ME-QC

Quality control metrics for data processed using Prantik Kundu's Multi-Echo Independent Components Analysis (ME-ICA).

Usage

ME-ICA minimally requires (1) acquired echo times (in milliseconds) and (2) functional datasets equal to the number of acquired echoes. But you can supply many other options, viewable with meica.py -h.

Here's one example use case:

meica.py -d 'sub-001_task-rest_echo-[1,2,3]_run-01_meep1.nii.gz' -e 15,30,45 -b 12s -a 'sub-001_T1w.nii.gz' --MNI

Where:

-e 15,30,45 are the echo times in milliseconds
-d 'sub-001_task-rest_echo-[1,2,3]_run-01_meep1.nii.gz' are the 4-D multi-echo fMRI datasets. Can supply each dataset individually or with bash shell expansion
-a 'sub-001_T1w.nii.gz' is an anatomical image with skull
-b 12s means drop first 12 seconds of data for equilibration
--MNI warp anatomical to MNI space using AFNI's MNI_caez_N27 (Colin 27) template.

Additional, optional parameters support situations such as: anatomical with no skull, applying FWHM smoothing, non-linear warping, etc.

Procedure

PROCEDURE 1 : Preprocess multi-echo datasets and apply multi-echo ICA based on spatial concatenation -Check arguments, input filenames, and filesystem for dependencies -Calculation of motion parameters based on images with highest constrast -Application of motion correction and coregistration parameters -Misc. EPI preprocessing (temporal alignment, smoothing, etc) in appropriate order -Compute PCA and ICA in conjuction with TE-dependence analysis

Output

• *_medn.nii.gz : 'Denoised' BOLD time series after: basic preprocessing, T2* weighted averaging of echoes (i.e. 'optimal combination'), ICA denoising. Use this dataset for task analysis and resting state time series correlation analysis.
• *_tsoc.nii.gz : 'Raw' BOLD time series dataset after: basic preprocessing and T2* weighted averaging of echoes (i.e. 'optimal combination'). 'Standard' denoising or task analyses can be assessed on this dataset (e.g. motion regression, physio correction, scrubbing, blah...) for comparison to ME-ICA denoising.
• *_mefc.nii.gz : Component maps (in units of \delta S) of accepted BOLD ICA components. Use this dataset for ME-ICR seed-based connectivity analysis.
• *_mefl.nii.gz : Component maps (in units of \delta S) of ALL ICA components.
• *_ctab.nii.gz : Table of component Kappa, Rho, and variance explained values, plus listing of component classifications.

Some Notes

• Make sure your datasets have slice timing information in the header. If not sure, specify a --tpattern option to meica.py. Check AFNI documentation of 3dTshift to see slice timing codes.
• FWHM smoothing is not recommended. tSNR boost is provided by optimal combination of echoes. For better overlap of 'blobs' across subjects, use non-linear standard space normalization instead with meica.py ... --qwarp

Citations

If you use ME-ICA in publications, please cite:

Kundu, P., Brenowitz, N.D., Voon, V., Worbe, Y., Vertes, P.E., Inati, S.J., Saad, Z.S., Bandettini, P.A. & Bullmore, E.T. Integrated strategy for improving functional connectivity mapping using multiecho fMRI. PNAS (2013). http://dx.doi.org/10.1073/pnas.1301725110

Kundu, P., Inati, S.J., Evans, J.W., Luh, W.M. & Bandettini, P.A. Differentiating BOLD and non-BOLD signals in fMRI time series using multi-echo EPI. NeuroImage (2011). http://dx.doi.org/10.1016/j.neuroimage.2011.12.028