سامانه همایش های دانشگاه آزاد اسلامی واحد نجف آباد

صفحه اصلی / International Conference on Artificial Intelligence; City, Industry and Health

Artificial Intelligence in the Diagnosis and Prognosis of Neurodegenerative Disorders: A Systematic Review of Algorithms, Challenges, and Future Directions

نویسندگان :

Donya Forghani¹ Mohamad Shahgholi²

1- 1Department of Biomedical Engineering, Na.C., Islamic Azad University, Najafabad, Iran 2- Department of Mechanical Engineering, Na.C., Islamic Azad University, Najafabad, Iran

کلمات کلیدی :

Artificial Intelligence،Deep Learning،Explainable AI (XAI)،Multimodal Data Integration،Neurodegenerative Disorders،brain tumor،multiple sclerosis،Alzheimer،Parkinson

چکیده :

Neurodegenerative disorders like Alzheimer's disease, Parkinson's disease, multiple sclerosis, and brain tumors rank among the most challenging 21st-century health abnormalities. The limitations of traditional diagnostic methods, combined with clinical complexity of the conditions, have grown to direct research attention toward AI-based alternatives. Particularly, machine learning (ML) and deep learning (DL) methods have emerged as excellent candidates for identifying hidden patterns and generating highly accurate predictions. This review evaluates recent research that has utilized artificial intelligence for diagnosing and predicting neurodegenerative diseases. The data used in these studies are magnetic resonance imaging (MRI), positron emission tomography (PET), electroencephalography (EEG), audio files, genetic data, and clinical features. Various algorithms like convolutional neural networks (CNN), long short-term memory (LSTM) networks, generative adversarial networks (GAN), support vector machines (SVM), and hybrid architecture have been employed for processing and analysis of data. Results from some research demonstrate that multimodal models—above all, imaging and non-imaging data combination models—have recorded excellent diagnostic performance, up to 99.47%. However, there are common challenges across many studies including lack of diversity in data, limited model interpretability, and poor external validation. This paper highlights the need to develop transparent, generalizable, and ethical AI systems, and identifies some key future research avenues.