Structural MR Brain Image based diagnosis of Alzheimer’s disease using Extreme Learning Machine

Alzheimer’s disease is a neurodegenerative disorder that results in memory degradation and impairment of cognitive function in about 55 million people (WHO 2023 reports), mostly targeting the elderly community. There is a rising need for early diagnosis in order to achieve better treatment outcomes. This study proposes an extreme learning machine (ELM) classifier to differentiate between various stages of Alzheimer’s disorder. It is carried out using selectively extracted brain regions from structural MR brain images. The diverse datasets from Alzheimer’s Disease Neuroimaging Initiative (ADNI) [2200 patient data] fortifies the approach of this study. Preprocessing and skull stripping the MR brain images, segmentation of the desired bio markers are done. Essential geometric features and textural features are extracted for different stages of Alzheimer’s disorder namely Mild cognitive impairment (MCI), Late mild cognitive impairment (LMCI) and Alzheimer’s disorder (AD). Extreme learning machine (ELM) classifier is utilized to discern the various stages. Once the model is trained and validated, it can be used to predict the Alzheimer's disease stage of new, unseen data. The model's output could indicate the likelihood or stage of Alzheimer's disease based on the input features. This study aims to contribute to the development of Alzheimer’s disorder diagnostics and emphasize on the potential of using machine learning techniques for analysis of structural MR brain images.