Development of an artificial neural network using texture and shape features extracted from the hippocampal regions on magnetic resonance imaging to aid in the diagnosis of Alzheimer's Disease

Abstract

The Alzheimer's disease (AD) is the most common cause of dementia in the world population, comprising around 60% of all cases and affecting 20% of the population over the age of 80 years. It is an irreversible neurodegenerative disease that causes loss of mental function due to deterioration of the brain tissue, affecting people in different manners. The most common effects include the difficulty remembering new information, solve simple problems and complete day-to-day functions at home. The magnetic resonance imaging has been routinely used in clinical practice as an aid to diagnosis and accompaniment of the disease due to its excellent contrast between the soft tissues and ability to provide information about the shape of the organs, thus allowing, the detection of changes in the brain induced by the AD. The diagnosis of the AD in early stages is fundamental to try to preserve to the maximum the intellectual abilities and prolong the patient's quality of life. Therefore, the primary objective of this research is to develop an artificial neural network, using texture and shape features extracted from the hippocampal regions in MR images, to aid the detection of AD by classifying MR images in healthy control (HC) and AD.