USE OF ALGORITHMS FOR EARLY DIAGNOSIS OF ALZHEIMER'S, PARKINSON'S AND MULTIPLE SCLEROSIS

Abstract

The application of artificial intelligence (AI) algorithms in the early diagnosis of neurodegenerative diseases, such as Alzheimer's, Parkinson's, and multiple sclerosis, has made significant advances. This integrative review analyzed 22 studies published between 2013 and 2023, using databases such as PubMed, IEEE Xplore, and Scopus. The results highlight the use of Convolutional Neural Networks (CNNs) in neuroimaging, such as magnetic resonance imaging and PET scans, with an accuracy of over 90% in diagnosing Alzheimer's. Multimodal approaches that integrate clinical and genetic data have demonstrated increasing efficacy. For Parkinson's, algorithms that analyze vocal signals and tremors have a sensitivity between 85% and 92%, while deep learning tools allow the detection of minimal motor changes. In the case of multiple sclerosis, models that combine magnetic resonance imaging and immunological profiles show high accuracy in the early detection of brain lesions. Despite the advances, challenges persist, including the standardization of databases, large-scale validation, and interpretation of results by health professionals. The limitations of this study include the lack of methodological uniformity in the articles analyzed and the scarcity of data from large clinical studies. It is proposed that future research invest in the integration of different data sources, expansion of population samples, and development of more transparent algorithms, facilitating its clinical adoption. It is concluded that AI has great potential to transform early diagnosis, allowing more effective and personalized interventions, but it requires refinement to consolidate its practical applicability.