Differential Identification of Prodromal Stages of Alzheimer’s Disease Using Tissue Probability Map (TPM) based Network

A lot of efforts have been made by researchers for easy detection of the prodromal phase of Alzheimer’s disease (AD) and other dementia to enable curative measures. Among the leading approaches that show promising results is the use of complex network theory on neuroimaging data such as functional magnetic resonance imaging (fMRI), diffusion tensor imaging (DTI), magnetoencephalogram (MEG), electroencephalogram (EEG) etc. However, exploring the network theory using the tissue probability Map (TPM) of magnetic resonance imaging (MRI) data has been quite unexplored. Therefore, in this paper, we developed the generalized improved multiscale permutation entropy (GIMPE) for the computation of complexity of grey matter (GM) TPM for all the considered region of interests (ROIs). In order to formulate a well defined network, the vectors, GIMPEs of all ROIs are taken as nodes and the edges between the nodes are defined by the Euclidean distance between the corresponding vectors (GIMPEs). The validation of our approach on MRI data accentuates the importance of the proposed approach as well as the significance of TPM based brain networks for the discrimination and differential diagnosis of normal aging, prodromal phase and the later phase of AD.