Order from us for quality, customized work in due time of your choice.
Abstract
Alzheimers disease (AD) is an progressive brain neurological disorder which destroys brain cells causing people to lose their memory, mental functions and ability to continue daily activities. Diagnostic symptoms are experienced by patients usually at later stages after irreversible neural damage occurs. Detection of Alzheimers Disease is challenging because sometimes the signs that distinguish Alzheimers Disease MRI data, can be found in normal healthy brain MRI data of older people. Even though this disease is not completely curable,earlier detection can help for proper treatment and to prevent permanent damage to brain tissues. Age and genetics are the greatest risk factors for this disease.
This paper reviews the latest reports on Alzheimers Disease detection based on different types of Neural Network Architectures.
Introduction
Alzheimers disease is a condition that affects the brain, even though the symptoms are mild at first it becomes more severe over time.Common symptoms of Alzheimers disease include memory loss, language problems, and impulsive or unpredictable behavior.As the symptoms get worse, it becomes hard for people to remember recent events and to recognize people they know. Alzheimers disease can range from a state of mild impairment, through to moderate impairment, before eventually reaching severe cognitive decline.
People with Mild Alzheimers disease develop memory problems and cognitive difficulties that may take longer than usual to perform daily tasks, wandering and getting lost. In Moderate Alzheimers disease, the parts of the brain responsible for language, senses, reasoning, and consciousness are damaged. In Severe Alzheimers disease, plaques and tangles are present throughout the brain, causing the brain tissue to shrink substantially.
Hippocampus is the responsible part of the brain for episodic and spatial memory.The reduction in hippocampus causes cell loss and damage specifically to synapses and neuron ends. So neurons cannot communicate anymore via synapses.
As a result, brain regions related to remembering (short term memory), thinking, planning, and judgment are affected.In elderly individuals over the age of 75, identifying differences between Alzheimers Disease brain and a normal functioning brain is difficult as they share similar brain patterns and image intensities.
As per the reports, the below pie chart gives us a clear picture of the people who are mostly affected by Alzheimers Disease considering Age as a factor :
Types of Neural Networks
Deep learning methods are used for classification and prediction have been applied in various fields, including computer vision and natural language processing, both of which demonstrate breakthroughs in performance. Although hybrid approaches have yielded relatively good results, they do not take full advantage of deep learning, which automatically extracts features from large amounts of neuroimaging data.
Artificial Neural Network (ANN)
Artificial Neural Network(ANN), is a group of multiple perceptrons/ neurons in every layer. ANN is also known as a Feed-Forward Neural network because the inputs are being processed in the forward direction only.
ANN consists of 3 layers Input, Hidden and Output layers. The input layer takes the inputs,hidden layer processes and analyses the inputs, and then further the output layer produces the result. Essentially, each layer tries to learn certain weights.
Recurrent Neural Networks (RNN)
Recurrent Neural Networks (RNN) is a special type of network, which unlike feedforward networks has recurrent connections.RNN has a recurrent connection in the hidden state. This looping constraint makes sure that sequential information is captured in the input data.
Therefore, A looping constraint on the hidden layer of ANN turns to RNN.
Convolutional Neural Network(CNN)
Convolutional neural network (CNN) is a deep feed-forward neural network (FNN) composed of multi-layer artificial neurons, with excellent performance in large-scale image processing, classification, segmentation and also for other auto correlated data.
The building blocks of CNNs are filters i.e,. kernels. Kernels are used to extract the relevant features from the input using the convolution operation. Lets try to grasp the importance of filters using images as input data.
Though convolutional neural networks were introduced to solve problems related to image data, they perform impressively on sequential inputs as well.
Survey Study
In our survey, we noticed that ADNI and OASIS open-access datasets were used. Alzheimer’s Disease Neuroimaging Initiative (ADNI) is a multisite study that aims to improve clinical trials for the prevention and treatment of Alzheimers disease. This cooperative study combines expertise and funding from the private and public sector to study subjects with AD, as well as those who may develop AD and controls with no signs of cognitive impairment and has made a global impact.
The primary goal of ADNI is to
- Detect the earliest signs of AD and to track the disease using biomarkers.
- Validate, standardize, and optimize biomarkers for clinical AD trials.
- Make all data and samples available for sharing with clinical trial designers and scientists worldwide.
Open Access Series of Imaging Studies (OASIS) is aimed at making neuroimaging datasets freely available to the scientific community and is hosted by the central.xnat.org provide the community with open access to a significant database of neuroimaging and processed imaging data.OASIS-3 is a longitudinal neuroimaging, clinical, cognitive, and biomarker dataset for normal aging and Alzheimers Disease.
In paper [1], an ensemble of three DenseNet styled models – DenseNet-121, DenseNet-161, and DenseNet-169 is used. For each MRI data, theyve created patches from three physical planes of imaging : Axial or horizontal plane, Coronal or frontal plane, and Sagittal or median plane. These patches are fed to the proposed network as input. Theyve applied transfer learning and the three models have been pre-trained with ImageNet dataset. The individual models are optimized with the Stochastic Gradient Descent (SGD) algorithm to achieve 83.18 overall accuracy.
In paper [2], Two independent datasets are used ADNI-1 as training, ADNI2 as testing, to yield accuracy. The back-propagation algorithm is used to calculate the error between the network output and the expected output in Gradient Computation. After the initial error value is calculated from the given random weight by the least squares method, the weights are updated until the differential value becomes 0.
To improve the performance, multimodal neuroimaging data such as MRI for brain structural atrophy, amyloid PET for brain amyloid-² accumulation, and FDG-PET for brain glucose metabolism have been used. Deep learning approaches have yielded accuracies of up to 86.0% for AD classification and 84.2% for MCI conversion prediction.
In paper [3], This architecture is built using Keras with TensorFlow backend. In Data preprocessing all the data are transformed into a standardized structure by performing co-registration with a standard template and skull stripping.
A 3D CNN model is created inspired by VGG-16 architecture. The model has been trained with categorical cross-entropy loss and the Adam optimizer. 3D models are used here to avoid information loss. The average accuracy of the model achieves 73.4% on ADNI dataset and 69.9%classification accuracy on the OASIS dataset.
In paper [4], First the Data Type Analysis is done, where the proper types of data and ROIs are determined. To verify the effect of segmentation, they segmented the AD and cognitively unimpaired subjects of T1-MR images with the MALP-EM algorithm and obtained the Segmented datasets.
Then, a set of VGG-like Multi-Modality AD classifiers is constructed, which considers both T1-MRI and FDG-PET data as inputs and provides predictions. Then theyve trained and tested the networks with the pMCI and sMCI data.
This network is then programmed based on TensorFlow. Training procedures of the networks are conducted on a personal computer with a Nvidia GTX1080Ti GPU.
In paper [5], the first step is data preprocessing and augmentation, the second stage is feature extraction from input images, and the third step is the classification of dementia classes. They have developed a CNN – based approach inspired by VGG-16 for the classification of dementia stages.
In paper [6], they have used a very deep CNN structure adopted for binary classification method. The shift and scale invariant features are extracted from different layers of CNN architecture resulting in the highly accurate trained model. Furthermore, extensive and unique preprocessing strategies utilized in this work improved the quality of the data fed into LeNet and GoogleNet which ultimately positively impacted the classifier performance.
In paper [7], the random datasets were marked for binary classification and the percentage data 75% for data training and 25% for data testing purpose. The dataset was preprocessed before through training and testing. The architecture of neural networks is using Alexnet architecture with ve layer of convolution. Compared to other journal results, the study method mostly uses ADNI database and LeNet or GoogleNet architecture.
In paper [8], MRI scans are provided in the form of 3D Nifti volumes. At first, skull stripping and grey matter(GM) segmentation is carried out on an axial scans through spatial normalization bias correction and modulation using SPM-g* tool. GM volumes are then converted to JPEG slices using the Python Nibabel package. Slices from start and end which contain no information and discarded from the data set.
Conclusion
From the survey, we can draw a conclusion that there are various technologies and methodologies used for detection of Alzheimers disease at an earlier stage where each individual methodology has a variable precision and accuracy. The two primary datasets, namely ADNI and OASIS are being used where each dataset.Convolutional Neural Network (CNN) based Classification model is used to predict Alzheimers Disease affected-brain v/s a normal aging brain and was able to do so with higher accuracy.This could be used for clinical decision making processes to detect and classify different stages of Alzheimers Disease.
References
- Islam, Jyoti, and Yanqing Zhang. ‘An ensemble of deep convolutional neural networks for Alzheimer’s disease detection and classification.’ arXiv preprint arXiv:1712.01675 (2017).
- Jo, Taeho, Kwangsik Nho, and Andrew J. Saykin. ‘Deep learning in Alzheimers disease: diagnostic classification and prognostic prediction using neuroimaging data.’ Frontiers in aging neuroscience 11 (2019): 220.
- Yagis, Ekin, et al. ‘3D Convolutional Neural Networks for Diagnosis of Alzheimers Disease via structural MRI.’ (2020).
- Huang, Yechong, et al. ‘Diagnosis of Alzheimers disease via multi-modality 3D convolutional neural network.’ Frontiers in Neuroscience 13 (2019): 509.
- Mehmood, Atif, et al. ‘A Deep Siamese Convolutional Neural Network for Multi-Class Classification of Alzheimer Disease.’ Brain Sciences 10.2 (2020): 84
- Sarraf, Saman, and Ghassem Tofighi. ‘Classification of alzheimer’s disease structural MRI data by deep learning convolutional neural networks.’ arXiv preprint arXiv:1607.06583 (2016).
- Al-azdi, Faransi, et al. ‘Design of A Convolutional Neural Network System to Increase Diagnostic Efficiency of Alzheimers Disease.’ IOP Conference Series: Materials Science and Engineering. Vol. 648. No. 1. IOP Publishing, 2019.
- Farooq, Ammarah, et al. ‘A deep CNN based multi-class classification of Alzheimer’s disease using MRI.’ 2017 IEEE International Conference on Imaging systems and techniques (IST). IEEE, 2017.
- Liu, Sheng, et al. ‘On the design of convolutional neural networks for automatic detection of Alzheimers disease.’ Machine Learning for Health Workshop. PMLR, 2020.
- Islam, Jyoti, and Yanqing Zhang. ‘Brain MRI analysis for Alzheimers disease diagnosis using an ensemble system of deep convolutional neural networks.’ Brain informatics 5.2 (2018): 2.
- Khvostikov, Alexander, et al. ‘3D Inception-based CNN with sMRI and MD-DTI data fusion for Alzheimer’s Disease diagnostics.’ arXiv preprint arXiv:1809.03972 (2018)
- Oh, Kanghan, et al. ‘Classification and visualization of Alzheimers disease using volumetric convolutional neural network and transfer learning.’ Scientific Reports 9.1 (2019): 1-16.
- Sarraf, Saman, Ghassem Tofighi, and Alzheimers Disease Neuroimaging Initiative. ‘DeepAD: Alzheimers disease classification via deep convolutional neural networks using MRI and fMRI.’ BioRxiv (2016): 070441.
- Wang, Michael. ‘Interpretable 2D and 3D Convolutional Neural Networks for Alzheimers Disease in Brain Scans.’
- Jain, Rachna, et al. ‘Convolutional neural network based Alzheimers disease classification from magnetic resonance brain images.’ Cognitive Systems Research 57 (2019): 147-159.
Order from us for quality, customized work in due time of your choice.