Ricci Flow-Based Approach for Early Diagnosis of Alzheimer's Disease

Early diagnosis of Alzheimer's disease (AD) is increasingly important due to its rising prevalence and significant impact on individuals, families, and healthcare systems. Hippocampal atrophy is a well-established and significant biomarker for AD. Advanced techniques like MRI imaging and surface parameterization have shown considerable promise in improving the accuracy and speed of AD diagnosis. This study aims to utilize the Ricci flow method to map the 3D hippocampal surface to a 2D sphere and extract relevant features for early AD detection. The process involves several key steps: inputting an MRI scan and preprocessing to isolate the hippocampal surface, applying the Ricci flow to map this surface to a sphere, constructing a feature vector using Linear Discriminant Analysis (LDA) and Kernel LDA, and employing various classifiers to diagnose AD, with model evaluation based on the ADNI dataset. Experimental results reveal that combining Ricci flow-based feature extraction with Kernel LDA significantly improves diagnostic accuracy. The model achieves classification accuracies of 97.28% (NC/AD), 96.14% (NC/EMCI), 96.45% (NC/MCI), 94.83% (EMCI/LMCI), 95.84% (MCI/AD), and 95.37% (LMCI/AD). Additionally, it attains 93.65% and 92.30% accuracy in three-way and four-way classification tasks, respectively. These results outperform most reviewed studies and are comparable to others. This research highlights the potential of merging advanced 3D imaging techniques with mathematical models to enhance diagnostic precision, emphasizing the critical role of early detection in the effective treatment and management of Alzheimer's disease.