Search

Finding the biomarkers associated with autism spectrum disorder (ASD) is helpful for understanding the underlying roots of the disorder and can lead to earlier diagnosis and more targeted treatments. In essence, we are faced with two challenges (i) how to learn a node representation and a clean graph structure from original graph data with high dimensionality and (ii) how to jointly model the procedure of node representation learning, structure learning and graph classification.We propose FSL-BrainNet, an interpretable graph convolution network (GCN) model for jointly Learning of node Features and clean Structures in brain networks for automatic brain network classification and interpretation. We formulate an end-to-end trainable and interpretable framework for graph classification and biomarkers (salient brain regions and potential subnetworks) identification.The experimental results on the ABIDE dataset show that our proposed methods not only achieve improved prediction performance compared with the state-of-the-art methods, but also find a compact set of highly suggestive biomarkers including relevant brain regions and subnetworks to ASD.Through node feature learning and structure learning, our model can simultaneously select important brain regions and identify subnetworks.

Alzheimer's disease (AD) is highly prevalent and a significant cause of dementia and death in elderly individuals. Motivated by breakthroughs of multi-task learning (MTL), efforts have been made to extend MTL to improve the Alzheimer's disease cognitive score prediction by exploiting structure correlation. Though important and well-studied, three key aspects are yet to be fully handled in an unified framework: (i) appropriately modeling the inherent task relationship; (ii) fully exploiting the task relatedness by considering the underlying feature structure. (iii) automatically determining the weight of each task. To this end, we present the Bi-Graph guided self-Paced Multi-Task Feature Learning (BGP-MTFL) framework for exploring the relationship among multiple tasks to improve overall learning performance of cognitive score prediction. The framework consists of the two correlation regularization for features and tasks, ℓ

Accurate diagnosis of autism spectrum disorder (ASD) plays a key role in improving the condition and quality of life for patients. In this study, we mainly focus on ASD diagnosis with functional brain networks (FBNs). The major challenge for brain networks modeling is the high dimensional connectivity in brain networks and limited number of subjects, which hinders the classification capability of graph convolutional networks (GCNs).To alleviate the influence of the limited data and high dimensional connectivity, we introduce a unified three-stage graph learning framework for brain network classification, involving multi-graph clustering, graph generation and graph classification. The framework combining Graph Generation, Clustering and Classification Networks (GraphCGC-Net) enhances the critical connections by multi-graph clustering (MGC) with a supervision scheme, and generates realistic brain networks by simultaneously preserving the global consistent distribution and local topology properties.To demonstrate the effectiveness of our approach, we evaluate the performance of the proposed method on the Autism Brain Imaging Data Exchange (ABIDE) dataset and conduct extensive experiments on the ASD classification problem. Our proposed method achieves an average accuracy of 70.45% and an AUC of 72.76% on ABIDE. Compared with the traditional GCN model, the proposed GraphCGC-Net obtains 9.3%, and 10.64% improvement in terms of accuracy and AUC metrics, respectively.The comprehensive experiments demonstrate that our GraphCGC-Net is effective for graph classification in brain disorders diagnosis. Moreover, we find that MGC can generate biologically meaningful subnetworks, which is highly consistent with the previous neuroimaging-derived biomarker evidence of ASD. More importantly, the promising results suggest that applying generative adversarial networks (GANs) in brain networks to improve the classification performance is worth further investigation.

Coronary artery segmentation in coronary computed tomography angiography (CTA) images plays a crucial role in diagnosing cardiovascular diseases. However, due to the complexity of coronary CTA images and coronary structure, it is difficult to automatically segment coronary arteries accurately and efficiently from numerous coronary CTA images.In this study, an automatic method based on symmetrical radiation filter (SRF) and D-means is presented. The SRF, which is applied to the three orthogonal planes, is designed to filter the suspicious vessel tissue according to the features of gradient changes on vascular boundaries to segment coronary arteries accurately and reduce computational cost. Additionally, the D-means local clustering is proposed to be embedded into vessel segmentation to eliminate noise impact in coronary CTA images.The results of the proposed method were compared against the manual delineations in 210 coronary CTA data sets. The average values of true positive, false positive, Jaccard measure, and Dice coefficient were [Formula: see text], [Formula: see text], [Formula: see text], and [Formula: see text], respectively. Moreover, comparing the delineated data sets and public data sets showed that the proposed method is better than the related methods.The experimental results indicate that the proposed method can perform complete, robust, and accurate segmentation of coronary arteries with low computational cost. Therefore, the proposed method is proven effective in vessel segmentation of coronary CTA images without extensive training data and can meet clinical applications.

Complete left bundle branch block (cLBBB) is an electrical conduction disorder associated with cardiac disease. Septal flash (SF) involves septal leftward contraction during early systole followed by a lengthening motion toward the right ventricle and affects several patients with cLBBB. It has been revealed that cLBBB patients with SF may be at risk of cardiac function reduction and poor prognosis. Therefore, accurate identification of SF may play a vital role in counseling patients about their prognosis. Generally, Septal flash is identified by echocardiography using visual "eyeballing". However, this conventional method is subjective as it depends on operator experience. In this study, we build a linear attention cascaded net (LACNet) capable of processing echocardiography to identify SF automatically. The proposed method consists of a cascaded CNN-based encoder and an LSTM-based decoder, which extract spatial and temporal features simultaneously. A spatial transformer network (STN) module is employed to avoid image inconsistency and linear attention layers are implemented to reduce data complexity. Moreover, the left ventricle (LV) area-time curve calculated from segmentation results can be considered as a new independent disease predictor as SF phenomenon leads to transient left ventricle area enlargement. Therefore, we added the left ventricle area-time curve to LACNet to enrich input data diversity. The result shows the possibility of using echocardiography to diagnose cLBBB with SF automatically.

One of the most laborious for drug discovery is to select compounds from a library for experimental evaluation. Hence, we propose a machine learning model only needs to be trained on a small dataset to predict the inhibition constant (Ki) and half maximal inhibitory concentration (IC50) for a compound. We transfer the prediction task to a simpler binary classification task based on a naive but effective idea that we only need the related rank of a compound to determine whether to take it for further examination. To achieve this, we design a data augmentation strategy to effectively leverage the relationship between the compounds in the training set. After that, we formulate a new reward function for deep reinforcement learning to balance the feature selection and the accuracy. We employ a particle swarm optimized support vector machine for the binary classification task. Finally, a soft voting mechanism is introduced to solve the contradiction from the binary classification. Sufficient experiments show that our model achieves high and reliable accuracy, and is capable of ranking compounds based on a selected set of molecular descriptors. The current results show that our model provides a potential ligand-based in silico approach for prioritizing chemicals for experimental studies.

Major depressive disorder (MDD) is a common mental illness characterized by persistent feeling of depressed mood and loss of interest. It would cause, in a severe case, suicide behaviors. In clinical settings, automatic MDD detection is mainly based on electroencephalogram (EEG) signals with supervised learning techniques. However, supervised-based MDD detection methods encounter two ineviTable bottlenecks: firstly, such methods rely heavily on an EEG training dataset with MDD labels annotated by a physical therapist, leading to subjectivity and high cost; secondly, most of EEG signals are unlabeled in a real scenario. In this paper, a novel semisupervised-based MDD detection method named MDD-TSVM is presented. Specifically, the MDD-TSVM utilizes the semisupervised method of transductive support vector machine (TSVM) as its backbone, further dividing the unlabeled penalty item of the TSVM objective function into two pseudo-labeled penalty items with or without MDD. By such improvement, the MDD-SVM can make full use of labeled and unlabeled datasets as well as alleviate the class imbalance problem. Experiment results showed that our proposed MDD-TSVM achieved F1 score of 0.85 ± 0.05 and accuracy of 0.89 ± 0.03 on identifying MDD patients, which is superior to the state-of-the-art methods.

Novel coronavirus pneumonia (NCP) has become a global pandemic disease, and computed tomography-based (CT) image analysis and recognition are one of the important tools for clinical diagnosis. In order to assist medical personnel to achieve an efficient and fast diagnosis of patients with new coronavirus pneumonia, this paper proposes an assisted diagnosis algorithm based on ensemble deep learning. The method combines the Stacked Generalization ensemble learning with the VGG16 deep learning to form a cascade classifier, and the information constituting the cascade classifier comes from multiple subsets of the training set, each of which is used to collect deviant information about the generalization behavior of the data set, such that this deviant information fills the cascade classifier. The algorithm was experimentally validated for classifying patients with novel coronavirus pneumonia, patients with common pneumonia (CP), and normal controls, and the algorithm achieved a prediction accuracy of 93.57%, sensitivity of 94.21%, specificity of 93.93%, precision of 89.40%, and F1-score of 91.74% for the three categories. The results show that the method proposed in this paper has good classification performance and can significantly improve the performance of deep neural networks for multicategory prediction tasks.

Purpose Recently, brain connectivity networks have been used for the classification of neurological disorder, such as Autism Spectrum Disorders (ASD) or Alzheimer's disease (AD). Network analysis provides a new way for exploring the association between brain functional deficits and the underlying structural disruption related to brain disorders. Network embedding learning that aims to automatically learn low-dimensional representations for brain networks has drawn increasing attention in recent years. Method In this work we build upon graph neural network in order to learn useful representations for graph classification in an end-to-end fashion. Specifically, we propose a hierarchical GCN framework (called hi-GCN) to learn the graph feature embedding while considering the network topology information and subject's association at the same time. Results To demonstrate the effectiveness of our approach, we evaluate the performance of the proposed method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset and Autism Brain Imaging Data Exchange (ABIDE) dataset. Extensive experiments on ABIDE and ADNI datasets have demonstrated competitive performance of the hi-GCN model. Specifically, we obtain an average accuracy of 73.1%/78.5% as well as AUC of 82.3%/86.5% on ABIDE/ADNI. The comprehensive experiments demonstrate that our hi-GCN is effective for graph classification with brain disorders diagnosis. Conclusion The proposed hi-GCN method performs the graph embedding learning from a hierarchical perspective while considering the structure in individual brain network and the subject's correlation in the global population network, which can capture the most essential embedding features to improve the classification performance of disease diagnosis. Moreover, the proposed jointly optimizing strategy also achieves faster training and easier convergence than both the hi-GCN with pre-training and two-step supervision.

Objective Diabetic retinopathy (DR) is one of the most serious complications of diabetes. Early detection and treatment of DR are key public health interventions that can significantly reduce the risk of vision loss. How to effectively screen and diagnose the retinal fundus image in order to identify retinopathy in time is a major challenge. In the traditional DR screening system, the accuracy of micro-aneurysm (MA) and hemorrhagic (H) lesion detection determines the final screening performance. The detection method produced a large number of false positive samples for guaranteeing high sensitivity, and the classification model was not effective in removing false positives since the suspicious lesions lack label information. Methods In order to solve the problem of supervised learning in the diagnosis of DR, we formulate weakly supervised multi-class DR grading as a multi-class multi-instance problem where each image (bag) is labeled as healthy or abnormal and consists of unlabeled candidate lesion regions (instances). Specifically, we proposed a multi-kernel multi-instance learning method based on graph kernel. Moreover, we develop an under-sampling from instance level and over-sampling from bag level to improve the performance of the multi-instance learning in the diagnosis of DR. Results Through empirical evaluation and comparison with different baselinemethods and the state-of-the-art methods on data from Messidor, we illustrate that the proposed method reports favorable results, with an overall classification accuracy of 0.916 and an AUC of 0.957. Conclusions The experiments results demonstrate that the proposed multi-kernel multi-instance learning framework with bi-level re-sampling can solve the problem in the imbalanced and weakly supervised data for grading diabetic retinopathy, and it improves the diagnosis performance over several state-of-the-art competing methods.

Objective Alzheimers disease (AD) is characterized by gradual neurodegeneration and loss of brain function, especially for memory during early stages. Regression analysis has been widely applied to AD research to relate clinical and biomarker data such as predicting cognitive outcomes from Magnetic Resonance Imaging (MRI) measures. Recently, the multi-task feature learning (MTFL) methods have been widely studied to predict cognitive outcomes and select the discriminative feature subset from MRI features by incorporating inherent correlations among multiple clinical cognitive measures. However, the existing MTFL assumes the correlation among all the tasks is uniform, and the task relatedness is modeled by encouraging a common subset of features with neglecting the inherent structure of tasks and MRI features. Methods In this paper, we proposed a generalized fused group lasso (GFGL) regularization to model the underlying structures, involving (1) a graph structure within tasks and (2) a group structure among the image features. Then, we present a multi-task learning framework (called GFGL–MTFL), combining the l2, 1-norm with the GFGL regularization, to model the flexible structures. Results Through empirical evaluation and comparison with different baseline methods and the state-of-the-art MTL methods on data from Alzheimer’s Disease Neuroimaging Initiative (ADNI) database, we illustrate that the proposed GFGL–MTFL method outperforms other methods in terms of both Mean Squared Error (nMSE) and weighted correlation coefficient (wR). Improvements are statistically significant for most scores (tasks). Conclusions The experimental results with real and synthetic data demonstrate that incorporating the two prior structures by the generalized fused group lasso norm into the multi task feature learning can improve the prediction performance over several state-of-the-art competing methods, and the estimated correlation of the cognitive functions and the identification of cognition relevant imaging markers are clinically and biologically meaningful.

Alzheimer's disease (AD) has been not only a substantial financial burden to the health care system but also an emotional burden to patients and their families. Making accurate diagnosis of AD based on brain magnetic resonance imaging (MRI) is becoming more and more critical and emphasized at the earliest stages. However, the high dimensionality and imbalanced data issues are two major challenges in the study of computer aided AD diagnosis. The greatest limitations of existing dimensionality reduction and over-sampling methods are that they assume a linear relationship between the MRI features (predictor) and the disease status (response). To better capture the complicated but more flexible relationship, we propose a multi-kernel based dimensionality reduction and over-sampling approaches. We combined Marginal Fisher Analysis with l2,1-norm based multi-kernel learning (MKMFA) to achieve the sparsity of region-of-interest (ROI), which leads to simultaneously selecting a subset of the relevant brain regions and learning a dimensionality transformation. Meanwhile, a multi-kernel over-sampling (MKOS) was developed to generate synthetic instances in the optimal kernel space induced by MKMFA, so as to compensate for the class imbalanced distribution. We comprehensively evaluate the proposed models for the diagnostic classification (binary class and multi-class classification) including all subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. The experimental results not only demonstrate the proposed method has superior performance over multiple comparable methods, but also identifies relevant imaging biomarkers that are consistent with prior medical knowledge.

Alzheimer's disease (AD) is a severe neurodegenerative disorder that usually starts slowly and progressively worsens. Predicting the progression of Alzheimer's disease with longitudinal analysis on the time series data has recently received increasing attention. However, training an accurate progression model for brain network faces two major challenges: missing features, and the small sample size during the follow-up study. According to our analysis on the AD progression task, we thoroughly analyze the correlation among the multiple predictive tasks of AD progression at multiple time points. Thus, we propose a multi-task learning framework that can adaptively impute missing values and predict future progression over time from a subject's historical measurements. Progression is measured in terms of MRI volumetric measurements, trajectories of a cognitive score and clinical status. To this end, we propose a new perspective of predicting the AD progression with a multi-task learning paradigm. In our multi-task learning paradigm, we hypothesize that the inherent correlations exist among: (i). the prediction tasks of clinical diagnosis, cognition and ventricular volume at each time point; (ii). the tasks of imputation and prediction; and (iii). the prediction tasks at multiple future time points. According to our findings of the task correlation, we develop an end-to-end deep multi-task learning method to jointly improve the performance of assigning missing value and prediction. We design a balanced multi-task dynamic weight optimization. With in-depth analysis and empirical evidence on Alzheimer's Disease Neuroimaging Initiative (ADNI), we show the benefits and flexibility of the proposed multi-task learning model, especially for the prediction at the M60 time point. The proposed approach achieves 5.6%, 5.7%, 4.0% and 11.8% improvement with respect to mAUC, BCA and MAE (ADAS-Cog13 and Ventricles), respectively.

The cardiac Left Ventricle (LV) segmentation is still challengeable due to the complex anatomical structure surrounding the LV and intensity overlaps caused by intensity inhomogeneities, which widely exists in cardiac Magnetic Resonance (MR) images. To overcome these problems, a shape prior constrained two-layer level set method is proposed to simultaneously extract the endocardium and epicardium of the LV from cardiac short-axis MR images. The proposed method is validated on MR images from the LV challenge of MICCAI 2009. Quantitative and qualitative experiments and clinical metrics demonstrate that the proposed method has better performance than some representative methods in term of Dice Similarity Coefficient (DSC), Average Perpendicular Distance (APD), Ejection Fraction (EF), and LV mass. Experimental results demonstrate that the proposed method has the priority of preserving cardiac anatomical structure and improving segmentation accuracy.

BACKGROUND Differential demon is a fast and efficient registration algorithm. It drives the floating image to deform using the force based on the gradient between the reference and floating image. But it will cause abnormal deformation when the driving force approaches zero,which limits its practical applications. OBJECTIVE This paper proposed an improved differential demon algorithm, which aimed to enhance the registration performance of the existing demon algorithm. METHODS Firstly, we review the original differential demon algorithm. Then, we propose an improved differential demon algorithm and the process of mathematical deduction. Finally, we use experiment to prove that the improved differential demon algorithm is effective and it can improve the accuracy of registration. RESULTS We tested our method on data sets provided by Xuanwu Hospital Capital Medical University. The registration performance proved to be better than the original demon algorithm in terms of mutual information, normalized correlation coefficient, mean square error and iteration number. CONCLUSIONS Experiment results demonstrate the superiority of method proposed in this paper to the original demon algorithm.

BACKGROUND Existing methods may fail to locate and segment the optic disc (OD) due to imprecise boundaries, inconsistent image contrast and deceptive edge features in retinal images. OBJECTIVE To locate the OD and detect the OD boundary accurately. METHODS The method exploits a multi-stage strategy in the detection procedure. Firstly, OD location candidate regions are identified based on high-intensity feature and vessels convergence property. Secondly, a line operator filter for circular brightness feature detection is designed to locate the OD accurately on candidates. Thirdly, an initialized contour is obtained by iterative thresholding and ellipse fitting based on the detected OD position. Finally, a region-based active contour model in a variational level set formulation and ellipse fitting are employed to estimate the OD boundary. RESULTS The proposed methodology achieves an accuracy of 98.67% for OD identification and a mean distance to the closest point of 2 pixels in detecting the OD boundary. CONCLUSION The results illuminate that the proposed method is effective in the fast, automatic, and accurate localization and boundary detection of the OD. The present work contributes to the more effective evaluation of the OD and realizing automatic screening system for early eye diseases to a large extent.

Vessel segmentation has always been a considerable challenge task due to the presence of varying thickness of the vessels and weak contrasts of medical image intensities. In this paper, an effective method is proposed, which consists of four steps. Firstly, the input images are converted into gray ones with predetermined weightings aiming at increasing image contrast if they are colorful. Secondly, the image intensities are expanded from regions of interest to the whole image domain with a mirroring operation to avoid introducing undesired boundaries by image filtering operations existing in the next step. Thirdly, an improved multi-scale enhancement method inspired by the Frangi filtering is proposed to enhance image contrast between blood vessels and other objects in the image. Finally, an improved level set model is proposed to segment blood vessels from the enhance images and the original gray images. The proposed method has been evaluated on two retinal vessel image repositories, namely, DRIVE and STARE. Experimental results and comparison with 13 existing methods show that the proposed method produces similar segmentation accuracy at the same level with representative methods in the literature. Its effectiveness on segmentation of other type vessels is also discussed at the end of this paper.

Classification plays a critical role in false positive reduction (FPR) in lung nodule computer aided detection (CAD). The difficulty of FPR lies in the variation of the appearances of the nodules, and the imbalance distribution between the nodule and non-nodule class. Moreover, the presence of inherent complex structures in data distribution, such as within-class imbalance and high-dimensionality are other critical factors of decreasing classification performance. To solve these challenges, we proposed a hybrid probabilistic sampling combined with diverse random subspace ensemble. Experimental results demonstrate the effectiveness of the proposed method in terms of geometric mean (G-mean) and area under the ROC curve (AUC) compared with commonly used methods.