addecode
Bass Connections project on Exercise Therapy and Alzheimer's Disease
This repository contains the code for reading in MRI k-space data from the Bruker machine and reconstructing images from undersampled k-space data. Additional toolboxes are also needed and can be found in BassData/MRI Toolboxes on Box. Previously extracted k-space data is also available in BassData, saved as brukerdata.mat.
The main script in this repository is run_bart.m, which walks you through reading in data, generating a sampling pattern, and reconstructing images from undersampled data, all using the Berkeley Advanced Reconstruction Toolbox. You may elect to use our data or use data from the espirit workshop, in which case the BART folder at https://github.com/mikgroup/espirit-matlab-examples must be downloaded.
Reconstruction parameters may be optimized by varying iter_type and setting the range of values to explore. Reconstruction quality is assessed using relative L2-norm error, peak signal-to-noise ratio, and global structural similarity index. Various plots to compare reconstruction performance (images, quality metrics) may be generated (set 'gen_', 'plot_' to true)