There has been significant amount of effort in the literature in measuring the hypothesized widespread structural and functional connectivity alterations in MCI by diffusion tensor imaging (DTI) and/or resting state fMRI (R-fMRI). For instance, the ongoing ADNI-2 project already released dozens of DTI and R-fMRI datasets for early MCI patients. However, a fundamental question arises when attempting to map connectivities in MCI: how to define and localize the best possible network nodes, or Regions of Interests (ROIs), for brain connectivity mapping, and how to perform accurate comparisons of those connectivities across different brains and populations? These still remain as open and urgent problems. Approaches: Our recently developed novel data-driven approach has discovered a map of Dense Individualized and Common Connectivity-based Cortical Landmarks (DICCCOL) in healthy brains. These landmarks possess intrinsically-established correspondences across brains, while their locations were defined in each individual's local image space. In this project, we propose to create a universal and individualized ROI reference system for MCI specifically, by predicting and optimizing the DICCCOL map in well-characterized MCI subjects to be recruited from Duke Medical Center. The resulted DICCCOL map in MCI, named DICCCOL-M, will be annotated into functional networks by concurrent task-based fMRI, R-fMRI, DTI and MRI data. We propose to predict DICCCOL-M in ADNI-2 subjects based on DTI/MRI data and assess the hypothesized large-scale connectivity alterations in ADNI-2 subjects and their longitudinal changes for the purpose of MCI conversion prediction. Significance: 1) The created DICCCOL-M map can be considered and used as a next-generation brain atlas, which will have much finer granularity and better functional homogeneity than the Brodmann brain atlas that has been used in the brain science field for over 100 years. 2) The algorithms will be developed and released based on the open source platform of Insight Toolkit (ITK). The dissemination of the algorithms and associated datasets to the community will significantly contribute to numerous applications in brain imaging that rely on accurate localization of ROIs. 3) Despite recent DTI and R-fMRI studies in the literature to assess brain connectivities in MCI/AD, connectivity alterations in large-scale networks, e.g., over 358 DICCCOL ROIs, and their relationships to AD progression are largely unknown. This knowledge gap will be significantly bridged in this project by assessing these large-scale networks represented by DICCCOL-M in Duke and ADNI-2 subjects.

Public Health Relevance

The proposed studies will develop, validate and apply predictive models of large-scale brain networks derived from well-characterized MCI and control neuroimaging data (Duke Bryan Alzheimer's Disease Research Center) on the ADNI-2 dataset. This project will develop and validate connectivity-based imaging parameters that are predictive of AD progression for early diagnosis of AD.

National Institute of Health (NIH)
National Institute on Aging (NIA)
Research Project (R01)
Project #
Application #
Study Section
Special Emphasis Panel (ZRG1-SBIB-Q (80))
Program Officer
Hsiao, John
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
University of Georgia
Biostatistics & Other Math Sci
Schools of Arts and Sciences
United States
Zip Code
Thung, Kim-Han; Yap, Pew-Thian; Shen, Dinggang (2017) Multi-stage Diagnosis of Alzheimer's Disease with Incomplete Multimodal Data via Multi-task Deep Learning. Deep Learn Med Image Anal Multimodal Learn Clin Decis Support ( 10553:160-168
Liu, Mingxia; Zhang, Jun; Yap, Pew-Thian et al. (2017) View-aligned hypergraph learning for Alzheimer's disease diagnosis with incomplete multi-modality data. Med Image Anal 36:123-134
Zhang, Wei; Jiang, Xi; Zhang, Shu et al. (2017) Connectome-scale functional intrinsic connectivity networks in macaques. Neuroscience 364:1-14
Zhao, Qinghua; Li, Will X Y; Jiang, Xi et al. (2017) Functional brain networks reconstruction using group sparsity-regularized learning. Brain Imaging Behav :
Cheng, Bo; Liu, Mingxia; Shen, Dinggang et al. (2017) Multi-Domain Transfer Learning for Early Diagnosis of Alzheimer's Disease. Neuroinformatics 15:115-132
Zhang, Jun; Liu, Mingxia; Shen, Dinggang (2017) Detecting Anatomical Landmarks From Limited Medical Imaging Data Using Two-Stage Task-Oriented Deep Neural Networks. IEEE Trans Image Process 26:4753-4764
Wang, Yan; Ma, Guangkai; An, Le et al. (2017) Semisupervised Tripled Dictionary Learning for Standard-Dose PET Image Prediction Using Low-Dose PET and Multimodal MRI. IEEE Trans Biomed Eng 64:569-579
Hou, Yingkun; Park, Sang Hyun; Wang, Qian et al. (2017) Enhancement of Perivascular Spaces in 7?T MR Image using Haar Transform of Non-local Cubes and Block-matching Filtering. Sci Rep 7:8569
Li, Yujie; Chen, Hanbo; Jiang, Xi et al. (2017) Transcriptome Architecture of Adult Mouse Brain Revealed by Sparse Coding of Genome-Wide In Situ Hybridization Images. Neuroinformatics 15:285-295
Wang, Jun; Wang, Qian; Peng, Jialin et al. (2017) Multi-task diagnosis for autism spectrum disorders using multi-modality features: A multi-center study. Hum Brain Mapp 38:3081-3097

Showing the most recent 10 out of 199 publications