IQMs for functional images

Measures for the structural information

Definitions are given in the summary of structural IQMs.

Measures for the temporal information

  • DVARS - D referring to temporal derivative of timecourses, VARS referring to RMS variance over voxels ([Power2012] dvars_nstd) indexes the rate of change of BOLD signal across the entire brain at each frame of data. DVARS is calculated with nipype after motion correction:

    \[\text{DVARS}_t = \sqrt{\frac{1}{N}\sum_i \left[x_{i,t} - x_{i,t-1}\right]^2}\]


    Intensities are scaled to 1000 leading to the units being expressed in x10 \(\%\Delta\text{BOLD}\) change.


    MRIQC calculates two additional standardized values of the DVARS. The dvars_std metric is normalized with the standard deviation of the temporal difference time series. The dvars_vstd is a voxel-wise standardization of DVARS, where the temporal difference time series is normalized across time by that voxel standard deviation across time, before computing the RMS of the temporal difference [Nichols2013].

  • Global Correlation (gcor(), gcor) calculates an optimized summary of time-series correlation as in [Saad2013]:

    \[\text{GCOR} = \frac{1}{N}\mathbf{g}_u^T\mathbf{g}_u\]

    where \(\mathbf{g}_u\) is the average of all unit-variance time series in a \(T\) (# timepoints) \(\times\) \(N\) (# voxels) matrix.

  • Temporal SNR (tSNR, tsnr) is a simplified interpretation of the tSNR definition [Kruger2001]. We report the median value of the tSNR map calculated like:

    \[\text{tSNR} = \frac{\langle S \rangle_t}{\sigma_t},\]

    where \(\langle S \rangle_t\) is the average BOLD signal (across time), and \(\sigma_t\) is the corresponding temporal standard-deviation map.

Measures for artifacts and other

  • Framewise Displacement (mean_fd): expresses instantaneous head-motion. Rotational displacements are calculated as the displacement on the surface of a sphere of radius 50 mm [Power2012]:

    \[\text{FD}_t = |\Delta d_{x,t}| + |\Delta d_{y,t}| + |\Delta d_{z,t}| + |\Delta \alpha_t| + |\Delta \beta_t| + |\Delta \gamma_t|\]

    Along with the base framewise displacement, MRIQC reports the number of timepoints above FD threshold (num_fd), and the percent of FDs above the FD threshold w.r.t. the full timeseries. In both cases, the threshold is set at 0.20mm.

  • Ghost to Signal Ratio (gsr(), ghost_*): along the two possible phase-encoding axes x, y:

    \[\text{GSR} = \frac{\mu_G - \mu_{NG}}{\mu_S}\]
  • Outlier fraction (outlier) - Mean fraction of outliers per fMRI volume as given by AFNI’s 3dToutcount.

  • Quality index (quality) - Mean quality index as computed by AFNI’s 3dTqual.


[Atkinson1997]Atkinson et al., Automatic correction of motion artifacts in magnetic resonance images using an entropy focus criterion, IEEE Trans Med Imag 16(6):903-910, 1997. doi:10.1109/42.650886.
[Friedman2008]Friedman, L et al., Test–retest and between‐site reliability in a multicenter fMRI study. Hum Brain Mapp, 29(8):958–972, 2008. doi:10.1002/hbm.20440.
[Giannelli2010]Giannelli et al., Characterization of Nyquist ghost in EPI-fMRI acquisition sequences implemented on two clinical 1.5 T MR scanner systems: effect of readout bandwidth and echo spacing. J App Clin Med Phy, 11(4). 2010. doi:10.1120/jacmp.v11i4.3237.
[Jenkinson2002]Jenkinson et al., Improved Optimisation for the Robust and Accurate Linear Registration and Motion Correction of Brain Images. NeuroImage, 17(2), 825-841, 2002. doi:10.1006/nimg.2002.1132.
[Kruger2001]Krüger et al., Physiological noise in oxygenation-sensitive magnetic resonance imaging, Magn. Reson. Med. 46(4):631-637, 2001. doi:10.1002/mrm.1240.
[Nichols2013]Nichols, Notes on Creating a Standardized Version of DVARS, 2013.
[Power2012](1, 2) Power et al., Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion, NeuroImage 59(3):2142-2154, 2012, doi:10.1016/j.neuroimage.2011.10.018.
[Saad2013](1, 2) Saad et al. Correcting Brain-Wide Correlation Differences in Resting-State FMRI, Brain Conn 3(4):339-352, 2013, doi:10.1089/brain.2013.0156.

mriqc.qc.functional module

mriqc.qc.functional.gcor(func, mask=None)[source]

Compute the GCOR [Saad2013].

  • func (numpy.ndarray) – input fMRI dataset, after motion correction
  • mask (numpy.ndarray) – 3D brain mask

the computed GCOR value

mriqc.qc.functional.gsr(epi_data, mask, direction='y', ref_file=None, out_file=None)[source]

Computes the GSR [Giannelli2010]. The procedure is as follows:

  1. Create a Nyquist ghost mask by circle-shifting the original mask by \(N/2\).
  2. Rotate by \(N/2\)
  3. Remove the intersection with the original mask
  4. Generate a non-ghost background
  5. Calculate the GSR


This should be used with EPI images for which the phase encoding direction is known.

  • epi_file (str) – path to epi file
  • mask_file (str) – path to brain mask
  • direction (str) – the direction of phase encoding (x, y, all)

the computed gsr