Download Gannet v3.5.0
New features
- Gannet is now compatible with BIDS
bids-matlab, a MATLAB toolkit for interfacing with BIDS datasets, is embedded in Gannet, and users can now input BIDS datasets directly intoGannetLoad.m(see documentation for usage instructions)
- Added image QA metrics based on MRIQC
GannetSegment.mnow calculates and saves the following QA metrics for each anatomical image:- Signal-to-noise ratio (SNR) of GM, WM, CSF, and the whole brain (two variants: one using the median absolute deviation of the air background based on Dietrich et al. (2007), and one using the standard deviation of the noise within the tissue class)
- Contrast-to-noise ratio (CNR) between GM and WM
- Coefficient of variation (CV) of GM and WM
- Coefficient of joint variation (CJV) between GM and WM
- Entropy-focus criterion (EFC)
- Foreground-background energy ratio (FBER)
- WM to maximum intensity ratio (WM2MAX): the median intensity within WM over the 95% percentile of the full intensity distribution, that captures the existence of long tails due to hyper-intensity of the carotid vessels and fat
- These metrics are saved in
MRS_struct.out.QA
- Added GitHub Actions for continuous integration with
bids-matlab,dicm2nii, andexport_figto ensure compatibility with the latest versions of these dependencies - Added
dependabot.ymlto automatically check for updates to GitHub Actions - Water-unfiltered (but aligned) spectra are now saved in
MRS_struct - All version strings (e.g.,
MRS_struct.version.*) are now stored underMRS_struct.info.version.*, and datetime stamps for processing steps are added asMRS_struct.info.datetime.*
Major changes
- When calculating GM, WM, and CSF fractions in
GannetSegment.m, the probabilistic tissue maps are now thresholded at 0.9 before calculating the fractions. This should improve the accuracy of tissue fraction estimates
Minor changes
- Modified
SiemensTWIXRead.mto use iceParam forpointsBeforeEchoin MEGA-sLASER, aligning with MEGA-PRESS handling for consistency - Changed the naming convention for water residual fit results from
ResidWatertoresid_waterinGannetFit.mfor consistency - Improved input argument validation and error messages, including checks for cell array inputs and file existence, with clearer error output for missing files
- Updated
export_figto v3.52 - Updated
dicm2niito v2025.08.21 - MANY COSMETIC CHANGES to improve code readability and maintainability
Full Changelog: v3.4.0...v3.5.0