Detection of Alzheimer's disease using features of brain region-of-interest-based individual network constructed with the sMRI image

Ranking convolutional neural network for Alzheimer's disease mini-mental state examination prediction at multiple time-points.

Alzheimer’s disease (AD) is the most common form of dementia, causing progressive cognitive decline. Radiomic features obtained from structural magnetic resonance imaging (sMRI) have shown a great potential in predicting this disease. However, radiomic features based on the whole brain segmented regions have not been explored yet. In our study, we collected sMRI data that include 80 patients with AD and 80 healthy controls (HCs). For each patient, the T1 weighted image (T1WI) images were segmented into 106 subregions, and radiomic features were extracted from each subregion. Then, we analyzed the radiomic features of specific brain subregions that were most related to AD. Based on the selective radiomic features from specific brain subregions, we built an integrated model using the best machine learning algorithms, and the diagnostic accuracy was evaluated. The subregions most relevant to AD included the hippocampus, the inferior parietal lobe, the precuneus, and the lateral occipital gyrus. These subregions exhibited several important radiomic features that include shape, gray level size zone matrix (GLSZM), and gray level dependence matrix (GLDM), among others. Based on the comparison among different algorithms, we constructed the best model using the Logistic regression (LR) algorithm, which reached an accuracy of 0.962. Conclusively, we constructed an excellent model based on radiomic features from several specific AD-related subregions, which could give a potential biomarker for predicting AD.

/pdf/investigation-of-underlying-association-between-whole-brain-39k42wtq.pdf

Investigation of Underlying Association Between Whole Brain Regions and Alzheimer’s Disease: A Research Based on an Artificial Intelligence Model

Imbalanced learning problems often occur in application scenarios and are additionally an important research direction in the field of machine learning. Traditional classifiers are substantially less effective for datasets with an imbalanced distribution, especially for high-dimensional longitudinal data structures. In the medical field, the imbalance of data problem is more common, and correctly identifying samples of the minority class can obtain important information. Moreover, class imbalance in imbalanced AD (Alzheimer’s disease) data presents a significant challenge for machine learning algorithms that assume the data are evenly distributed within the classes. In this paper, we propose a random forest-based feature selection algorithm for imbalanced neuroimaging data classification. The algorithm employs random forest to evaluate the value of each feature and combines the correlation matrix to choose the optimal feature subset, which is applied to imbalanced MRI (magnetic resonance imaging) AD data to identify AD, MCI (mild cognitive impairment), and NC (normal individuals). In addition, we extract multiple features from AD images that can represent 2D and 3D brain information. The effectiveness of the proposed method is verified by the experimental evaluation using the public ADNI (Alzheimer’s neuroimaging initiative) dataset, and results demonstrate that the proposed method has a higher prediction accuracy and AUC (area under the receiver operating characteristic curve) value in NC-AD, MCI-AD, and NC-MCI group data, with the highest accuracy and AUC value for the NC-AD group data.

/pdf/enhancing-feature-selection-for-imbalanced-alzheimers-281r3bwc.pdf

Enhancing Feature Selection for Imbalanced Alzheimer’s Disease Brain MRI Images by Random Forest

Severity classification is the most pivotal task in Alzheimer’s disease diagnosis. Detection of brain structural changes from brain MR images is crucial for Alzheimer’s classification. In this paper, we have proposed a transfer learning and residual separable convolution network for the classification of Alzheimer’s. The proposed network includes three separable convolution layers with two average pooling layers. An upsampling has been performed to regain its spatial resolution for the residual connection. The main intuition of separable convolution is to optimize parameters with depth-wise convolution. Similarly, the residual connection has been used to reduce the vanishing gradient problem. Finally, a three-layer fully connected dense network has been used for the four-class Alzheimer’s classification. Kaggle dataset has been utilized for the experiments to report results. We have achieved an accuracy of 97.32% on the dataset with five-fold cross-validation. Our model has reported an improvement of 1% in jaccard similarity and outperforms the competing models in all vital metrics.

Alzheimer’s severity classification using Transfer Learning and Residual Separable Convolution Network

Structural magnetic resonance imaging (sMRI)-based Alzheimer's disease (AD) classification has attracted a lot of attention and been widely investigated in recent years. However, owing to high dimensionality problem, regions of interest (ROI) of a brain are not characterized properly in spatial domain, which has been a main cause of weakening the discriminating ability of the extracted features. In this study, we propose the ROI-based contourlet subband energy (ROICSE) feature to represent the sMRI image in the frequency domain for AD classification. Specifically, a preprocessed sMRI image is firstly segmented into 90 ROIs by a constructed brain mask. Instead of extracting features from the 90 ROIs in the spatial domain, the contourlet transform is separately performed on these ROIs to obtain their subbands. In order to capture energy information, subband energy (SE) feature vector is constructed based on subbands of an ROI. Afterwards, the SE feature vectors of 90 ROIs are concatenated to form a ROICSE feature of the sMRI image. Finally, SVM classifier is selected to classify 880 subjects from the ADNI and OASIS databases using the ROICSE feature. Experimental results show that the ROICSE approach outperforms six state-of-the-art methods, demonstrating that energy and contour information of the ROI are important to capture differences between the sMRI images of AD and HC subjects. Meanwhile, brain regions related to AD can also be found using the ROICSE feature, indicating that the ROICSE feature can be a promising assistant imaging marker for the AD diagnosis via the sMRI image.

Shaowu Zhang

Papers

Extracting ROI-Based Contourlet Subband Energy Feature from the sMRI Image for Alzheimer's Disease Classification.