Alzheimer's Disease Classification Model Based on MED-3D Transfer Learning

Abstract Background Alzheimer’s disease (AD) is the most common type of dementia, typically characterized by memory loss followed by progressive cognitive decline and functional impairment. Many clinical trials of potential therapies for AD have failed, and there is currently no approved disease-modifying treatment. Biomarkers for early detection and mechanistic understanding of disease course are critical for drug development and clinical trials. Amyloid has been the focus of most biomarker research. Here, we developed a deep learning-based framework to identify informative features for AD classification using tau positron emission tomography (PET) scans. Results The 3D convolutional neural network (CNN)-based classification model of AD from cognitively normal (CN) yielded an average accuracy of 90.8% based on five-fold cross-validation. The LRP model identified the brain regions in tau PET images that contributed most to the AD classification from CN. The top identified regions included the hippocampus, parahippocampus, thalamus, and fusiform. The layer-wise relevance propagation (LRP) results were consistent with those from the voxel-wise analysis in SPM12, showing significant focal AD associated regional tau deposition in the bilateral temporal lobes including the entorhinal cortex. The AD probability scores calculated by the classifier were correlated with brain tau deposition in the medial temporal lobe in MCI participants (r = 0.43 for early MCI and r = 0.49 for late MCI). Conclusion A deep learning framework combining 3D CNN and LRP algorithms can be used with tau PET images to identify informative features for AD classification and may have application for early detection during prodromal stages of AD.

Download Full-text

Classification of Sex and Alzheimer’s Disease via Brain Imaging-Based Deep Learning on 85,721 Samples

10.1101/2020.08.18.256594 ◽

2020 ◽

Author(s):

Bin Lu ◽

Hui-Xian Li ◽

Zhi-Kai Chang ◽

Le Li ◽

Ning-Xuan Chen ◽

...

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Deep Learning ◽

Individual Differences ◽

Brain Imaging ◽

Transfer Learning ◽

Brain Mri ◽

Classification Model ◽

Parahippocampal Gyrus ◽

Industrial Grade

AbstractBeyond detecting brain damage or tumors, little success has been attained on identifying individual differences and brain disorders with magnetic resonance imaging (MRI). Here, we sought to build industrial-grade brain imaging-based classifiers to infer two types of such inter-individual differences: sex and Alzheimer’s disease (AD), using deep learning/transfer learning on big data. We pooled brain structural data from 217 sites/scanners to constitute the largest brain MRI sample to date (85,721 samples from 50,876 participants), and applied a state-of-the-art deep convolutional neural network, Inception-ResNet-V2, to build a sex classifier with high generalizability. In cross-dataset-validation, the sex classification model was able to classify the sex of any participant with brain structural imaging data from any scanner with 94.9% accuracy. We then applied transfer learning based on this model to objectively diagnose AD, achieving 88.4% accuracy in cross-site-validation on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset and 91.2% / 86.1% accuracy for a direct test on two unseen independent datasets (AIBL / OASIS). Directly testing this AD classifier on brain images of unseen mild cognitive impairment (MCI) patients, the model correctly predicted 63.2% who eventually converted into AD, versus predicting 22.1% as AD who did not convert into AD during follow-up. Predicted scores of the AD classifier correlated significantly with illness severity. By contrast, the transfer learning framework was unable to achieve practical accuracy for psychiatric disorders. To improve interpretability of the deep learning models, occlusion tests revealed that hypothalamus, superior vermis, thalamus, amygdala and limbic system areas were critical for predicting sex; hippocampus, parahippocampal gyrus, putamen and insula played key roles in predicting AD. Our trained model, code, preprocessed data and an online prediction website have been openly-shared to advance the clinical utility of brain imaging.

Download Full-text

Alzheimer’s disease Classification from Brain MRI based on transfer learning from CNN

2018 11th Biomedical Engineering International Conference (BMEiCON) ◽

10.1109/bmeicon.2018.8609974 ◽

2018 ◽

Cited By ~ 2

Author(s):

Bijen Khagi ◽

Chung Ghiu Lee ◽

Goo-Rak Kwon

Keyword(s):

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Transfer Learning ◽

Brain Mri ◽

Disease Classification

Download Full-text

An Alzheimer’s Disease Identification and Classification Model Based on the Convolutional Neural Network with Attention Mechanisms

Traitement du signal ◽

10.18280/ts.380533 ◽

2021 ◽

Vol 38 (5) ◽

pp. 1557-1564

Author(s):

Yin Chen

Keyword(s):

Neural Network ◽

Alzheimer’S Disease ◽

Alzheimer's Disease ◽

Deep Learning ◽

Convolutional Neural Network ◽

Brain Mri ◽

Brain Regions ◽

Classification Model ◽

Model Based ◽

Disease Identification

MRI image analysis of brain regions based on deep learning can effectively reduce the workload of doctors in reading films and improve the accuracy of diagnosis. Therefore, deep learning models have great application prospects in the classification and prediction of Alzheimer’s patients and normal people. However, the existing research has ignored the correlation between small abnormalities in local brain regions and changes in brain tissues. To this end, this paper studies an Alzheimer’s disease identification and classification model based on the convolutional neural network (CNN) with attention mechanisms. In this paper, the attention mechanisms were introduced from the regional level and the feature level, and the information of brain MRI images was fused from multiple levels to find out the correlation between the slices in brain MRI images. Then, a spatio-temporal graph CNN with dual attention mechanisms was constructed, which made the network model more attentive to the salient channel features while eliminating the impact of certain noise features. The experimental results verified the effectiveness of the constructed model in identification and classification of Alzheimer’s disease.

Download Full-text