Unified voxel- and tensor-based morphometry (UVTBM) using registration confidence

Abstract

Keywords: MRI Brain registration Morphometry Diffeomorphisms Atlases Voxel-based morphometry Tensor-based morphometry a b s t r a c t Voxel-based morphometry (VBM) and tensor-based morphometry (TBM) both rely on spatial normalization to a template and yet have different requirements for the level of registration accuracy. VBM requires only global alignment of brain structures, with limited degrees of freedom in transformation, whereas TBM performs best when the registration is highly deformable and can achieve higher registration accuracy. In addition, the registration accuracy varies over the whole brain, with higher accuracy typically observed in subcortical areas and lower accuracy seen in cortical areas. Hence, even the determinant of Jacobian of registration maps is spatially varying in their accuracy, and combining these with VBM by direct multiplication introduces errors in VBM maps where the registration is inaccurate. We propose a unified approach to combining these 2 morphometry methods that is motivated by these differing requirements for registration and our interest in harnessing the advantages of both. Our novel method uses local estimates of registration confidence to determine how to weight the influence of VBM-and TBM-like approaches. Results are shown on healthy and mild Alzheimer's subjects (N ¼ 150) investigating age and group differences, and potential of differential diagnosis is shown on a set of Alzheimer's disease (N ¼ 34) and frontotemporal dementia (N ¼ 30) patients compared against controls (N ¼ 14). These show that the group differences detected by our proposed approach are more descriptive than those detected from VBM, Jacobian-modulated VBM, and TBM separately, hence leveraging the advantages of both approaches in a unified framework.

Functional Activation-Informed Structural Changes during Stroke Recovery: A Longitudinal MRI Study

“…Tensor-based morphometry (TBM) approach could outperform the VBM in some situations. However, according to previous studies, TBM, which is a high degree of freedom method, relies on a high degree of registration accuracy and may be underpowered in situations where registration accuracy is lower [ 63 , 64 ]. Our study is a disease-specific analysis, which results in very different spatial atrophy profiles due to lesions, and, thus, increases the uncertainty in registration.…”

“…Our study is a disease-specific analysis, which results in very different spatial atrophy profiles due to lesions, and, thus, increases the uncertainty in registration. A VBM approach would be suitable in the case with more uncertainty in registration [ 64 ]. Therefore, we used a VBM approach rather than a TBM approach in this study.…”

A Systematic Characterization of Structural Brain Changes in Schizophrenia

“…It should be noted that, although we further divided the TBM results into GM and WM by using GM and WM masks for cortical surface visualization and better comparison between TBM and VBM, this GM-WM division did not apply at the voxel level. As the Jacobian matrix of each displacement vector maps a point in the template image to the corresponding point in the individual image, it may be more prone to registration errors than VBM [43]. Thus, it has been suggested that VBM is more sensitive and powerful for detecting local anatomical differences and TBM contains more noise in regions where there is more uncertainty in registration, while TBM contains more information and would be preferred in regions where there is less uncertainty in registration [24,[44][45][46].…”

Systems Biology, Neuroimaging, Neuropsychology, Neuroconnectivity and Traumatic Brain Injury

“…From this kind of volumetric approach using automated image analysis techniques, several 100 brain ROIs can be derived, but the individual compilation of such measures becomes cumbersome when attempting to visualize where traumatic changes have occurred. Another approach is to use 3-D techniques that warp the individual brain to a normative template where a Jacobian-modulated extraction of how the patient brain has to be altered to fit the template graphically will display where differences reside in the individual patient compared to the control sample (Avants et al, 2011a , b ; Tustison et al, 2014 ; Khan et al, 2015 ). Using such an approach if there is substantial loss of parenchymal volume in the patient, the warping to fit the normative standard will depict volume reductions.…”