Luders, E. | bioRxiv Science

bioRxiv2022

CAT - A Computational Anatomy Toolbox for the Analysis of Structural MRI Data

A large range of sophisticated brain image analysis tools have been developed by the neuroscience community, greatly advancing the field of human brain mapping. Here we introduce the Computational Anatomy Toolbox (CAT) - a powerful suite of tools for brain morphometric analyses with an intuitive graphical user interface, but also usable as a shell script. CAT is suitable for beginners, casual users, experts, and developers alike providing a comprehensive set of analysis options, workflows, and integrated pipelines. The available analysis streams - illustrated on an example dataset - allow for voxel-based, surface-based, as well as region-based morphometric analyses. Notably, CAT incorporates multiple quality control options and covers the entire analysis workflow, including the preprocessing of cross-sectional and longitudinal data, statistical analysis, and the visualization of results. The overarching aim of this article is to provide a complete description and evaluation of CAT, while offering a citable standard for the neuroscience community.

neuroscience↗

bioRxiv2022

To smooth or not to smooth: One step closer to single-voxel accuracy without spatial smoothing

Traditionally, when conducting voxel- or vertex-wise analyses in neuroimaging studies, it seemed imperative that brain data are convoluted with a Gaussian kernel, a procedure known as "spatial smoothing". However, we suggest that - under certain conditions - smoothing may be omitted for the benefit of an improved regional specificity. We demonstrate the suitability of this omission by combining high-dimensional spatial registration and threshold-free cluster enhancement (TFCE) in a sample of 754 brains. Our findings revealed that, without smoothing, it is possible to capture brain atrophy within the hippocampal complex while dissociating neighboring areas (cornu ammonis, dentate gyrys, subiculum, and amygdala). In contrast, the traditional smoothing step would result in a single hippocampal cluster (the larger the smoothing kernel, the lower the specificity). Supplemental analyses not only varying the size of the smoothing kernel, but also the size of the sample, the signal-to-noise ratio, as well as the accuracy of the spatial registration confirm that no smoothing (or less smoothing) leads to increased specificity while maintaining sensitivity, at least for small-scale structures (e.g., hippocampus and amygdala). Nevertheless, classic analyses based on smoothed data will continue to provide important insights, especially for large-scale structures (e.g., cortical regions).

neuroscience↗