A fuzzy approach to assess the indication of dementia based on magnetic reasoning imaging

Alzheimer´s disease (AD), also known as geriatric dementia, is the most serious dementia occurred among the elders whose age is over 65. Cognitive functions or memory will decay with age and are irreversible. The cause of dementia remains a mystery and the dementia population normally increases within aged societies. It is not only hard to diagnose whether an elderly will suffer from dementia but also the mental disorders induced by AD results in a heavy burden to his family members. Consequently, there is a dire need to examine whether an elderly has an indication of AD and takes necessary prevention to delay the cognition decay. Literature reported that AD subjects show thinner cortical thickness on MRI than controlled subjects. However, there is no clear hint that associates cortical thickness and state of dementia. In this study we first locate the white matter, gray matter, Caudate nucleus, and Thalamus from MRI in order to extract the cortical thickness information, reconstruct the spherical brain models from normal subjects and AD by the Freesurfer software developed at Harvard University. Then, fuzzy c-means clustering is used to find centroids from persistence diagrams that come from cortical thickness. Finally, a fuzzy model is proposed to assess the indication of dementia from cortical thickness of MRI. Experimental results of controlled and AD subjects verify that the proposed fuzzy model can help examine whether an elderly suffers from dementia symptom.