LIFU-responsive nanomedicine enables acoustic droplet vaporization-induced apoptosis of macrophages for stabilizing vulnerable atherosclerotic plaques.

Due to the high risk of tearing and rupture, vulnerable atherosclerotic plaques would induce serious cardiovascular and cerebrovascular diseases. Despite the available clinical methods can evaluate the vulnerability of plaques and specifically treat vulnerable plaques before a cardiovascular event, but the efficiency is still low and undesirable. Herein, we rationally design and engineer the low-intensity focused ultrasound (LIFU)-responsive FPD@CD nanomedicine for the highly efficient treatment of vulnerable plaques by facilely loading phase transition agent perfluorohexane (PFH) into biocompatible PLGA-PEG-PLGA nanoparticles (PPP NPs) and then attaching dextran sulphate (DS) onto the surface of PPP NPs for targeting delivery. DS, as a typical macrophages-targeted molecule, can achieve the precise vaporization of NPs and subsequently controllable apoptosis of RAW 264.7 macrophages as induced by acoustic droplet vaporization (ADV) effect. In addition, the introduction of DiR and Fe3O4 endows nanomedicine with near-infrared fluorescence (NIRF) and magnetic resonance (MR) imaging capabilities. The engineered FPD@CD nanomedicine that uses macrophages as therapeutic targets achieve the conspicuous therapeutic effect of shrinking vulnerable plaques based on in vivo and in vitro evaluation outcomes. A reduction of 49.4% of vascular stenosis degree in gross pathology specimens were achieved throughout the treatment period. This specific, efficient and biosafe treatment modality potentiates the biomedical application in patients with cardiovascular and cerebrovascular diseases based on the relief of the plaque rupture concerns.

Linking Multi-Layer Dynamical GCN With Style-Based Recalibration CNN for EEG-Based Emotion Recognition.

Electroencephalography (EEG)-based emotion computing has become one of the research hotspots of human-computer interaction (HCI). However, it is difficult to effectively learn the interactions between brain regions in emotional states by using traditional convolutional neural networks because there is information transmission between neurons, which constitutes the brain network structure. In this paper, we proposed a novel model combining graph convolutional network and convolutional neural network, namely MDGCN-SRCNN, aiming to fully extract features of channel connectivity in different receptive fields and deep layer abstract features to distinguish different emotions. Particularly, we add style-based recalibration module to CNN to extract deep layer features, which can better select features that are highly related to emotion. We conducted two individual experiments on SEED data set and SEED-IV data set, respectively, and the experiments proved the effectiveness of MDGCN-SRCNN model. The recognition accuracy on SEED and SEED-IV is 95.08 and 85.52%, respectively. Our model has better performance than other state-of-art methods. In addition, by visualizing the distribution of different layers features, we prove that the combination of shallow layer and deep layer features can effectively improve the recognition performance. Finally, we verified the important brain regions and the connection relationships between channels for emotion generation by analyzing the connection weights between channels after model learning.

Data Augmentation for EEG-Based Emotion Recognition Using Generative Adversarial Networks.

One of the greatest limitations in the field of EEG-based emotion recognition is the lack of training samples, which makes it difficult to establish effective models for emotion recognition. Inspired by the excellent achievements of generative models in image processing, we propose a data augmentation model named VAE-D2GAN for EEG-based emotion recognition using a generative adversarial network. EEG features representing different emotions are extracted as topological maps of differential entropy (DE) under five classical frequency bands. The proposed model is designed to learn the distributions of these features for real EEG signals and generate artificial samples for training. The variational auto-encoder (VAE) architecture can learn the spatial distribution of the actual data through a latent vector, and is introduced into the dual discriminator GAN to improve the diversity of the generated artificial samples. To evaluate the performance of this model, we conduct a systematic test on two public emotion EEG datasets, the SEED and the SEED-IV. The obtained recognition accuracy of the method using data augmentation shows as 92.5 and 82.3%, respectively, on the SEED and SEED-IV datasets, which is 1.5 and 3.5% higher than that of methods without using data augmentation. The experimental results show that the artificial samples generated by our model can effectively enhance the performance of the EEG-based emotion recognition.

Visualization of microRNA-21 Dynamics in Neuroblastoma Using Magnetic Resonance Imaging Based on a microRNA-21-Responsive Reporter Gene.

BACKGROUND: MicroRNAs (miRs) have been shown to be closely associated with the occurrence and development of tumors and to have potential as diagnostic and therapeutic targets. The detection of miRs by noninvasive imaging technology is crucial for deeply understanding their biological functions. Our aim was to develop a novel miR-21-responsive gene reporter system for magnetic resonance imaging (MRI) visualization of the miR-21 dynamics in neuroblastoma. METHODS: The reporter gene ferritin heavy chain (FTH1) was modified by the addition of 3 copies of the sequence completely complementary to miR-21 (3xC_miR-21) to its 3'-untranslated region (3' UTR) and transduced into SK-N-SH cells to obtain SK-N-SH/FTH1-3xC_miR-21 cells. Then, the antagomiR-21 was delivered into cells by graphene oxide functionalized with polyethylene glycol and dendrimer. Before and after antagomiR-21 delivery, FTH1 expression, MRI contrast and intracellular iron uptake were assayed in vitro and in vivo. RESULTS: In the SK-N-SH/FTH1-3xC_miR-21 cells, FTH1 expression was in an "off" state due to the combination of intratumoral miR-21 with the 3' UTR of the reporter gene. AntagomiR-21 delivered into the cells bound to miR-21 and thereby released it from the 3' UTR of the reporter gene, thus "switching on" FTH1 expression in a dose-dependent manner. This phenomenon resulted in intracellular iron accumulation and allowed MRI detection in vitro and in vivo. CONCLUSION: MRI based on the miR-21-responsive gene reporter may be a potential method for visualization of the endogenous miR-21 activity in neuroblastoma and its response to gene therapy.

P3-MSDA: Multi-Source Domain Adaptation Network for Dynamic Visual Target Detection.

Single-trial electroencephalogram detection has been widely applied in brain-computer interface (BCI) systems. Moreover, an individual generalized model is significant for applying the dynamic visual target detection BCI system in real life because of the time jitter of the detection latency, the dynamics and complexity of visual background. Hence, we developed an unsupervised multi-source domain adaptation network (P3-MSDA) for dynamic visual target detection. In this network, a P3 map-clustering method was proposed for source domain selection. The adversarial domain adaptation was conducted for domain alignment to eliminate individual differences, and prediction probabilities were ranked and returned to guide the input of target samples for imbalanced data classification. The results showed that individuals with a strong P3 map selected by the proposed P3 map-clustering method perform best on the source domain. Compared with existing schemes, the proposed P3-MSDA network achieved the highest classification accuracy and F1 score using five labeled individuals with a strong P3 map as the source domain. These findings can have a significant meaning in building an individual generalized model for dynamic visual target detection.

MRI detection of the malignant transformation of stem cells through reporter gene expression driven by a tumor-specific promoter.

BACKGROUND: Existing evidence has shown that mesenchymal stem cells (MSCs) can undergo malignant transformation, which is a serious limitation of MSC-based therapies. Therefore, it is necessary to monitor malignant transformation of MSCs via a noninvasive imaging method. Although reporter gene-based magnetic resonance imaging (MRI) has been successfully applied to longitudinally monitor MSCs, this technique cannot distinguish the cells before and after malignant transformation. Herein, we investigated the feasibility of using a tumor-specific promoter to drive reporter gene expression for MRI detection of the malignant transformation of MSCs. METHODS: The reporter gene ferritin heavy chain (FTH1) was modified by adding a promoter from the tumor-specific gene progression elevated gene-3 (PEG3) and transduced into MSCs to obtain MSCs-PEG3-FTH1. Cells were induced to undergo malignant transformation via indirect coculture with C6 glioma cells, and these transformed cells were named MTMSCs-PEG3-FTH1. Western blot analysis of FTH1 expression, Prussian blue staining and transmission electron microscopy (TEM) to detect intracellular iron, and MRI to detect signal changes were performed before and after malignant transformation. Then, the cells before and after malignant transformation were inoculated subcutaneously into nude mice, and MRI was performed to observe the signal changes in the xenografts. RESULTS: After induction of malignant transformation, MTMSCs demonstrated tumor-like features in morphology, proliferation, migration, and invasion. FTH1 expression was significantly increased in MTMSCs-PEG3-FTH1 compared with MSCs-PEG3-FTH1. Prussian blue staining and TEM showed a large amount of iron particles in MTMSCs-PEG3-FTH1 but a minimal amount in MSCs-PEG3-FTH1. MRI demonstrated that the T2 value was significantly decreased in MTMSCs-PEG3-FTH1 compared with MSCs-PEG3-FTH1. In vivo, mass formation was observed in the MTMSCs-PEG3-FTH1 group but not the MSCs-PEG3-FTH1 group. T2-weighted MRI showed a significant signal decrease, which was correlated with iron accumulation in the tissue mass. CONCLUSIONS: We developed a novel MRI model based on FTH1 reporter gene expression driven by the tumor-specific PEG3 promoter. This approach could be applied to sensitively detect the occurrence of MSC malignant transformation.

A Sleeping rs-fMRI Study of Preschool Children with Autism Spectrum Disorders.

OBJECTIVES: The brain functional network of autism spectrum disorders (ASDs) in the earlier stages of life has been almost unknown due to difficulties in obtaining a resting-state functional magnetic resonance imaging (rs-fMRI). This study aimed to perform rs-MRI under a sedated sleep state and reveal possible alterations in the brain functional network. METHODS: Rs-fMRI was performed in a group of preschool children (aged 2-6 years, 53 with ASD, 63 as controls) under a sedated sleeping state. Based on graph theoretical analysis, global and local topological metrics were calculated to investigate alterations in brain functional networks. Besides, correlation analyses were conducted between the abnormal attribute values and the Childhood Autism Rating Scale (CARS) scores. RESULTS: The graph theoretical analysis showed that the nodal degree of the right medial frontal gyrus and the nodal efficiency of the right lingual gyrus in the ASD group were higher than those in the control group (P<0.05). There was a statistically significant positive correlation (R=0.318, P<0.05) between the right midfrontal gyrus nodal degree values and CARS scores in the ASD patients. CONCLUSION: Alterations of some nodal attributes in the brain network occurred in preschool autistic children which could serve as potential imaging biomarkers for evaluating ASD in earlier stages.

Two-Level Domain Adaptation Neural Network for EEG-Based Emotion Recognition.

Emotion recognition plays an important part in human-computer interaction (HCI). Currently, the main challenge in electroencephalogram (EEG)-based emotion recognition is the non-stationarity of EEG signals, which causes performance of the trained model decreasing over time. In this paper, we propose a two-level domain adaptation neural network (TDANN) to construct a transfer model for EEG-based emotion recognition. Specifically, deep features from the topological graph, which preserve topological information from EEG signals, are extracted using a deep neural network. These features are then passed through TDANN for two-level domain confusion. The first level uses the maximum mean discrepancy (MMD) to reduce the distribution discrepancy of deep features between source domain and target domain, and the second uses the domain adversarial neural network (DANN) to force the deep features closer to their corresponding class centers. We evaluated the domain-transfer performance of the model on both our self-built data set and the public data set SEED. In the cross-day transfer experiment, the ability to accurately discriminate joy from other emotions was high: sadness (84%), anger (87.04%), and fear (85.32%) on the self-built data set. The accuracy reached 74.93% on the SEED data set. In the cross-subject transfer experiment, the ability to accurately discriminate joy from other emotions was equally high: sadness (83.79%), anger (84.13%), and fear (81.72%) on the self-built data set. The average accuracy reached 87.9% on the SEED data set, which was higher than WGAN-DA. The experimental results demonstrate that the proposed TDANN can effectively handle the domain transfer problem in EEG-based emotion recognition.