Artificial Intelligence based Neurodegenerative Disease Diagnosis and Research Analysis using Functional MRI (FMRI): A Review

Biomedical image processing is gaining importance along with the various noninvasive imaging modalities in the medical field for both clinical and research purposes. Medical image processing algorithms have an important role in noise and artifacts filtering, image segmentation, feature extraction and classification, that can accelerate disease identification. This review paper evaluates functional magnetic resonance imaging (fMRI), its related techniques (rs-fMRI) and task fMRI (t-MRI) for the diagnosis, analysis and treatment of brain disorders, neurodegenerative diseases and cognitive impairment. In fMRI, an incredible amount of behavioral response analysis makes use of a contrast imaging method called Blood Oxygen Level Dependent (BOLD). Comparisons on resting state fMRI, task fMRI, aging and defective brain, brain cancers and other brain pathologies were done here. However, fMRI still face a number of challenges such as limited sample size, bias and variance, low temporal resolution, over fitting and requirement of longitudinal research. In addition to reflecting neural activity, BOLD signals are highly sensitive to non-neuronal changes. Studies of brain tumors, lesions, and other forms of pathology in humans have not yet been investigated using fMRI.