An Interpretable Framework for Identifying Cerebral Microbleeds and Alzheimer's Disease Severity using Multimodal...

Area of Science:

• Neuroimaging and computational biology
• Artificial intelligence in medicine
• Biomarker discovery for neurodegenerative diseases

Background:

• Cerebral microbleeds (CMBs) are crucial indicators for cerebrovascular disorders like stroke and Alzheimer's disease (AD).
• Automating CMB detection and enhancing prediction reliability for early chronic disease diagnosis remains a significant challenge.
• Existing methods using spike neural networks (SNNs) and decision trees lack interpretability.

Purpose of the Study:

• To develop an explainable artificial intelligence (XAI) system for automated detection of CMBs in MRI images.
• To determine Alzheimer's disease (AD) severity using gene expression data.
• To enhance the reliability and interpretability of diagnostic and prognostic predictions.

Main Methods:

• Utilized a spike neural network (SNN) for CMB identification in MRI images.
• Employed a decision tree for analyzing gene expression data to assess AD severity.
• Developed and applied Pixel Density Analysis (PDA) for MRI data interpretation.
• Implemented Probabilistic Graphical Models (PGM) for gene expression data interpretation.

Main Results:

• Successfully identified cerebral microbleeds (CMBs) in MRI scans.
• Determined the severity of Alzheimer's disease (AD) through gene expression analysis.
• Achieved enhanced reliability in prediction outputs via explainable AI (XAI) methods.
• Provided interpretable insights into diagnostic and prognostic decision-making processes.

Conclusions:

• Explainable AI (XAI) methods, specifically PDA and PGM, significantly improve the reliability of CMB detection and AD severity assessment.
• The developed XAI system offers interpretable outputs, addressing the limitations of complex 'black-box' models.
• This approach holds promise for earlier and more accurate diagnosis and management of neurodegenerative conditions.