An efficient ranking-based ensembled multiclassifier for neurodegenerative diseases classification using deep...

Area of Science:

• Neuroscience
• Artificial Intelligence
• Medical Imaging

Background:

• Neurodegenerative diseases like Alzheimer's (AD) and Parkinson's (PD) cause progressive neuron loss, impacting cognition and motor function.
• Current diagnostic methods, including machine learning, have accuracy limitations.
• Accurate and early diagnosis is crucial for managing these debilitating conditions and improving patient outcomes.

Purpose of the Study:

• To develop and validate a novel ranking-based ensemble approach using deep learning for improved diagnosis of AD and PD.
• To enhance the accuracy and generalization performance of neurodegenerative disease prediction.
• To compare the proposed method against existing diagnostic approaches.

Main Methods:

• A three-phase modeling procedure involving data preprocessing, classification using five deep learning models, and a ranking-based ensemble strategy.
• Utilized Magnetic Resonance Imaging (MRI) datasets: Alzheimer's Disease Neuroimaging Initiative (ADNI) for AD and Parkinson's Progressive Marker Initiative (PPMI) for PD.
• Employed weighted strategies and rank calculations to ensemble predictions from individual deep learning models.

Main Results:

• Achieved high classification accuracies: AD-CN (97.89%), AD-MCI (99.33%), CN-MCI (99.44%), PD-CN (99.22%), PD-SWEDD (97.56%), CN-SWEDD (98.22%).
• Demonstrated promising multi-class classification accuracy: 97.18% for AD and 97.85% for PD.
• The proposed deep learning ensemble method outperformed existing approaches in both binary and multi-class classification tasks.

Conclusions:

• The developed deep learning-based ensemble technique offers a competitive and accurate solution for predicting Alzheimer's and Parkinson's diseases.
• The approach significantly enhances generalization performance in diagnosing neurodegenerative diseases.
• This method represents a promising advancement in the early and accurate detection of AD and PD.