detect TADC self-assembled species. However, the successful monitoring of TADC aggregation kinetics was limited to three fluorophores (X-34, Bis-ANS, and pFTAA) which yielded sigmoidal curves but different aggregation half-times, hinting to different species being detected. Altogether, this study highlights the potential of using multiple extrinsic fluorescent probes, alone or in combination, as tools to further clarify mechanisms behind the aggregation of amyloidogenic proteins. Full article

The coordination of zinc ions by histidine residues of amyloid-beta peptide (Aβ) plays a critical role in the zinc-induced Aβ aggregation implicated in Alzheimer’s disease (AD) pathogenesis. The histidine to arginine substitution at position 6 of the Aβ sequence (H6R, English mutation) leads to an early onset of AD. Herein, we studied the effects of zinc ions on the aggregation of the Aβ42 peptide and its isoform carrying the H6R mutation (H6R-Aβ42) by circular dichroism spectroscopy, dynamic light scattering, turbidimetric and sedimentation methods, and bis-ANS and thioflavin T fluorescence assays. Zinc ions triggered the occurrence of amorphous aggregates for both Aβ42 and H6R-Aβ42 peptides but with distinct optical properties. The structural difference of the formed Aβ42 and H6R-Aβ42 zinc-induced amorphous aggregates was also supported by the results of the bis-ANS assay. Moreover, while the Aβ42 peptide demonstrated an increase in the random coil and β-sheet content upon complexing with zinc ions, the H6R-Aβ42 peptide showed no appreciable structural changes under the same conditions. These observations were ascribed to the impact of H6R mutation on a mode of zinc/peptide binding. The presented findings further advance the understanding of the pathological role of the H6R mutation and the role of H6 residue in the zinc-induced Aβ aggregation. Full article

Multi-wavelength fluorescence spectroscopy was evaluated in this work as tool for real-time monitoring of antibody aggregation in CHO fed-batch cultivations via partial least square (PLS) modeling. Therefore, we used the extrinsic fluorescence dyes 1-anilinonaphthalene-8-sulfonate (ANS), 4,4′-bis-1-anilinonaphthalene-8-sulfonate (Bis-ANS), or Thioflavin T (ThT) as medium additives. This is a new application area, since these dyes are commonly used for aggregate detection during formulation development. We determined the half maximum inhibitory concentrations of ANS (203 ± 11 µmol·L⁻¹), Bis-ANS (5 ± 0.5 µmol·L⁻¹), and ThT (3 ± 0.2 µmol·L⁻¹), and selected suitable concentrations for this application. The results showed that the emission signals of non-covalent dye antibody aggregate interaction superimposed the fluorescence signals originating from feed medium and cell culture. The fluorescence datasets were subsequently used to build PLS models, and the dye-related elevated fluorescence signals dominated the model calibration. The soft sensors based on ANS and Bis-ANS signals showed high predictability with a low error of prediction (1.7 and 2.3 mg·mL⁻¹ aggregates). In general, the combination of extrinsic dye and used concentration influenced the predictability. Furthermore, the ThT soft sensor indicated that the intrinsic fluorescence of the culture might be sufficient to predict antibody aggregation online. Full article