Classification of Alzheimers Disease Structural MRI Data by Deep Learning Convolutional Neural Networks

📝 Abstract

Recently, machine learning techniques especially predictive modeling and pattern recognition in biomedical sciences from drug delivery system to medical imaging has become one of the important methods which are assisting researchers to have deeper understanding of entire issue and to solve complex medical problems. Deep learning is a powerful machine learning algorithm in classification while extracting low to high-level features. In this paper, we used convolutional neural network to classify Alzheimer’s brain from normal healthy brain. The importance of classifying this kind of medical data is to potentially develop a predict model or system in order to recognize the type disease from normal subjects or to estimate the stage of the disease. Classification of clinical data such as Alzheimer’s disease has been always challenging and most problematic part has been always selecting the most discriminative features. Using Convolutional Neural Network (CNN) and the famous architecture LeNet-5, we successfully classified structural MRI data of Alzheimer’s subjects from normal controls where the accuracy of test data on trained data reached 98.84%. This experiment suggests us the shift and scale invariant features extracted by CNN followed by deep learning classification is most powerful method to distinguish clinical data from healthy data in fMRI. This approach also enables us to expand our methodology to predict more complicated systems.

💡 Analysis

Recently, machine learning techniques especially predictive modeling and pattern recognition in biomedical sciences from drug delivery system to medical imaging has become one of the important methods which are assisting researchers to have deeper understanding of entire issue and to solve complex medical problems. Deep learning is a powerful machine learning algorithm in classification while extracting low to high-level features. In this paper, we used convolutional neural network to classify Alzheimer’s brain from normal healthy brain. The importance of classifying this kind of medical data is to potentially develop a predict model or system in order to recognize the type disease from normal subjects or to estimate the stage of the disease. Classification of clinical data such as Alzheimer’s disease has been always challenging and most problematic part has been always selecting the most discriminative features. Using Convolutional Neural Network (CNN) and the famous architecture LeNet-5, we successfully classified structural MRI data of Alzheimer’s subjects from normal controls where the accuracy of test data on trained data reached 98.84%. This experiment suggests us the shift and scale invariant features extracted by CNN followed by deep learning classification is most powerful method to distinguish clinical data from healthy data in fMRI. This approach also enables us to expand our methodology to predict more complicated systems.

📄 Content

1 Classification of Alzheimer’s Disease Structural MRI Data by Deep Learning Convolutional Neural Networks Saman Sarraf 1,2 , Ghassem Tofighi 3 samansarraf@ieee.org , gtofighi@ryerson.ca 1 Department of Electrical and Computer Engineering, McMaster University, Hamilton, ON, Canada 2 Rotman Research Institute at Baycrest, University of Toronto, ON, Canada 3 Electrical and Computer Engineering Department, Ryerson University, Toronto, ON, Canada arXiv:1603.08631v1 [cs.CV] 22 Jul 2016 2

Abstract
Recently, machine learning techniques especially predictive modeling and pattern recognition in biomedical sciences from drug delivery system to medical imaging has become one of the important methods which are assisting researchers to have deeper understanding of entire issue and to solve complex medical problems. Deep learning is a powerful machine learning algorithm in classification while extracting low to high-level features. In this paper, we used convolutional neural network to classify Alzheimer’s brain from normal healthy brain. The importance of classifying this kind of medical data is to potentially develop a predict model or system in order to recognize the type disease from normal subjects or to estimate the stage of the disease. Classification of clinical data such as Alzheimer’s disease has been always challenging and most problematic part has been always selecting the most discriminative features. Using Convolutional Neural Network (CNN) and the famous architecture LeNet-5, we successfully classified structural MRI data of Alzheimer’s subjects from normal controls where the accuracy of test data on trained data reached 98.84%. This experiment suggests us the shift and scale invariant features extracted by CNN followed by deep learning classification is most powerful method to distinguish clinical data from healthy data in fMRI. This approach also enables us to expand our methodology to predict more complicated systems. Introduction Alzheimer’s Disease Alzheimer’s disease is a neurological, irreversible, progressive brain disorder and multifaceted disease that slowly destroys brain cells causing memory and thinking skills loss, and ultimately the ability to carry out the simplest tasks. The cognitive decline caused by this disorder ultimately leads to dementia. For instance, the disease begins with mild deterioration and gets progressively worse in a neurodegenerative type of dementia. Diagnosing Alzheimer’s disease requires very careful medical 3

assessments such as patients’ history, Mini Mental State Examination (MMSE) and physical and neurobiological exam. Structural imaging based on magnetic resonance is an integral part of the clinical assessment of patients with suspected Alzheimer dementia. Atrophy of medial temporal structures is now considered to be a valid diagnostic marker at the mild cognitive impairment stage. Structural imaging is also included in diagnostic criteria for the most prevalent non-Alzheimer dementias, reflecting its value in differential diagnosis. In addition, rates of whole-brain and hippocampal atrophy are sensitive markers of neurodegeneration, and are increasingly used as outcome measures in trials of potentially disease- modifying therapies [1]. Clinical diagnostic criteria are currently based on the clinical examination and neuropsychological assessment, with the identification of dementia and then of the Alzheimer’s phenotype [2]. Development of an assistant tool or algorithm to classify structural MRI data and more importantly to recognize brain disorder data from healthy subjects has been always clinicians ‘interests. Any machine learning algorithm which is able to classify Alzheimer’s disease assists scientists and clinicians to diagnose this brain disorder. In this work, the convolutional neural network (CNN) which is one of the Deep Learning Network architecture is utilized in order to classify the Alzheimer’s brains and healthy brains and to produce a trained and predictive model. Deep Learning Hierarchical or structured deep learning is a modern branch of machine learning that was inspired by human brain. This technique has been developed based on complicated algorithms that model high-level features and extract those abstractions from data by using similar neural network architecture but much complicated. The neuroscientists discovered the “neocortex” which is a part of the cerebral cortex concerned with sight and hearing in mammals, process sensory signals by propagating them through a complex hierarchy over time. That was the main motivation to develop the deep machine learning focusing on computational models for information representation that exhibit similar characteristics to that of the neocortex [3] [4] [5]. Convolutional Neural Networks which are inspired by human visual system are similar to classic neural networks. This architecture has been particularly designed based on 4

the explicit

This content is AI-processed based on ArXiv data.