Bobcat-optimized hybrid quantum-classical spike-driven network for MRI-based Alzheimer's stage prediction

Area of Science:

• Neuroimaging
• Artificial Intelligence
• Computational Neuroscience

Background:

• Alzheimer's disease (AD) diagnosis relies on accurate staging, but conventional MRI methods face limitations in noise sensitivity and modeling complex neuroimaging patterns.
• Automated, accurate staging of Alzheimer's disease is crucial for effective clinical intervention and treatment planning.
• Existing MRI-based approaches struggle with noise, manual interpretation, and capturing intricate non-linear relationships across AD stages.

Purpose of the Study:

• To propose and validate a novel Hybrid Quantum-Classical Spike-Driven Network optimized with the Bobcat Optimization Algorithm (HQSDNet-BOA) for automated, multistage Alzheimer's disease classification using MRI data.
• To enhance MRI data processing through adaptive noise suppression, robust brain tissue segmentation, and discriminative anatomical representation learning.
• To leverage quantum-classical computing and advanced AI techniques for precise identification of non-linear and spatio-temporal AD patterns.

Main Methods:

• Development of the HQSDNet-BOA framework integrating Square Root Sage-Husa Adaptive Robust Kalman Filtering (SRS-HARKF) for noise reduction and Graph-Enhanced Fuzzy Clustering (GEFC) for brain segmentation.
• Implementation of a Hybrid Structural Graph Attention Network (HSGAN) to learn local-global anatomical features and a Hybrid Quantum-Classical Spike-Driven Network (HQSDNet) incorporating quantum learning and spike-driven transformers.
• Optimization of network parameters using the Bobcat Optimization Algorithm (BOA) for enhanced convergence and computational efficiency on ADNI and OASIS-3 datasets.

Main Results:

• The HQSDNet-BOA achieved high classification performance on the ADNI dataset, with 98.8% accuracy, 98.0% precision, 98.5% recall, and 98.25% F1-score.
• The proposed framework demonstrated superior performance compared to existing techniques in accurately classifying multiple stages of Alzheimer's disease.
• The system exhibited improved computational speed, indicating efficient processing of neuroimaging data for AD diagnosis.

Conclusions:

• The HQSDNet-BOA provides a robust, efficient, and clinically relevant solution for accurate Alzheimer's disease stage prediction.
• The integrated quantum-classical approach effectively addresses limitations of conventional MRI analysis for AD diagnosis.
• This framework offers significant potential for advancing neuroimaging-based diagnostic support systems for Alzheimer's disease.