fMRI Timeseries Extraction

doi  10.25663/brainlife.app.369
   Registered  5/29/2020    103  Users    18853  Requests    1 - 2 min

fmriprep outputs to connectivity matrices

49.1% Success Rate


This App uses the following input/output datatypes

Input
fmri
Validator
func/task

from fMRIPrep

fMRI Task

fmri: bold.nii.gz • events: events.tsv • events_json: events.json • sbref: sbref.nii.gz • sbref_json: sbref.json • physio: physio.tsv.gz • physio_json: physio.json
parc
parcellation/volume

From multi-atlas transfer app; should be aligned to preprocessed bold.nii.gz (i.e. same space). The internal functions of this app will re-slice the func and the parc to be on the same voxel grid.

(aka "segmentation") parc.nii.gz containing volumetric parcellation / atlas, along with a key.txt file indicating the coorespondance between parc.nii.gz labels and human readable region names.

parc: parc.nii.gz • key: key.txt • label: label.json
confounds
regressors

from fMRIPrep

Confounds (or nuisance regressors) are variables representing fluctuations with a potential non-neuronal origin. Such non-neuronal fluctuations may drive spurious results in fMRI data analysis, including standard activation GLM and functional It is possible to minimize confounding effects of non-neuronal signals by including them as nuisance regressors in the GLM design matrix or regressing them out from the fMRI data - a procedure known as denoising. connectivity analyses.

confounds: regressors.tsv • json: regressors.json
mask
Validator
mask
func

from fMRIPrep

Volume Mask

mask: mask.nii.gz
Output
output_regress
Validator
func/task
+ all fmri tags

regressed output bold

fMRI Task

bold: bold.nii.gz • events: events.tsv • events_json: events.json • sbref: sbref.nii.gz • sbref_json: sbref.json • physio: physio.tsv.gz • physio_json: physio.json
out_ts
Validator
timeseries
time series

new timeseries datatype

A time series is a chronologically ordered series of numeric values. Time series will generally be stored as tables, with a row of column headers indicating the name of the series.

tsv: timeseries.tsv.gz • json: timeseries.json

Configuration
strategy: enum = 36P
  • 36P Satterthwaite 36 parameter strategy Default
  • 24aCompCor Anatomical CompCor strategy
  • globalsig Just global signal regression, nothing else
  • linear Just linear de-trending
  • 24aCompCorGsr Anatomical CompCor strategy with global signal

Strategy of nuisance regression, all strategies include linear de-trending

tr?: number

If you know the TR, you can set it here. Script will also try to automatically read it from BL info (try this first).

smoothfwhm?: number

FWHM of spatial smoothing kernel applied to bold.nii.gz (put 0 for no smoothing)

lowpass?: number = 0.08

Low-pass temporal filter

highpass?: number = 0.008

High-pass temporal filter

addlinear: boolean = true

Add linear regressor to nuisance regression. This makes your time series generally centered around 0.

spikethresh: number = 9999

If you want a spike regressor for frames that exceed a certain FD threshold, set that threshold here. Threshold is typically in the 0.5 to 1 range; else, "9999" means no spike regressor.

inspace: enum = data
  • data compute time series from data grid, resample labels Default
  • label computer time series from label grid, resample data

In almost ALL cases, this should be set to 'data'. It will take a super long time to run the app if you set it to 'label' and you might run out of memory.

savets: boolean = true

nomatrix: boolean = true

Maintainers

List of users who currently maintains this App.

Josh Faskowitz


Contributors

List of code contributors.(faskowit/app-fmri-2-mat).

Josh Faskowitz