anibalsolon/ica-aroma

A data-driven method to identify and remove motion-related independent components from functional MRI data.

ICA-AROMA

ICA-AROMA (ICA-based Automatic Removal Of Motion Artefacts) is a data-driven method to identify and remove motion-related independent components from fMRI data. To that end it exploits a small, but robust set of theoretically motivated features, avoiding the need for classifier re-training and thus providing direct and easy applicability. This version requires python (version 2.7 or 3.5 upwards) and an installation of FSL. Please see the manual for a description of how to run ICA-AROMA.

Change Log

Note that some earlier versions of the original ICA-AROMA scripts (v0.1-beta and v0.2-beta) contained a mistake at the denoising stage of the method that led to incorrect output. The issue was resolved in version v0.3-beta (27th of April 2015) onwards.

v0.3-beta to v0.4.0

• port to python 2/3
• general refactor
• speed up and simplify feature calculations
• add test harness (nose)
• remove restrictions due to irregular path handling
• use nibabel for accessing nifti file info
• move documentation to markdown
• replace fsl_regfilt with an internal routine
• allow installation as package via pip or directly as standalone script
• gnu standard short/long forms for command line arguments

0.2-beta to v0.3-beta

• Correct off by one error in the definition of the string of indices of the components to be removed by fsl_regfilt
• Take the maximum of the absolute value of the correlation between the component time-course and set of realignment parameters
• Correct for the fact that the defined frequency-range used for the high-frequency content feature, in few cases, did not include the final Nyquist frequency due to limited numerical precision

Dependencies

This is tested with python versions 2.7 and 3.5, specifically those distributed in the anaconda python distribution. It should also work with the system python (generally 2.7) on recent versions of Linux. In addition to a python installation the following is required:

• The FSL software suite.
• The numpy and nibabel python packages

These are available on ubuntu and debian systems via the neurodebian repository or may be installed explicitly via downloads and/or pip install.

Licensing

ICA-AROMA is distributed under the Apache licence.