Dementia involves both changes in local cerebral blood flow and metabolic function. Two types of molecular imaging, single-photon emission computed tomography (SPECT) and fluorodeoxyglucose positron emission tomography (FDG-PET), are currently powerful tools in the clinical diagnosis of dementia. By analyzing the patterns of cerebral blood flow and glucose metabolism, a high correlation between them has been found. Therefore, this study aims to use SPECT and its corresponding clinical cognitive function norms, employing AI computation, to predict cerebral glucose metabolic function molecular imaging using structural brain imaging MRI, i.e., zero-dose FDG-PET. This assists in the early diagnosis of individuals at high risk of dementia while reducing radiation exposure and medical expenses. It also addresses the current situation where the prevalence of positron emission imaging is much lower than magnetic resonance imaging.
