Early Detection of Alzheimer's Disease Using Gamified Self Administrative Gerocognitive Exam and Deep Learning

This study uses a convolutional neural network (CNN) for multiclass classification in a gamified version of the SAGE (Self-Administered Gerocognitive Examination) test to introduce a distinctive approach for the early identification of Alzheimer's disease (AD). The application features three interactive games: a card flipping and matching game, a “set the time” game with a maximum of four attempts, and a timed challenge to name 12 animals within three minutes, verified through an API (Application Programming Interface). These engaging activities reduce anxiety associated with traditional assessments while providing valuable insights into cognitive states. Results are collected, and if a specific threshold is reached, the system recommends further evaluation via MRI (Magnetic Imaging Resonance) testing. Additionally, the application can analyze MRI scans to monitor disease progression. Key benefits include early detection through non-invasive methods, enhanced user engagement, real-time feedback for timely intervention, and comprehensive monitoring of cognitive health. Our findings suggest that combining gamification with advanced deep learning (DL) techniques can significantly improve early detection strategies for cognitive decline.