Using Fuzzy Systems to Infer Memory Impairment from MRI

Abstract

Alzheimer’s Disease (AD) is a common form of dementia which mostly affects elderly people. Gradual loss in memory and declining cognitive functions are core symptoms associated with AD. Conventional brain images do not provide sufficient information to diagnose AD at an early stage. To delay the progression of memory impairment there is a dire need to develop systems capable of early AD diagnosis. This paper describes a proposed fuzzy method for inferring the risk of dementia using the brain cortical thickness and hippocampus thickness. The aim is to develop a reliable index that allows the evaluation of brain health. The dementia index poses potential to become a biologically-based biomarker for the clinical assessment of patient’s dementia. Results show that the inference value of patient with mild cognitive impairment (MCI) is significantly higher than that of healthy (Control) or schizophrenia (SCZ) patients. Our results suggest that a higher inference value indicates that the patient is at higher risk and is more likely to eventually progress to AD. The system was also tested with age-associated memory impairment (AAMI) patients. The results confirm that our model is able to distinguish between these four patient groups.