Distinctions among real and apparent respiratory motions in human fMRI data
Jonathan Power, Charles Lynch, Benjamin Silver, Marc Dubin, Alex Martin, Rebecca Jones
Neuroimage 2019 Nov 1; 201:116041
Pubmed link
Figures (pptx)
Below are movies and files associated with the article "Distinctions among real and apparent respiratory motions in human fMRI data". All movies are 1080p and will look best at full-screen resolutions. The movies stream from YouTube. Click on links to download movies or files. Files to download are hosted on Dropbox Pro; if the links don't work our traffic has exceeded its 200GB/day limit and the links will be re-enabled the following day.
Data quality check:
A movie of the physiology traces of the 900-subject release Human Connectome Project (HCP) and decisions about whether data quality was sufficient for use was part of the supplemental material of another paper and can be found here. These traces and decisions are the basis of selecting 440 of the 900 subjects for analysis.
Movie 1: (.mov - 136 MB)
This movie, for each subject, shows the 4 runs of respiratory belt traces, the corresponding X (left-right) position traces, and the power spectra of both traces with the peak respiratory position shown. The peaks are algorithmically determined and are occassionally not the desired respiratory peaks; analyses of the paper excluded and/or fixed these peaks, but they are retained here to show what an unsupervised peak-finder will identify.
Movie 2: (.mov - 6 MB)
This movie shows centers of mass from 10 MyConnectome scans and 10 HCP subjects (40 scans). Unprocessed images were analyzed for centers of mass at each volume, yielding X, Y, and Z axis coordinates, plotted below. The MyConnectome data do not have accompanying respiratory records; respiratory records for HCP scans are shown at the bottom. Units are voxels and scales are constant across each dataset. Ongoing respiratory fluctuations are seen in all 3 axes in the HCP data, always in Z and X and usually in Y. In MyConnectome scans, fluctuations are always in Z, often in Y, and infrequently in X.
Movie 3: (.mov - 1.9 GB)
This movie shows gray plots illustrating the revised motion estimates of the paper. Top panel has realignment parameters in gray, in bold red the FDfiltered,4-TR, and in the thin light red line FDfiltered,1-TR. Second panel has in light green DVARS in minimally processed data, in darker green DVARS in FIX-ICA-denoised data. Third panel has the respiratory belt trace in blue, and then ENV in red, RV in blue, and RVT in black. Bottom panel shows all in-brain voxel timeseries on a -2 to 2% grayscale, with voxels ordered vertically by FreeSurfer-defined anatomical compartment. The bright green line separates gray matter (above) from white matter (below).