Accurate and robust brain image alignment using boundary-based registration

Abstract

The fine spatial scales of the structures in the human brain represent an enormous challenge to the successful integration of information from different images for both within-and between-subject analysis. While many algorithms to register image pairs from the same subject exist, visual inspection shows that their accuracy and robustness to be suspect, particularly when there are strong intensity gradients and/or only part of the brain is imaged. This paper introduces a new algorithm called Boundary-Based Registration, or BBR. The novelty of BBR is that it treats the two images very differently. The reference image must be of sufficient resolution and quality to extract surfaces that separate tissue types. The input image is then aligned to the reference by maximizing the intensity gradient across tissue boundaries. Several lower quality images can be aligned through their alignment with the reference. Visual inspection and fMRI results show that BBR is more accurate than correlation ratio or normalized mutual information and is considerably more robust to even strong intensity inhomogeneities. BBR also excels at aligning partial brain images to whole brain images, a domain in which existing registration algorithms frequently fail. Even in the limit of registering a single slice, we show the BBR results to be robust and accurate.

Cortical pathology in multiple sclerosis detected by the T1/T2‐weighted ratio from routine magnetic resonance imaging

“…The T2w image was registered to the T1w image using FreeSurfer's bbregister, which performs within‐subject registration using a boundary‐based cost function 28. Given the cortical segmentation of one image (T1w), this routine processes images of different modalities and resolutions through resampling of the images by trilinear interpolation, resulting in images with the standard resolution of FreeSurfer (1 × 1 × 1mm 3 ).…”

Functional brain network centrality is related to APOE genotype in cognitively normal elderly

“…Time series were high‐pass filtered at a cutoff frequency of 182 s (100 TR). The RS‐fMRI data of each subject were mapped to the native‐space structural MR image using boundary‐based registration (Greve & Fischl, 2009), after which the standard space‐mapping parameters of the structural image were used to map them to MNI standard space at a sampling resolution of 4 mm isotropic. Of the initially selected study of 269 subjects with fMRI data, eight had to be discarded due to bad image quality (excessive motion, e.g., too high FD values as evaluated and reported by the preprocessing software (mean displacement > 0.5 mm), missing data and/or failed registration to the anatomical scans), leaving 261 preprocessed fMRI data sets.…”

Differential patterns of age‐related cortical and subcortical functional connectivity in 6‐to‐10 year old children: A connectome‐wide association study

“…Details about the FreeSurfer data processing and quality control in the Generation R Study are described elsewhere (Mous et al., 2014). The FreeSurfer image, including the cortical and subcortical labels were registered to the rs‐fMRI data by applying the transformation matrix resulting from a 12 degree of freedom affine registration of the T 1 ‐weighted image to the rs‐fMRI data (Greve & Fischl, 2009). Thus, all time‐series for analyses were extracted from native fMRI space.…”