
UK BIOBANK BRAIN IMAGING - CROSS-SUBJECT AVERAGE TEMPLATES FOR MULTIPLE MODALITIES
30 January 2016 - Stephen Smith & Fidel Alfaro Almagro, FMRIB Oxford

All averaged images are in 1mm or 2mm MNI152 co-ordinate space
Averages are across approximately 4000-4500 subjects (depending on specific modality)
Subjects have average age 62y (range 45-78)
For more information see: http://biobank.ctsu.ox.ac.uk/crystal/docs/brain_mri.pdf

STRUCTURAL MRI
  T1 = T1-weighted whole-head image after FAST bias correction
  T1_preunbiased = T1 brain image without bias correction
  T2_FLAIR = T2-weighted FLAIR whole-head image after FAST bias correction
  T2_FLAIR_preunbiased = T2_FLAIR brain image without bias correction

SUSCEPTIBILITY-WEIGHTED IMAGING
  SWI = SWI "venogram" with SWI magnitude image modulated by filtered phase image
  T2star = T2*-weighted image

RESTING-STATE fMRI
  MNI152 standard-space group-ICA maps at two dimensionalities (detail of parcellation):
    rfMRI_ICA_d25.nii.gz
    rfMRI_ICA_d100.nii.gz
  List of group-ICA components that are not clearly artefactual
      (listed numbers refer to volume numbers in the above 4D group-ICA map files, with counting starting at 1):
    rfMRI_GoodComponents_d25_v1.txt
    rfMRI_GoodComponents_d100_v1.txt
  Group-averaged network-matrices (full and partial correlation):
    rfMRI_GroupMeanNetmat_d25_fullcorr_v1.txt
    rfMRI_GroupMeanNetmat_d25_partialcorr_v1.txt
    rfMRI_GroupMeanNetmat_d100_fullcorr_v1.txt
    rfMRI_GroupMeanNetmat_d100_partialcorr_v1.txt

TASK fMRI
  Group-average activation contrasts (fixed-effect z-statistics):
    tfMRI_zstat1_FE.nii.gz ("Shapes")
    tfMRI_zstat2_FE.nii.gz ("Faces")
    tfMRI_zstat5_FE.nii.gz ("Faces-Shapes")
  Group-average-based ROI masks:
    tfMRI_mask1.nii.gz  ("Shapes")
    tfMRI_mask2.nii.gz  ("Faces")
    tfMRI_mask5.nii.gz  ("Faces-Shapes")
    tfMRI_mask5a.nii.gz ("Faces-Shapes" in amygdala)

DIFFUSION MRI
  Tensor fit
    dti_FA.nii.gz = fractional anisotropy
    dti_MD.nii.gz = mean diffusivity
    dti_MO.nii.gz = tensor mode
    dti_L1.nii.gz dti_L2.nii.gz dti_L3.nii.gz = tensor eigenvalues
    dti_V1.nii.gz dti_V2.nii.gz dti_V3.nii.gz = tensor eigenvectors
  NODDI modelling
    NODDI_ICVF.nii.gz  = intra-cellular volume fraction
    NODDI_ISOVF.nii.gz = isotropic or free water volume fraction
    NODDI_OD.nii.gz    = orientation dispersion index
  Probabilistic tractography - voxelwise modelling
    dyads1.nii.gz dyads2.nii.gz dyads3.nii.gz = tract directions for fibre populations 1,2,3
    dyads1_thr.nii.gz dyads2_thr.nii.gz dyads3_thr.nii.gz = dyads after thresholding according to volume fraction (likelihood of that tract being present)
    dyads1_dispersion.nii.gz dyads2_dispersion.nii.gz dyads3_dispersion.nii.gz = uncertainties on tract directions
    mean_f1samples.nii.gz mean_f2samples.nii.gz mean_f3samples.nii.gz = volume fractions of fibre populations 1,2,3 ("strength")
    sum_f123samples.nii.gz = volume fraction sum across all 3 fibres
  Probabilistic tractography
    tracts/* = AutoPtx tract outputs for 27 separate tracts
    TractsSummed.nii.gz = summation across all 27 tracts

