Deep-learning-based application of imaging genomics in brain diseases
Approved Research ID: 75310
Approval date: November 1st 2021
Aim:The project aims to develop a computer method for predicting brain disease from multi-modal brain image and genomic data. This method can explore the mechanisms of brain diseases and provide biomarkers for treatments.
Scientific rationale: Patients with different kinds of cognitive disorders differ in the brain structure, functional connections and brain tissue. For example, the volume changes in hippocampus brain region is proved to be closely correlated with the progression of Alzheimer Disease; the functional connections in patients with cognitive disease is very different compared with the normal people. All these changes can be observed by the neuroimaging techniques, such as fMRI ,DTI and T1-weighted MRI. In addition, the cognitive disease is greatly affected by genetic inheritance, and studies have confirmed that there are multiple gene loci strongly correlated with changes in hippocampus volume. Therefore, through the analysis of brain images and gene data, we can study the mechanism of brain circuits for the treatment of brain cognitive diseases, and provide biological markers for the diagnosis and early treatment of diseases.
Project duration:The project will last for 3 years
Public health impact:This project will produce a new computational method for diagnosis, typing, prediction and prevention of brain diseases. The results of this project will probably provide substantial new insights into the genetic landscape of the brain and offer a scientific value that could advance application on normal brain development and neurological disorders and improve the diagnosis of brain disease.