Align and Bias Correct Your Brain!

Once you’re files are ready to by preprocessed (Yes, there’s a word for the processing that takes place before the processing). They need to be aligned to a common image space.
Here’s my handy script:

#!/bin/bash

#SBATCH --time=30:00
#SBATCH --ntasks=1
#SBATCH --nodes=1
#SBATCH --mem-per-cpu=8192M
#SBATCH -o /fslhome/zach8769/logfiles/oasis_longitudinal/output_prepT1_0.txt
#SBATCH -e /fslhome/zach8769/logfiles/oasis_longitudinal/error_prepT1_0.txt
#SBATCH -J "prepT1_0"
#SBATCH --mail-user=GalileoHumpkins@gmail.com
#SBATCH --mail-type=BEGIN
#SBATCH --mail-type=END
#SBATCH --mail-type=FAIL

export PBS_NODEFILE=`/fslapps/fslutils/generate_pbs_nodefile`
export PBS_JOBID=$SLURM_JOB_ID
export PBS_O_WORKDIR="$SLURM_SUBMIT_DIR"
export PBS_QUEUE=batch
export OMP_NUM_THREADS=$SLURM_CPUS_ON_NODE

ROOT=/fslhome/GalileoHumpkins/
DATASET=${ROOT}compute/oasis_longitudinal/
TEMPLATE=${ROOT}compute/templates/MICCAI2012-Multi-Atlas-Challenge-Data/
export ANTSPATH=/fslhome/zach8769/bin/antsbin/bin/
PATH=${ANTSPATH}:${PATH}

IFS=$'\n'
array=( $(find ${DATASET}/*/t1/*/ -type f -name t1.nii.gz) )
for i in 0;
  do
  SUB=$(dirname ${array[$i]})

  ${ANTSPATH}antsRegistrationSyNQuick.sh -d 3 \
    -f ${TEMPLATE}T_template0.nii.gz \
    -m ${SUB}/t1.nii.gz \
    -o ${SUB}/rigid_ \
    -t r

  N4BiasFieldCorrection -d 3 \
    -i ${SUB}/rigid_Warped.nii.gz \
    -o ${SUB}/n4.nii.gz \
    -s 8 -b [200] -c [50x50x50x50,0.000001]
    
  N4BiasFieldCorrection -d 3 \
    -i ${SUB}/n4.nii.gz \
    -o ${SUB}/n4.nii.gz \
    -s 4 -b [200] -c [50x50x50x50,0.000001]

  N4BiasFieldCorrection -d 3 \
    -i ${SUB}/n4.nii.gz \
    -o ${SUB}/n4.nii.gz \
    -s 2 -b [200] -c [50x50x50x50,0.000001]

  ResampleImage 3 \
    ${SUB}/n4.nii.gz \
    ${SUB}/n4_resliced.nii.gz 1x1x1

  rm ${SUB}/rigid_0GenericAffine.mat
  rm ${SUB}/rigid_InverseWarped.nii.gz
done

Th N4BiasFieldCorrection command accounts for different intensities in the scan that result from bias in the signal as a result of their location in 3 dimensions (there are a lot more details but they’re beyond the scope of this post). The ResampleImage command ensures that all images have the same voxel spacing. As a whole, this script ensures that image spacing and intensity is as comparable as possible before any further processes that compare anatomical differences are used.

After

Before