Reducing inter-subject anatomical variation: Effect of normalization method on sensitivity of functional magnetic resonance imaging data analysis in auditory cortex and the superior temporal region

Abstract

Conventional group analysis of functional MRI (fMRI) data usually involves spatial alignment of anatomy across participants by registering every brain image to an anatomical reference image. Due to the high degree of inter-subject anatomical variability, a low-resolution average anatomical model is typically used as the target template, and/or smoothing kernels are applied to the fMRI data to increase the overlap among subjects’ image data. However, such smoothing can make it difficult to resolve small regions such as subregions of auditory cortex when anatomical morphology varies among subjects. Here, we use data from an auditory fMRI study to show that using a high-dimensional registration technique (HAMMER) results in an enhanced functional signal-to-noise ratio (fSNR) for functional data analysis within auditory regions, with more localized activation patterns. The technique is validated against DARTEL, a high-dimensional diffeomorphic registration, as well as against commonly used low-dimensional normalization techniques such as the techniques provided with SPM2 (cosine basis functions) and SPM5 (unified segmentation) software packages. We also systematically examine how spatial resolution of the template image and spatial smoothing of the functional data affect the results. Only the high-dimensional technique (HAMMER) appears to be able to capitalize on the excellent anatomical resolution of a single-subject reference template, and, as expected, smoothing increased fSNR, but at the cost of spatial resolution. In general, results demonstrate significant improvement in fSNR using HAMMER compared to analysis after normalization using DARTEL, or conventional normalization such as cosine basis function and unified segmentation in SPM, with more precisely localized activation foci, at least for activation in the region of auditory cortex.

Local landmark alignment for high-resolution fMRI group studies: Toward a fine cortical investigation of hand movements in human

“…Indeed, the slight improvement in CS realignment obtained with DiDa has a significant positive effect on group statistics (Figures 7 and 8). This confirms that any change in deformation has an impact on functional group analysis, as reported in (Viceic et al, 2009) or (Tahmasebi et al, 2009(Tahmasebi et al, , 2012 for auditory cortex. Note that effective smoothness as provided by SPM8 was not statistically different between the registration methods used (average volumetric resolution = 3.2, σ =0.2).…”

“…We used a high spatial resolution after smoothing equal to 1.5x1.5x1.5mm 3 . As expected, a larger resolution (3x3x3mm 3 ), after smoothing, comparable to image resolution (3.3x3.3x4mm 3 ) used in Tahmasebi et al (2009) improved the number of activated voxels (see Supplementary data, Figures 13 and 14) but hampered the separation of different foci of activation (see Supplementary data, Figures 11 and 12). In this case, the difference in activation detectability between DARTEL and Disco is no more significant (see Supplementary data, Figure 15).…”

“…However, the standard spatial smoothing (6-mm full width at half maximum) they used limited the accuracy of the assignment of activated clusters to precise anatomical regions. In (Tahmasebi et al, 2009) the effect of spatial smoothing was investigated in auditory cortex, showing that smoothing degrades the localization of the activation and increases the magnitude of the peak activation for all studied methods. We used a high spatial resolution after smoothing equal to 1.5x1.5x1.5mm 3 .…”

“…This bias is likely to occur with any registration strategy relying on the arbitrary choice of a reference. In (Tahmasebi et al, 2009), the authors concluded that HAMMER (Shen and Davatzikos, 2002) outperformed DARTEL in the auditory cortex but the sensitivity of the former to the chosen reference was not reported. An extension of DDe (or HAMMER) to iteratively compute a mean image as a template, similarly to DARTEL or DiDa, could be considered for the method improvement.…”

“…In all these studies a set of quantitative measures was used to compare the deformed structural MR source images to the target. The effects of the registration on brain activation detection were investigated, to the best of our knowledge, mainly in auditory cortex (Viceic et al, 2009;Tahmasebi et al, 2009Tahmasebi et al, , 2012. In this paper, in the context of motor cortex studies, we explore the effects of non-linear registration methods on the detection of statistically significant activated clusters and on their localization.…”