Evaluating dementia risk from MRI by fuzzy inference model

It is well-known that individuals´ cognitive functions and memory will decay with age, an irreversible process that can result in various diseases. Among them, Alzheimer´s disease (AD) is the most common mental disease which affects elderly people whose age is over 65. Though the main causes remain a mystery till nowadays, there is no doubt that this disease appears repeatedly in aged societies. Thus, there is a strong need to improve our current knowledge in order to figure out whether a patient suffers from AD or not, which is not even easy to understand. Since the literature reports that AD subjects present a thin cortical thickness, this study continues our previous one towards this direction. In this study we improved the previous fuzzy c-means clustering which is used to find centroids from persistence diagrams, this one coming from the cortical thickness. As last step we defined a new fuzzy model to estimate whether a subject is affected by dementia or not, based on cortical thickness. New and more experiments conducted on normal, schizophrenia and AD patients gather that our system is reliable and indeed can help to find the Alzheimer´s disease in patients.