Ensemble Based MRI Analysis for Early Detection of Alzheimer’s Disease

The neurodegenerative disorder Alzheimer’s disease is indicated by cognitive deficits and memory loss. Early identification and diagnosis of the disease contribute positively to therapeutic intervention opportunities, and can improve patient outcomes following diagnosis. Structural brain abnormalities linked with Alzheimer’s can be reliably measured through magnetic resonance imaging, making it a strong imaging modality for cognitive impairment assessment, including Alzheimer’s. This study develops an ensemble learning strategy to diagnose patients with Alzheimer’s using MRI data. Specifically, an ensemble strategy uses the strengths of multiple less complex machine learning models to improve predictive accuracy and robustness locally or across populations, as compared to using a single model. The proposed ensemble diagnostic model utilizes pre-processed MRI images in order to extract relevant features of interest, and the software stack publishes the assessment of various machine learning models. The proposed method exhibits key performance metrics as accuracy-0.99, F1 score-0.98, recall-1.00 and precision-0.96. It shows that ensemble learning is a very powerful machine learning method to use for machine learning predictions when compared to traditional and individual approaches.