Alotaibi Shoayee Dlaim, Alharbi Abeer A K
Department of Artificial Intelligence and Data Science, College of Computer Science and Engineering, University of Hail, Hail, Saudi Arabia.
King Salman Center for Disability Research, 11614, Riyadh, Saudi Arabia.
Sci Rep. 2025 Jul 9;15(1):24659. doi: 10.1038/s41598-025-09124-6.
Dementia is a degenerative and chronic disorder, increasingly prevalent among older adults, posing significant challenges in providing appropriate care. As the number of dementia cases continues to rise, delivering optimal care becomes more complex. Machine learning (ML) plays a crucial role in addressing this challenge by utilizing medical data to enhance care planning and management for individuals at risk of various types of dementia. Magnetic resonance imaging (MRI) is a commonly used method for analyzing neurological disorders. Recent evidence highlights the benefits of integrating artificial intelligence (AI) techniques with MRI, significantly enhancing the diagnostic accuracy for different forms of dementia. This paper explores the use of AI in the automated detection and classification of dementia, aiming to streamline early diagnosis and improve patient outcomes. Integrating ML models into clinical practice can transform dementia care by enabling early detection, personalized treatment plans, and more effectual monitoring of disease progression. In this study, an Enhancing Automated Detection and Classification of Dementia in Thinking Inability Persons using Artificial Intelligence Techniques (EADCD-TIPAIT) technique is presented. The goal of the EADCD-TIPAIT technique is for the detection and classification of dementia in individuals with cognitive impairment using MRI imaging. The EADCD-TIPAIT method performs preprocessing to scale the input data using z-score normalization to obtain this. Next, the EADCD-TIPAIT technique performs a binary greylag goose optimization (BGGO)-based feature selection approach to efficiently identify relevant features that distinguish between normal and dementia-affected brain regions. In addition, the wavelet neural network (WNN) classifier is employed to detect and classify dementia. Finally, the improved salp swarm algorithm (ISSA) is implemented to choose the WNN technique's hyperparameters optimally. The stimulation of the EADCD-TIPAIT technique is examined under a Dementia prediction dataset. The performance validation of the EADCD-TIPAIT approach portrayed a superior accuracy value of 95.00% under diverse measures.
痴呆症是一种退行性慢性疾病，在老年人中越来越普遍，给提供适当护理带来了重大挑战。随着痴呆症病例数量持续上升，提供最佳护理变得更加复杂。机器学习（ML）通过利用医学数据来加强对各类痴呆症风险个体的护理规划和管理，在应对这一挑战中发挥着关键作用。磁共振成像（MRI）是分析神经疾病的常用方法。最近的证据凸显了将人工智能（AI）技术与MRI相结合的益处，显著提高了对不同形式痴呆症的诊断准确性。本文探讨了AI在痴呆症自动检测和分类中的应用，旨在简化早期诊断并改善患者预后。将ML模型整合到临床实践中，可以通过实现早期检测、个性化治疗方案以及对疾病进展更有效的监测来改变痴呆症护理。在本研究中，提出了一种使用人工智能技术增强思维能力障碍者痴呆症自动检测和分类（EADCD-TIPAIT）技术。EADCD-TIPAIT技术的目标是使用MRI成像对认知障碍个体中的痴呆症进行检测和分类。EADCD-TIPAIT方法进行预处理，使用z分数归一化对输入数据进行缩放以实现此目的。接下来，EADCD-TIPAIT技术执行基于二元灰雁优化（BGGO）的特征选择方法，以有效识别区分正常和受痴呆症影响的脑区的相关特征。此外，采用小波神经网络（WNN）分类器来检测和分类痴呆症。最后，实施改进的樽海鞘群算法（ISSA）以最佳地选择WNN技术的超参数。在痴呆症预测数据集下检验了EADCD-TIPAIT技术的效果。EADCD-TIPAIT方法的性能验证在各种测量下呈现出95.00%的卓越准确率值。
