Identification and Evaluation of Multimodal Connectomics in Early Alzheimer’s Dementia

Alzheimer’s disease is a common neurodegenerative disease with a long disease course and is one of the main causes of dementia. To investigate the longitudinal disorder of multimodal brain networks in the process of cognitive decline and construct a connectome-based identification model for questionable dementia, we construct functional and structural connectivity of brain networks by using magnetic resonance imaging (MRI) data from the OASIS database. The discovery group of longitudinal database includes 113 healthy aging individuals, 65 healthy subjects at the time of scanning but diagnosed as questionable dementia during follow-up, and 39 subjects with questionable dementia at the time of scanning. The results showed that the default mode network of the potential dementia group had structural and functional abnormalities as early as the cognitive normal stage. Furthermore, a logistic regression model was further constructed using the connectomics features of 455 subjects in identification group to construct an objective mapping between the imaging biomarkers and the clinical dementia scale. The average classification accuracy was about 85%. The above findings may indicate the potential changes of the central nervous system in the early stage of dementia, and serve as an auxiliary and reference for clinicians in the early diagnosis of dementia.