A practical modification to a resting state fMRI protocol for improved characterization of cerebrovascular function (https://doi.org/10.1016/j.neuroimage.2021.118306)
The breathing task code can be found in this repo: https://github.com/RayStick/BreathingTasks_PsychoPy. The settings used for the breathing task code are listed below.
scan_trigger = 5; doRest = 2; tResting = 480 ; trialnum = 3; tPace = 24 ; tBreathPace = 6; tHold = 15; tExhale = 2; tRecover = 4; BH_instructions = 'BREATH-HOLD task \n \nFollow the breathing instructions \n \nBreathe through your nose'; end_exp_key = 'escape'
scan_trigger = 5; doRest = 2; tResting = 480; trialnum = 2; tStartRest = 28; tGetReady = 2; tCDB = 8; tCDBPace = 4; tFree = 43; CDB_instructions = 'DEEP BREATHING task \n \nTake deep breaths IN and OUT when cued \n \nBreathe through your nose'; end_exp_key = 'escape'
RestDuration = 480; end_exp_key = 'escape'
PLEASE NOTE: the analysis code in this repository has been shared alongside the peer-reviewed manuscript in order to be transparent with what analysis was performed to produce the results. The code in this repository is not written as a package that is generalizable to different study set-ups, and unfortunately we do not have the permission to share the data. If you use this code, or parts of it, an acknowledgment would be welcome. You are welcome to use and edit it for your data set-up. If you are looking for citable fMRI pre-processing pipelines check out something like fMRIprep (https://github.com/nipreps/fmriprep). For a package that processes physiological data, to prepare them for analysis with fMRI, see the physiopy community github (https://github.com/physiopy).
Used the default options (except for input and output directories).
Ran three times for each data segment.
inputfile_txt = physiological text file from the whole scan session
time_col = 1
vol_col = 2
sf = 1000
TR = 1.2
n_TRs = 534 (BH+REST) or 513 (CDB+REST) or 400 (REST)
prefix = subjectID_BHREST or subjectID_CDBREST or subjectID_REST
extra_TRs = 17
inputfile_json = 0
Ran three times for each data segment.
filename = output from Step 2
fs = 1000
vol_col = 1
CO2_col = 2
prefix = subjectID_BHREST or subjectID_CDBREST or subjectID_REST
HRFconv= 1
output_hires = 1
td = 0
demean = 0
peak_check = 1
fMRI_file = input is the output from Step 1. CO2_file = input is the output from Step 3.
Ran in five different ways to make five data segments of length 7 mins 50 seconds:
-
Filename in 'BH+REST'; filename out 'BH+REST';
cut_TRs_start=10;
cut_TRs_end=134; -
Filename in 'CDB+REST'; filename out 'CDB+REST';
cut_TRs_start=10;
cut_TRs_end=113; -
Filename in 'BH+REST'; filename out 'rest_bh';
cut_TRs_start=144;
cut_TRs_end=0; -
Filename in 'CDB+REST'; filename out 'rest_cdb';
cut_TRs_start=123;
cut_TRs_end=0; -
Filename in 'rest'; filename out 'rest';
cut_TRs_start=10;
cut_TRs_end=0;
Remaining inputs (for the main analysis):
TR = 1.2
extra_TRs_half = 17
sf = 1000
nf = 40
bulk_shift = 0
save_bulk_hires = 0
fine_shift = 1
save_fine_hires = 0
shift_unit_s = 0.3
shift_max_s = 15
prefix = subjectID_BHREST_BH_CO2
Also ran with bulk_shift=1 and fine_shift=0 to explore the bulk shift (Fig4).
Ran for each data segment, using the same start and end time defined in Step 4.
Input 1 = Output directory of PreProc_MRI
Input 2 = Output file from PreProc_MRI
Input 3 = Output file from PreProc_MRI
Input 4 = Output file from PreProc_MRI
Input 5 = Output file from Shift_hires_CO2regressor
Input 6 = Output file from Shift_hires_CO2regressor (demeaned with 1d_tool.py)
Input 7 = 100
Input 8 = 0.3
Input 9 = 1
Inputs 10 and 11 = Same start and end times specified for the 5 segments in Step 4
Inputs 11 = 1.2
(x.3dDeconvolve_shifts was not run, but is an alternative to x.3dREMLfit_shifts)
Ran the code for each each data segment.
Input 1 = Output directory (Concat_3dREMLfit) from Step 5
Input 2 = File prefix used for the output files from Step 5
Input 3 = Output file from PreProc_MRI
Input 4 = Output file from PreProc_MRI
Input 5 = 5.1280e-04
Input 6 = 390
Thanks goes to these wonderful people (emoji key):
Rachael Stickland 💻 🤔 📖 |
Kristina Zvolanek 🤔 👀 |
Molly Bright 🤔 |
Stefano Moia 🤔 |
Cesar Caballero Gaudes 🤔 |
This project follows the all-contributors specification. Contributions of any kind welcome!