Skip to content

Structural MRI

FSL contains a range of tools for working with and segmenting structural MRI images.

  • BET can be used to reliably perform brain extraction (a.k.a. skull-stripping) on images of any modality.
  • FAST can be used to segment a brain image into different tissue types (usually grey matter, white matter, and cerebrospinal fluid).
  • FIRST and MIST can be used to generate segmentations of sub-cortical structures.
  • BIANCA and TRUENET identify white matter hyperintensities (WMHs) or small white matter lesions in MRI images, and can work with a range of modalities including T1, T2 and T2-FLAIR.

The fsl_anat script incorporates several of the above tools into a general pipeline for processing anatomical images.

A range of other tools for analysing and working with structural image data are available within FSL:

  • FSL-VBM can be used to compare voxelwise differences in the local grey matter related to (e.g.) one clinical score, or across populations.
  • SIENA and SIENAX can be used for single-time-point ("cross-sectional") and two-time-point ("longitudinal") analysis of brain change from T1 images, in particular the estimation of atrophy (volumetric loss of brain tissue).
  • lesion filling can be used to fill in small white matter pathologies (as already defined by a mask) with intensities that match surrounding healthy tissue, which is helpful to allow other structural tools to process the image as if no pathologies were present, since this often improves performance for registration and segmentation.