Note
This page is a reference documentation. It only explains the function signature, and not how to use it. Please refer to the user guide for the big picture.
7.5.5. nilearn.image.high_variance_confounds¶

nilearn.image.
high_variance_confounds
(imgs, n_confounds=5, percentile=2.0, detrend=True, mask_img=None)¶ Return confounds signals extracted from input signals with highest variance.
Parameters: imgs: Niimglike object
mask_img: Niimglike object
See http://nilearn.github.io/manipulating_images/input_output.html If provided, confounds are extracted from voxels inside the mask. If not provided, all voxels are used.
n_confounds: int
Number of confounds to return
percentile: float
Highestvariance signals percentile to keep before computing the singular value decomposition, 0. <= percentile <= 100. mask_img.sum() * percentile / 100. must be greater than n_confounds.
detrend: bool
If True, detrend signals before processing.
Returns: v: numpy.ndarray
highest variance confounds. Shape: (number of scans, n_confounds)
Notes
This method is related to what has been published in the literature as ‘CompCor’ (Behzadi NeuroImage 2007).
The implemented algorithm does the following:
 compute sum of squares for each signals (no mean removal)
 keep a given percentile of signals with highest variance (percentile)
 compute an svd of the extracted signals
 return a given number (n_confounds) of signals from the svd with highest singular values.