Gowsikraja P, Geetha K, Rajan C
Department of Computer Science and Design, Kongu Engineering College, Erode, Tamil Nadu, India.
Department of Computer Science and Engineering, Excel Engineering College, Komarapalayam, Tamil Nadu, India.
NMR Biomed. 2025 Mar;38(3):e5323. doi: 10.1002/nbm.5323.
Alzheimer's disease (AD) is the most prevalent form of dementia, characterized by progressive memory loss and cognitive decline, often affecting behavior and speech. Early detection of AD remains a challenge due to its symptomatic overlap with normal aging and other cognitive disorders, necessitating precise classification methods. This paper proposes a novel Skill Al-Biruni Earth Radius Optimization-enabled Deep Spiking Neural Network (SBERO_Deep SNN) for AD classification using magnetic resonance imaging (MRI). Initially, input MRI images undergo enhancement through thresholding transformations. The segmentation is done using UNeXT, which is optimized by the hybrid SBERO algorithm. The SBERO combines the Skill Optimization Algorithm (SOA) and Al-Biruni Earth Radius (BER). Statistical features, local binary patterns (LBP), and gradient directional patterns (GDP) are then extracted, and classification is performed using a Deep Spiking Neural Network (Deep SNN) trained with SBERO. The proposed method achieves 90.49% accuracy, 89.98% sensitivity, and 90.16% specificity, outperforming existing state-of-the-art techniques in AD classification. The qualitative analysis highlights the robustness of the model in differentiating AD from other cognitive disorders, particularly in early stages, providing a reliable tool for clinical diagnosis.
阿尔茨海默病（AD）是最常见的痴呆形式，其特征为进行性记忆丧失和认知衰退，常影响行为和言语。由于其症状与正常衰老及其他认知障碍存在重叠，AD的早期检测仍然是一项挑战，因此需要精确的分类方法。本文提出了一种新型的基于阿尔 - 比鲁尼地球半径优化的深度脉冲神经网络（SBERO_Deep SNN），用于利用磁共振成像（MRI）对AD进行分类。首先，通过阈值变换对输入的MRI图像进行增强。使用由混合SBERO算法优化的UNeXT进行分割。SBERO将技能优化算法（SOA）和阿尔 - 比鲁尼地球半径（BER）相结合。然后提取统计特征、局部二值模式（LBP）和梯度方向模式（GDP），并使用经SBERO训练的深度脉冲神经网络（Deep SNN）进行分类。所提出的方法实现了90.49%的准确率、89.98%的灵敏度和90.16%的特异性，在AD分类方面优于现有的最先进技术。定性分析突出了该模型在区分AD与其他认知障碍方面的稳健性，尤其是在早期阶段，为临床诊断提供了一个可靠的工具。
