Step 1. Registration of the MNI template to native T1 space

The atlas and region files are defined in Montreal Neurological Institute (MNI) space, whereas tractography is performed in each participant's diffusion space. Moving an MNI-space image into a participant therefore requires two transforms: a nonlinear warp from MNI to the participant's T1 image, computed in this step, and a linear transform from T1 to diffusion space, obtained during preprocessing and applied in Step 2.

A nonlinear registration is necessary because affine transforms account for global rotation, scaling and shear but not for individual differences in the size and position of subcortical structures. Positional errors of a few millimetres in the midbrain displace the ventral tegmental area (VTA) region outside the VTA.

Procedure

Registration is performed with antsRegistrationSyNQuick.sh (Avants et al., 2008), with the participant's skull-stripped T1 image as the fixed image and the skull-stripped MNI152 1 mm template as the moving image. Image coverage should match: registering a skull-stripped participant image to a whole-head template biases the solution.

antsRegistrationSyNQuick.sh -d 3 \
  -f "$PROJECT/anat/$s/${s}_T1w_brain.nii.gz" \
  -m "$FSLDIR/data/standard/MNI152_T1_1mm_brain.nii.gz" \
  -o "$OUT/$s/reg/mni2t1_" -n 4

The outputs are the affine component (mni2t1_0GenericAffine.mat), the forward warp from MNI to T1 (mni2t1_1Warp.nii.gz, used in Step 2), the inverse warp from T1 to MNI (mni2t1_1InverseWarp.nii.gz, used to return participant-level results to atlas space) and the template resampled into participant space (mni2t1_Warped.nii.gz).

If the T1 image has not been skull-stripped, this should be done first (Diffusion MRI Preprocessing); atlas construction used SynthStrip (Hoopes et al., 2022) and the example dataset used ANTs brain extraction. If the T1 image is already aligned to the diffusion image, this warp is the only transform required (T1_TO_DWI=header in the configuration).

The full script follows. Each participant requires approximately 10 to 15 min with four threads, and the script runs four participants concurrently.

Script 01_register_mni_to_t1.sh (52 lines)

01_register_mni_to_t1.sh
#!/bin/bash
# =============================================================================
# Step 1. ANTs SyN registration of the MNI152 template (moving) to each T1 (fixed)
# =============================================================================
# Step 2 applies the forward affine + warp to carry MNI-space seed, target and atlas into T1.
# In:  $PROJECT/anat/<subj>/<subj>_T1w_brain.nii.gz (skull-stripped), $MNI_TEMPLATE
# Out: $OUT/<subj>/reg/mni2t1_{0GenericAffine.mat,1Warp,1InverseWarp,Warped}; QC = Warped
# Run: bash 01_register_mni_to_t1.sh  (FORCE=1 redoes everyone). Hours; log in $OUT/logs/
# =============================================================================

source "$(dirname "$0")/00_config.sh"
start_log "$0"
read_subjects

# Register one participant; returns early if the T1 is missing or the warp already exists.
register_one() {
  local s=$1
  local t1="$PROJECT/anat/$s/${s}_T1w_brain.nii.gz"
  local d="$OUT/$s/reg"
  if [ ! -f "$t1" ]; then
    echo "!! $s missing $t1"
    return
  fi
  if [ "$FORCE" != 1 ] && [ -f "$d/mni2t1_1Warp.nii.gz" ]; then
    echo "== $s already registered (FORCE=1 to recompute)"
    return
  fi
  mkdir -p "$d"
  antsRegistrationSyNQuick.sh -d 3 \
    -f "$t1" \
    -m "$MNI_TEMPLATE" \
    -o "$d/mni2t1_" \
    -n "$ANTS_THREADS"
  echo ">> $s registered"
}

for s in "${SUBJECTS[@]}"; do
  register_one "$s" &
  throttle "$ANTS_JOBS"
done
# Not a bare "wait": that would also hang on the tee process start_log opened.
wait_for_jobs

# Audit. Step 2 skips any participant with a MISSING affine or warp, so fix those first.
printf "\nSubject\tAffine\tWarp\tInverseWarp\n"
for s in "${SUBJECTS[@]}"; do
  d="$OUT/$s/reg"
  printf "%s\t%s\t%s\t%s\n" "$s" \
    "$(present "$d/mni2t1_0GenericAffine.mat")" \
    "$(present "$d/mni2t1_1Warp.nii.gz")" \
    "$(present "$d/mni2t1_1InverseWarp.nii.gz")"
done

Verification

The script concludes with a per-participant table of the three transform files. Registration quality is assessed by overlaying mni2t1_Warped.nii.gz on the participant's T1 image; the ventricles, corpus callosum and brainstem outline should coincide. A registration that is globally acceptable but locally displaced in the midbrain will manifest in Step 2 as a misplaced VTA region, so this inspection supplements rather than replaces the region-level inspection described there.

Procedure​

Verification​

Procedure

Verification