/multiscanner_SBM

Generation of simulation data to demonstrate that our analyais strategy (scanner-specific strategy) outperforms the conventional one for T1-weighted structural MRI Data.

Primary LanguageMATLAB

Source-based morphometry for multi-centre MRI studies

Project Description

This repository We demonstrated that our proposed scanner-specific analysis strategy outperformed the conventional analysis strategy. Therefore, the proposed analysis strategy is recommended when one applies source-based morphometry in a multi-centre MRI study.

image

Citation

If you use the simulation code, please cite:"Ruiyang Ge, Shiqing Ding, Tyler Keeling, William G. Honer, Sophia Frangou & Fidel Vila-Rodriguez, SS-Detect: Development and Validation of a New Strategy for Source-Based Morphometry in Multi-Scanner Studies, Journal of Neuroimaging, In press (https://onlinelibrary.wiley.com/doi/abs/10.1111/jon.12814)".

Application

Neuroimaging research is a collaborative effort, collaborative networks of researchers working together on a range of large-scale studies have been initiated. A prominent issue with a multicentre study is the heterogeneity of the data from scanners with different manufacturers (Siemens, GE, Philips…) and field strength (1.5 T, 3.0 T…).

Usage

  1. To run the simulation, you need to have the SimTB toolbox which you can download from https://trendscenter.org/software/simtb/ or https://github.com/calhounlab/simtb.
  2. Download the three m-files from the 'resources' folder, and save them in the same folder.
  3. The file named 'demo_simulation' is the demo to generate 20 datasets which mimic MRI data from 20 different scanners. Please see (TBD) for more details of the parameters.