ReMiND: Recovery of Missing Neuroimaging using Diffusion Models with Application to Alzheimer's Disease.

Objective: Missing data is a significant challenge in medical research. In longitudinal studies of Alzheimer's disease (AD) where structural magnetic resonance imaging (MRI) is collected from individuals at multiple time points, participants may miss a study visit or drop out. Additionally, technical issues such as participant motion in the scanner may result in unusable imaging data at designated visits. Such missing data may hinder the development of high-quality imaging-based biomarkers. Furthermore, when imaging data are unavailable in clinical practice, patients may not benefit from effective application of biomarkers for disease diagnosis and monitoring. Methods: To address the problem of missing MRI data in studies of AD, we introduced a novel 3D diffusion model specifically designed for imputing missing structural MRI (Recovery of Missing Neuroimaging using Diffusion models (ReMiND)). The model generates a whole-brain image conditional on a single structural MRI observed at a past visit or conditional on one past and one future observed structural MRI relative to the missing observation. Results: Experimental results show that our method can generate high-quality individual 3D structural MRI with high similarity to ground truth, observed images. Additionally, images generated using ReMiND exhibit relatively lower error rates and more accurately estimated rates of atrophy over time in important anatomical brain regions compared with two alternative imputation approaches: forward filling and image generation using variational autoencoders. Conclusion: Our 3D diffusion model can impute missing structural MRI data at a single designated visit and outperforms alternative methods for imputing whole-brain images that are missing from longitudinal trajectories.

Synthesis of a water-soluble naphthalene-based macrocycle and its host-guest properties.

Reported here is the synthesis of a naphthalene-based macrocycle bearing anionic carboxylato groups on the rims along with its complexation with cationic guests in aqueous media. The macrocycle could strongly bind guests in a molecular clip model with association constants of 106-107 M-1.

Current Progress of Ratiometric Fluorescence Sensors Based on Carbon Dots in Foodborne Contaminant Detection.

Carbon dots (CDs) are widely used in the detection of foodborne contaminants because of their biocompatibility, photoluminescence stability, and ease of chemical modification. In order to solve the interference problem of complexity in food matrices, the development of ratiometric fluorescence sensors shows great prospects. In this review, the progress of ratiometric fluorescence sensors based on CDs in foodborne contaminant detection in recent years will be summarized, focusing on the functionalized modification of CDs, the fluorescence sensing mechanism, the types of ratiometric fluorescence sensors, and the application of portable devices. In addition, the outlook on the development of the field will be presented, with the development of smartphone applications and related software helping to better enable the on-site detection of foodborne contaminants to ensure food safety and human health.

Modular Introduction of endo-Binding Sites in a Macrocyclic Cavity towards Selective Recognition of Neutral Azacycles.

Macrocycles with a functionalized interior, which is a general cavity feature of bioreceptors, are relatively hard to synthesize. Here we report a modular strategy to customize diverse endo-binding sites in the macrocycle cavity. Only two steps are needed. First, one V-shaped functional module bearing an embedded binding site and two 2,5-dimethoxyphenyls as reaction modules are connected. Then the condensation of the resulting monomer and paraformaldehyde directly produces the designed macrocycle. V-shaped monomers are deliberately used to guarantee the binding sites equatorially directing inward into the cavity and 2,5-dimethoxyphenyls standing axially as macrocycle sidewalls. More than a dozen endo-functionalized macrocyclic receptors have been constructed. Host-guest complexation studies show that macrocycle BP1-decorated interior OH moieties can strongly encapsulate neutral azacycles by forming inner hydrogen bonds, giving a high association constant of 4.59×104  M-1 in non-polar media.

Loss-Based Attention for Interpreting Image-Level Prediction of Convolutional Neural Networks.

Although deep neural networks have achieved great success on numerous large-scale tasks, poor interpretability is still a notorious obstacle for practical applications. In this paper, we propose a novel and general attention mechanism, loss-based attention, upon which we modify deep neural networks to mine significant image patches for explaining which parts determine the image decision-making. This is inspired by the fact that some patches contain significant objects or their parts for image-level decision. Unlike previous attention mechanisms that adopt different layers and parameters to learn weights and image prediction, the proposed loss-based attention mechanism mines significant patches by utilizing the same parameters to learn patch weights and logits (class vectors), and image prediction simultaneously, so as to connect the attention mechanism with the loss function for boosting the patch precision and recall. Additionally, different from previous popular networks that utilize max-pooling or stride operations in convolutional layers without considering the spatial relationship of features, the modified deep architectures first remove them to preserve the spatial relationship of image patches and greatly reduce their dependencies, and then add two convolutional or capsule layers to extract their features. With the learned patch weights, the image-level decision of the modified deep architectures is the weighted sum on patches. Extensive experiments on large-scale benchmark databases demonstrate that the proposed architectures can obtain better or competitive performance to state-of-the-art baseline networks with better interpretability. The source codes are available on: https://github.com/xsshi2015/Loss-based-Attention-for-Interpreting-Image-level-Prediction-of-Convolutional-Neural-Networks.

A Modular Synthetic Strategy for Functional Macrocycles.

Reported here is a molecule-Lego synthetic strategy for macrocycles with functional skeletons, involving one-pot and high-yielding condensation between bis(2,4-dimethoxyphenyl)arene monomers and paraformaldehyde. By changing the blocks, variously functional units (naphthalene, pyrene, anthraquinone, porphyrin, etc.) can be conveniently introduced into the backbone of macrocycles. Interestingly, the macrocyclization can be tuned by the geometrical configuration of monomeric blocks. Linear (180°) monomer yield cyclic trimers and pentamers, while V-shaped (120°, 90° and 60°) monomers tend to form dimers. More significantly, even heterogeneous macrocycles are obtained in moderate yield by co-oligomerization of different monomers. This series of macrocycles have the potential to be prosperous in the near future.

Selective complexation and efficient separation of cis/trans-1,2-dichloroethene isomers by a pillar[5]arene.

The complexation and separation of industrially important cis- and trans-1,2-dichloroethene (cis- and trans-DCE) isomers using perethylated pillar[5]arene (EtP5) are described. EtP5 exhibits considerable binding capability for the trans-DCE isomer over the cis-DCE in organic solution. Furthermore, nonporous adaptive crystals (NACs) of EtP5 can efficiently separate trans-DCE from a 50 : 50 (v/v) cis/trans-isomer mixture.

Efficient Separation of cis- and trans-1,2-Dichloroethene Isomers by Adaptive Biphen[3]arene Crystals.

Reported here is the highly efficient separation of industrially important cis- and trans-1,2-dichloroethene (cis-DCE and trans-DCE) isomers by activated crystalline 2,2',4,4'-tetramethoxyl biphen[3]arene (MeBP3) materials, MeBP3α. MeBP3 can be synthesized in excellent yield (99 %), and a cyclic pentamer is also obtained when using 1,2-dichloroethane as the solvent. The structure of MeBP3 in the CH3 CN@MeBP3 crystal displays a triangle-shape topology, forming 1D channels through window-to-window packing. Desolvated crystalline MeBP3 materials, MeBP3α, preferentially adsorb cis-DCE vapors over its trans isomer. MeBP3α is able to separate cis-DCE from a 50:50 (v/v) cis/trans-isomer mixture, yielding cis-DCE with a purity of 96.4 % in a single adsorption cycle. Single-crystal structures and powder X-ray diffraction patterns indicate that the uptake of cis-DCE triggers a solid-state structural transformation of MeBP3, suggesting the adaptivity of MeBP3α materials during the sorption process. Moreover, the separation can be performed over multiple cycles without loss of separation selectivity and capacity.

Supervised graph hashing for histopathology image retrieval and classification.

In pathology image analysis, morphological characteristics of cells are critical to grade many diseases. With the development of cell detection and segmentation techniques, it is possible to extract cell-level information for further analysis in pathology images. However, it is challenging to conduct efficient analysis of cell-level information on a large-scale image dataset because each image usually contains hundreds or thousands of cells. In this paper, we propose a novel image retrieval based framework for large-scale pathology image analysis. For each image, we encode each cell into binary codes to generate image representation using a novel graph based hashing model and then conduct image retrieval by applying a group-to-group matching method to similarity measurement. In order to improve both computational efficiency and memory requirement, we further introduce matrix factorization into the hashing model for scalable image retrieval. The proposed framework is extensively validated with thousands of lung cancer images, and it achieves 97.98% classification accuracy and 97.50% retrieval precision with all cells of each query image used.

Characterization of the Pathogenesis of H10N3, H10N7, and H10N8 Subtype Avian Influenza Viruses Circulating in Ducks.

Three H10 subtype avian influenza viruses were isolated from domestic ducks in China, designated as SH602/H10N8, FJ1761/H10N3 and SX3180/H10N7, with an intravenous pathogenicity index (IVPI) of 0.39, 1.60, and 1.27, respectively. These H10 viruses showed a complex pathology pattern in different species, although full genome characterizations of the viruses could not identify any molecular determinant underlying the observed phenotypes. Our findings describe the pathobiology of the three H10 subtype AIVs in chickens, ducks, and mice. FJ1761/H10N3 evolved E627K and Q591K substitutions in the gene encoding the PB2 protein in infected mice with severe lung damage, suggesting that H10 subtype avian influenza viruses are a potential threat to mammals.