Caspian Journal of Internal Medicine

Background: By incorporating advanced signal processing and artificial intelligence (AI)/machine learning techniques (MLT) in an automated manner, this system can be trained using extensive patient data, physiological signals, and images. Using state-of-the-art AI models and high-quality clinical data has shown promise in improving prognostic and diagnostic models for neurological diseases. However, it is essential to recognize that AI, machine, and deep-learning algorithms are not universally applicable solutions for all clinical data and inquiries.
Methods: This paper offers a comprehensive review of cutting-edge research on the automated diagnosis of major neurological disorders, with a primary focus on the application of AI techniques. An extensive search was conducted in the PubMed and Scopus databases from September 2023 to January 2025; 220 articles were retrieved from the initial search. After eliminating duplicate entries, a total of 105 distinct articles remained. Subsequently, in March 2025, 4 more articles directly pertinent to the topic of interest were identified and included.
Results: AI methodologies, particularly Support Vector Machines (SVM) and XGBoost, have demonstrated considerable potential in improving the efficacy of screening tests. AI methods enhance neurological disease diagnosis by integrating advanced diffusion MRI techniques and next-generation PET tracers. Deep learning models, such as convolutional neural networks, improve diagnostic precision and facilitate early detection of neurodegeneration and inflammation in various disorders.
Conclusions: Artificial intelligence in combination with conventional methods serves as a valuable resource for the early, precise, and non-invasive identification of neurodegenerative disorders. Additionally, it aids in evaluating treatment efficacy and forecasting the progression of these diseases.