Alzheimer’s Aid: ML Detection and Therapeutic Games

Alzheimer's Disease (AD) is a neurodegenerative disease that affects an individual's cognitive and functional abilities over a gradual period of time. Therefore, early diagnosis is critical. In this regard, this research looks into two aspects, putting deep learning techniques to use for exact detection of Alzheimer’s disease through MRI scans, and developing cognitive rehabilitation therapy games. We worked with the "Augmented Alzheimer MRI Dataset v2" which is hosted on Kaggle. It has 6400 MRI scans of patients with moderate dementia which we arranged into Mild Demented, Very Mild Demented, Moderate Demented, and Non-Demented. Two convolutional neural networks, DenseNet21, and VGG19 were assessed. VGG19 outperformed DenseNet21, attaining the best accuracy score of 95.8%, sensitivity of 96.4%, and specificity of 94.2%. Other therapeutic attention and memory games also showed considerable improvement in cognitive skills. This research provides a more complete solution for Alzheimer’s disease by bridging the gap between diagnostics and rehabilitation.