Advancing Alzheimer's Disease Detection: A Multi-Modal Deep Learning Approach for Streamlined Clinical Diagnosis

Abstract

Alzheimer’s disease (AD) stands as the most prevalent chronic ailment among the elderly, exhibiting a high incidence rate [1]. Accurate early-stage identification of Alzheimer's disease is crucial for successful treatment and recovery, presenting a significant research challenge in precise diagnosis.Researchers have used different approaches to; diagnose AD but these approaches lack prediction accuracy by various researchers. In recent years, deep learning has become more successful and popular in the area of medical imaging, becoming the method of choice for imaging the medical images and fast growing interest to identify AD. Deep models have improved precision and efficiency in approaching this investigation compared to regular machine learning technologies. This review paper analyses the difference between other research approaches that aim at the early identification of AD; incorporation of ConvNets with Positron Emission Tomography (PET) and Magnetic Resonance Imaging (MRI). The finding is that the networks trained on multi-modal images are more accurate than the networks trained on single-modal images at generalization. The performance of the suggested strategy has been tested using the data set from the Alzheimer’s Disease Neuroimaging Initiative. Accuracy (AUC),Specificity (SPE) and sensitivity (SEN) were the parameters used to assess the performance of the model and all the results obtained were 83.81%,87.50%,75.76% respectively.