Applications of artificial intelligence in the diagnosis of neurodegenerative changes in Alzheimer's disease

Neurodegenerative diseases, particularly Alzheimer’s disease, remain one of the greatest challenges of modern medicine due to late clinical manifestation and lack of effective curative therapies. Traditional diagnostic methods, including neuropsychological tests and standard neuroimaging, are limited in detecting subtle early changes. Artificial intelligence, especially machine learning and deep learning, has demonstrated significant potential in improving early diagnosis, classification, and prediction of disease progression. Convolutional neural networks applied to MRI and PET data enable automated feature extraction and accurate differentiation between Alzheimer’s disease, mild cognitive impairment, and healthy controls. Integrative multimodal approaches that combine imaging, cerebrospinal fluid biomarkers provide higher diagnostic sensitivity and specificity compared to unimodal analyses. Despite the advantages of AI in efficiency, scalability, and early detection, challenges remain regarding generalisability, interpretability, cost, and clinical implementation. This review highlights the current applications of AI in Alzheimer’s disease diagnostics, contrasts them with classical methods, and discusses methodological limitations, ethical aspects, and perspectives for future integration into clinical practice.