Scholarlink

<p>The degenerative neurological condition known as Alzheimer’s disease (AD) is characterized by cognitive decline and memory loss. Since standard clinical procedures frequently fail to differentiate AD from normal aging and related illnesses, early and accurate diagnosis is still difficult to achieve. Electroencephalography (EEG) provides a cost-effective, non-invasive diagnostic tool; however, its signals are often noisy and difficult to interpret, particularly in the central lobe, where cognitive activity is most prominent. This study introduces a novel framework, MQCGAN-IOPA, for enhanced AD identification from central lobe EEG. For deep contextual learning, the system combines a Multi-aspect Graph Attention Network (MGAN) with a Hybrid Quantum-Classical Convolutional Neural Network (HQ2CNN), the Discrete Quaternion Quadratic-phase Fourier Transform (D2QFT) for spatial-frequency feature extraction, and Adaptive Side Window Joint Bilateral Filtering (AWJBF) for noise reduction. To enhance accuracy, model parameters are optimized using the Improved Orca Predation Algorithm (IOPA). Experiments on the CAUEEG dataset show better performance than benchmark models like ResNet and FFT-GPA, with 99.96% accuracy, 99.95% recall, and a mere 0.01% error rate. In support of future AI-assisted clinical diagnostics, these findings demonstrate the potential of the suggested quantum-classical graph attention framework as a formidable instrument for accurate and timely AD diagnosis.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhanced Alzheimer’s Disease Identification from Central Lobe EEG Using Multi-Aspect Quantum-Classical Graph Attention Networks Optimized by Improved Orca Predation

  • 期刊论文

  • 发表日期：

    2025年12月18日

查看全文
  • B. Ramesh,
  • Elangovan Muniyandy,
  • Krishna Prakash Arunachalam,
  • S. Mythili

摘要

The degenerative neurological condition known as Alzheimer’s disease (AD) is characterized by cognitive decline and memory loss. Since standard clinical procedures frequently fail to differentiate AD from normal aging and related illnesses, early and accurate diagnosis is still difficult to achieve. Electroencephalography (EEG) provides a cost-effective, non-invasive diagnostic tool; however, its signals are often noisy and difficult to interpret, particularly in the central lobe, where cognitive activity is most prominent. This study introduces a novel framework, MQCGAN-IOPA, for enhanced AD identification from central lobe EEG. For deep contextual learning, the system combines a Multi-aspect Graph Attention Network (MGAN) with a Hybrid Quantum-Classical Convolutional Neural Network (HQ2CNN), the Discrete Quaternion Quadratic-phase Fourier Transform (D2QFT) for spatial-frequency feature extraction, and Adaptive Side Window Joint Bilateral Filtering (AWJBF) for noise reduction. To enhance accuracy, model parameters are optimized using the Improved Orca Predation Algorithm (IOPA). Experiments on the CAUEEG dataset show better performance than benchmark models like ResNet and FFT-GPA, with 99.96% accuracy, 99.95% recall, and a mere 0.01% error rate. In support of future AI-assisted clinical diagnostics, these findings demonstrate the potential of the suggested quantum-classical graph attention framework as a formidable instrument for accurate and timely AD diagnosis.