Diagnosis of Alzheimers Disease Using Dynamic High-Order Brain Networks

Back to Main BRAIN TMM page

PI: Shen, Dinggang (contact); Yap, Pew-thian 

Email: dgshen@med.unc.edu

Institution: University of North Carolina Chapel Hill 

Title: Diagnosis of Alzheimers Disease Using Dynamic High-Order Brain Networks

Grant #: EB022880 

Status: Completed

Deliverables:

https://www.imagwiki.nibib.nih.gov/sites/default/files/Yap_Shen_UNC_Sof…

In this project, we have investigated the utility of high-order dynamic networks in the diagnosis of brain disorders.

Our main findings are summarized as follows:

1. Functional Connectivity networks (FCNs) derived from resting-state functional magnetic resonance imaging (rs-fMRI) data are useful for the analysis and diagnosis of brain diseases, such as AD and its prodrome, MCI.

2. Dynamic changes of FC, which may indicate changes in macroscopic neural activity patterns in cognitive and behavioral aspects, are useful for brain disease diagnosis.

3. Integrating both temporal and spatial properties of Dynamic Connectivity Networks (DCNs) can improve the performance of machine learning methods in automatic brain disease diagnosis.

4. Dynamic FC features can complement static FC features to improve diagnosis.

5. The fusion of low- and high-order FC improves diagnostic performance.

6. Deep learning provides a way to learn end-to-end diagnosis models, circumventing the limitations of conventional handcrafted feature extraction methods.

7. Different views (e.g., order, spatial, and temporal) of rs-fMRI data provide complementary information for disease diagnosis.

8. Incorporating information from different modalities (e.g., MRI, PET, and CSF) further improves disease diagnosis.

 

Back to Main BRAIN TMM page

Table sorting checkbox
Off