Search

Hippocampus is an important anatomical region for Alzheimer's Disease (AD) identification. In this paper, a multi-scale attention-based convolutional network is proposed for AD identification. The two dimensional (2D) images in three different planes of hippocampal subfields are used as input of three branches of the proposed network, which achieves effective extraction of three dimensional (3D) data features while reducing the network complexity and improving the computational efficiency. The end-to-end 2D multi-scale attention-based deep learning network improves the diversity of the extracted features and captures significance of various voxels for classification, which achieves significant classification performance without handcrafted feature extraction and model stacking. Experimental results illustrate the effectiveness of the proposed method on AD identification. The proposed method will be useful for further medical analysis on hippocampal subfields of the brain for diagnosis of neurodegenerative disease.

The present study aimed to explore the effect of computerized multi-domain cognitive training (MDCT) on brain gray matter volume and neuropsychological performance in patients with amnestic mild cognitive impairment (amnestic MCI). Twenty-one patients with amnestic MCI participated in a computerized MDCT program. The program targeted a broad set of cognitive domains via programs focused on reasoning, memory, visuospatial, language, calculation, and attention. Seventeen Participants completed the intervention and all completed a battery of neuropsychological tests to evaluate cognitive function while 12 out of 17 underwent 3 T MRI scanning before and after the intervention to measure gray matter (GM) volume. We examined correlations between the changes in neuropsychological scores and GM volumes across participants after the intervention. After training, we observed significant increases in GM volume in the right angular gyrus (AG) and other parietal subareas near the intraparietal sulcus (p p > 0.05). A correlation analysis revealed positive correlations between the changes in GM volume in the right AG and scores in the immediate recall component of the Hopkins Verbal Learning Test-Revised (HVLT-R) (r = 0.64, p = 0.024) and the Brief Visuospatial Memory Test–Revised (BVMT-R) (r = 0.67, p = 0.016). Our findings indicate that a computerized MDCT program may protect patients with amnestic MCI against brain GM volume loss and has potential in preserving general cognition. Thus, our non-pharmacological intervention may slow the rate of disease progression.

Convolutional neural networks (CNNs) have become an increasingly popular tool for brain lesion segmentation in recent years due to its accuracy and efficiency. However, CNN-based brain lesion segmentation generally requires a large amount of annotated training data, which can be costly for medical imaging. In many scenarios, only a few annotations of brain lesions are available. One common strategy to address the issue of limited annotated data is to transfer knowledge from a different yet relevant source task, where training data is abundant, to the target task of interest. Typically, a model can be pretrained for the source task, and then fine-tuned with the scarce training data associated with the target task. However, classic fine-tuning tends to make small modifications to the pretrained model, which could hinder its adaptation to the target task. Fine-tuning with increased model capacity has been shown to alleviate this negative impact in image classification problems. In this work, we extend the strategy of fine-tuning with increased model capacity to the problem of brain lesion segmentation, and then develop an advanced version that is better suitable for segmentation problems. First, we propose a vanilla strategy of increasing the capacity, where, like in the classification problem, the width of the network is augmented during fine-tuning. Second, because unlike image classification, in segmentation problems each voxel is associated with a labeling result, we further develop a spatially adaptive augmentation strategy during fine-tuning. Specifically, in addition to the vanilla width augmentation, we incorporate a module that computes a spatial map of the contribution of the information given by width augmentation in the final segmentation. For demonstration, the proposed method was applied to ischemic stroke lesion segmentation, where a model pretrained for brain tumor segmentation was fine-tuned, and the experimental results indicate the benefit of our method.

Alzheimer's Disease (AD) is a progressive irreversible neurodegeneration disease and thus timely identification is critical to delay its progression.In this work, we focus on the traditional branch to design discriminative feature extraction and selection strategies to achieve explainable AD identification. Specifically, a spatial pyramid based three-dimensional histogram of oriented gradient (3D-HOG) feature learning method is proposed. Both global and local texture changes are included in spatial pyramid 3D-HOG (SPHOG) features for comprehensive analysis. Then a modified wrapper-based feature selection algorithm is introduced to select the discriminative features for AD identification while reduce feature dimensions.Discriminative SPHOG histograms with various resolutions are selected, which can represent the atrophy characteristics of cerebral cortex with promising performance. As subareas corresponding to selected histograms are consistent with clinical experience, explanatory is emphasized and illustrated with Hippocampus.Experimental results illustrate the effectiveness of the proposed method on feature learning based on samples obtained from common dataset and a clinical dataset. The proposed method will be useful for further medical analysis as its explanatory on other region-of-interests (ROIs) of the brain for early diagnosis of AD.

Deep learning based methods have improved the estimation of tissue microstructure from diffusion magnetic resonance imaging (dMRI) scans acquired with a reduced number of diffusion gradients. These methods learn the mapping from diffusion signals in a voxel or patch to tissue microstructure measures. In particular, it is beneficial to exploit the sparsity of diffusion signals jointly in the spatial and angular domains, and the deep network can be designed by unfolding iterative processes that adaptively incorporate historical information for sparse reconstruction. However, the number of network parameters is huge in such a network design, which could increase the difficulty of network training and limit the estimation performance. In addition, existing deep learning based approaches to tissue microstructure estimation do not provide the important information about the uncertainty of estimates. In this work, we continue the exploration of tissue microstructure estimation using a deep network and seek to address these limitations. First, we explore the sparse representation of diffusion signals with a separable spatial-angular dictionary and design an improved deep network for tissue microstructure estimation. The procedure for updating the sparse code associated with the separable dictionary is derived and unfolded to construct the deep network. Second, with the formulation of sparse representation of diffusion signals, we propose to quantify the uncertainty of network outputs with a residual bootstrap strategy. Specifically, because of the sparsity constraint in the signal representation, we perform a Lasso bootstrap strategy for uncertainty quantification. Experiments were performed on brain dMRI scans with a reduced number of diffusion gradients, where the proposed method was applied to two representative biophysical models for describing tissue microstructure and compared with state-of-the-art methods of tissue microstructure estimation. The results show that our approach compares favorably with the competing methods in terms of estimation accuracy. In addition, the uncertainty measures provided by our method correlate with estimation errors and produce reasonable confidence intervals; these results suggest potential application of the proposed uncertainty quantification method in brain studies.

In this paper, we compare the performance of the derived anatomical features and the extracted feature parameters in Alzheimer's disease (AD) identification. The correlation relationship between them and clinical mini-mental state examination (MMSE) score is analyzed. Based on these feature parameters, the highly correlated combined feature vectors are built and used as variables for the presented modified elastic net (EN) classifier. Experimental results show that the extracted feature parameters can obtain similar identification performance with the cortical thickness and the volume of gray matter (GM) in AD identification. The highly correlated combined feature vectors show the best identification performance among all of feature parameters using the modified EN-based classifier.

Diffusion magnetic resonance imaging (dMRI) provides a noninvasive method for measuring brain tissue microstructure. q -Space deep learning( q -DL) methods have been developed to accurately estimate tissue microstructure from dMRI scans acquired with a reduced number of diffusion gradients. In these methods, deep networks are trained to learn the mapping directly from diffusion signals to tissue microstructure. However, the quality of tissue microstructure estimation can be limited not only by the reduced number of diffusion gradients but also by the low spatial resolution of typical dMRI acquisitions. Therefore, in this work we extend q -DL to super-resolved tissue microstructure estimation and propose super-resolved q -DL (SR- q -DL), where deep networks are designed to map low-resolution diffusion signals undersampled in the q -space to high-resolution tissue microstructure. Specifically, we use a patch-based strategy, where a deep network takes low-resolution patches of diffusion signals as input and outputs high-resolution tissue microstructure patches. The high-resolution patches are then combined to obtain the final high-resolution tissue microstructure map. Motivated by existing q -DL methods, we integrate the sparsity of diffusion signals in the network design, which comprises two functional components. The first component computes sparse representation of diffusion signals for the low-resolution input patch, and the second component maps the low-resolution sparse representation to high-resolution tissue microstructure. The weights in the two components are learned jointly and the trained network performs end-to-end tissue microstructure estimation. In addition to SR- q -DL, we further propose probabilistic SR- q -DL, which can quantify the uncertainty of the network output as well as achieve improved estimation accuracy. In probabilistic SR- q -DL, a deep ensemble strategy is used. Specifically, the deep network for SR- q -DL is revised to produce not only tissue microstructure estimates but also the uncertainty of the estimates. Then, multiple deep networks are trained and their results are fused for the final prediction of high-resolution tissue microstructure and uncertainty quantification. The proposed method was evaluated on two independent datasets of brain dMRI scans. Results indicate that our approach outperforms competing methods in terms of estimation accuracy. In addition, uncertainty measures provided by our method correlate with estimation errors, which indicates potential application of the proposed uncertainty quantification method in brain studies.

Background and Objective In recent years, some clinical parameters, such as the volume of gray matter (GM) and cortical thickness, have been used as anatomical features to identify Alzheimer’s disease (AD) from Healthy Controls (HC) in some feature-based machine learning methods. However, fewer image-based feature parameters have been proposed, which are equivalent to these clinical parameters, to describe the atrophy of regions-of-interest (ROIs) of the brain. In this study, we aim to extract effective image-based feature parameters to improve the diagnostic performance of AD with magnetic resonance imaging (MRI) data. Methods A new subspace-based sparse feature learning method is proposed, which builds a union-of-subspace representation model to realize feature extraction and disease identification. Specifically, the proposed method estimates feature dimensions reasonably, at the same time, it protects local features for the specified ROIs of the brain, and realizes image-based feature extraction and classification automatically instead of computing the volume of GM or cortical thickness preliminarily. Results Experimental results illustrate the effectiveness and robustness of the proposed method on feature extraction and classification, which are based on the sampled clinical dataset from Peking University Third Hospital of China and the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset. The extracted image-based feature parameters describe the atrophy of ROIs of the brain well as clinical parameters but show better performance in AD identification than clinical parameters. Based on them, the important ROIs for AD identification can be identified even for correlated variables. Conclusion The extracted features and the proposed identification parameters show high correlation with the volume of GM and the clinical mini-mental state examination (MMSE) score respectively. The proposed method will be useful in denoting the changes of cerebral pathology and cognitive function in AD patients.

Purpose To evaluate cerebral blood flow (CBF) in patients with Alzheimer's disease (AD) using a three-dimensional pseudocontinuous arterial spin labeling (PCASL). We aimed to study the effects of different post label delay on the resulting CBF maps and to investigate the characteristics and clinical applications of brain perfusion. Materials and methods Sixteen AD patients and nineteen healthy control subjects were recruited. 3D PCASL was performed using a 3.0 T MR scanner. ASL was performed twice with different post label delays (PLD). Comparisons of CBF were made between AD patients and healthy control subjects respectively with PLD of 1.5 s and PLD of 2.5 s. Relationship between the CBF values and cognition was investigated using correlation analysis. A receiver operating characteristic (ROC) curve was generated for CBF measurements in posterior cingulate region. Result AD patients with PLD of 1.5 s showed lower CBF values primarily in bilateral temporal lobes, precuneus, middle and posterior cingulate gyri, left inferior parietal gyrus, left angular gyrus and left superior frontal gyrus. Lowered cerebral values were also observed in similar regions with PLD of 2.5 s, but the clusters of voxel were smaller. CBF values were associated with cognition scores in most of gyri mentioned above. Conclusion Hypoperfusion areas were observed in AD patients. PLD of 1.5 s was sufficient to display CBF. Considering the complicated AD pathology, multiple PLDs are strongly recommended.

Magnetic resonance imaging (MRI) has been successfully applied to investigate neuron pathological changes. Since the high dimension of observation data, sparse feature learning plays an important role in overcoming the challenge of high variable dimension and low sample size problem among the disease identification. In this paper, sparse Elastic Net (EN) was used to extract low dimension features and to identify the Alzheimer's disease (AD). Compare with principal component analysis (PCA) method, the EN method can solve the problems of less samples and high correlations between variables. For those variables sharing the same biological phenomenon, it selected whole groups into the model automatically once one variable among them was selected. Unlike other subspace learning methods, the proposed method used less man-made feature setting. The problems of dimension reduction and classification were conducted into a similar formulation. Experimental results illustrated the effectiveness of the proposed method.

The authors prospectively compared single dose (0.1 mmol/kg bodyweight) gadobenate dimeglumine with double dose (0.2 mmol/kg bodyweight) gadopentetate dimeglumine for contrast-enhanced magnetic resonance angiography (CE-MRA) in patients with suspected or known steno-occlusive disease of the carotid, renal or peripheral vasculature using an intra-individual crossover study design. Twenty-eight patients with suspected or known steno-occlusive disease of the carotid (n = 16), renal (n = 5) or peripheral arteries (n = 7) were randomised to receive either 0.1 mmol/kg gadobenate dimeglumine or 0.2 mmol/kg gadopentetate dimeglumine for a first CE-MRA procedure. After 3–5 days all patients underwent a second identical CE-MRA procedure with the other contrast agent. Three blinded readers assessed images for vessel anatomical delineation, disease detection/exclusion, and global preference. Diagnostic performance for detection of ≥51 % stenosis was determined for 20/28 patients who also underwent digital subtraction angiography (DSA). Non-inferiority was assessed using the Wilcoxon signed rank, McNemar and Wald tests. Quantitative (signal-to-noise and contrast-to-noise ratio) enhancement based on 3D maximum intensity projection reconstructions was compared. No differences were noted for any qualitative parameter. Equivalence was reported for all diagnostic preference end-points. Superiority for gadobenate dimeglumine was reported by all readers for sensitivity for disease detection (80.8–86.5 vs. 75.0–82.7 %). Quantitative enhancement was similar for single dose gadobenate dimeglumine and double dose gadopentetate dimeglumine. Under identical examination conditions a single 0.1 mmol/kg body weight dose of gadobenate dimeglumine can fully replace a double 0.2 mmol/kg body weight dose of gadopentetate dimeglumine for routine CE-MRA procedures.

Cerebral microbleeds (CMBs) are bleeding events associated with cerebral small vessel disease (SVD). Strictly lobar CMBs and strictly deep CMBs are likely caused by cerebral amyloid angiopathy (CAA) and hypertensive arteriopathy, respectively. Leukoaraiosis (LA) reflects an ischaemic change in SVD, and LA severity has been correlated with CMBs. However, whether different locations (aetiologies) of CMBs correlate with LA is unknown.Patients receiving brain MRI and tbl2*-weighted gradient-recalled echo scans in a stroke outpatient department were screened for CMBs. The MRI results of the patients with CMBs were sent to investigators for further review and were evaluated using the Microbleed Anatomical Rating Scale. Cerebral microbleed severity was graded using a numerical scale. Leukoaraiosis severity was assessed using the Fazekas scale.Cerebral microbleeds were observed in 14.6% of the 1289 screened patients. The CMB incidence increased with age (in years, 50: 1.3%; 50-59: 10.7%; 60-69: 17.6% and ≥ 70: 23.6%; P = 0.000). The CMB locations were distributed as follows: 23.4% strictly lobar, 12.2% strictly deep, 6.4% strictly infratentorial and 58.0% mixed. Cerebral microbleed severity correlated with the total Fazekas scale score. The numbers of lobar, deep and infratentorial CMBs correlated with the total Fazekas scale score. The mixed CMB group displayed a significantly higher Fazekas scale score than the strictly lobar CMB group (P = 0.000).Cerebral microbleed incidence increased with age. Mixed CMB type displayed the highest incidence. The severity and number of CMBs at any location correlated with LA severity. These results suggested interactions between hypertension and CAA during LA progression.