Abstract
The corpus callosum includes the majority of fibers that connect the two cortical hemispheres. Studies of cross-sectional callosal morphometry and area have revealed developmental, gender, and hemispheric differences in healthy populations and callosal deficits associated with neurodegenerative disease and brain injury. However, accurate quantification of the callosum using magnetic resonance imaging is complicated by intersubject variability in callosal size, shape, and location and often requires manual outlining of the callosum in order to achieve adequate performance. Here we describe an objective, fully automated protocol that utilizes voxel-based images to quantify the area and thickness both of the entire callosum and of different callosal compartments. We verify the method's accuracy, reliability, robustness, and multisite consistency and make comparisons with manual measurements using public brain-image databases. An analysis of age-related changes in the callosum showed increases in length and reductions in thickness and area with age. A comparison of older subjects with and without mild dementia revealed that reductions in anterior callosal area independently predicted poorer cognitive performance after factoring out Mini-Mental Status Examination scores and normalized whole brain volume. Open-source software implementing the algorithm is available at www.nitrc.org/projects/c8c8.
Highlights
Neuroimaging of the corpus callosum has attracted great interest in both medical and neuroscience literature in the past few decades
We introduce methods for (1) automatically isolating and parcellating the callosum, (2) defining standard locations along the length of the midsagittal corpus callosum, and (3) estimating callosal thickness centered on those standard locations as well as quantifying areas within geometrically defined callosal compartments (Witelson, 1989; Hofer and Frahm, 2006)
MATERIALS AND METHODS Our overall approach is to first use standard algorithms to produce whole-brain white matter (WM) segmentations that are used for callosal quantification: standard spatial affine normalization algorithms applied to the T1W image are used to warp the WM segmentations into Montreal Neurological Institute (MNI) space
Summary
Neuroimaging of the corpus callosum has attracted great interest in both medical and neuroscience literature in the past few decades. Callosal changes due to brain atrophy have been characterized in Alzheimer’s disease (Tomaiuolo et al, 2007; Di Paola et al, 2010; Frederiksen et al, 2011b), multiple sclerosis (Hasan et al, 2012b), and Huntington’s disease (Di Paola et al, 2012) and callosal morphology has been related to symptom severity. Developmental disorders (Paul, 2011) including Williams syndrome (Luders et al, 2007a; Sampaio et al, 2012), autism (Tepest et al, 2010), attentiondeficit/hyperactivity disorder (Luders et al, 2009; Gilliam et al, 2011), and dyslexia (Hasan et al, 2012a) are associated with callosal abnormalities. The corpus callosum is vulnerable to diffuse axonal injury and atrophy following traumatic brain injury (Maller et al, 2010). Callosal changes are found during human development and aging (Sullivan et al, 2002; Hasan et al, 2008b; Luders et al, 2010b), with callosal morphology reflecting hemispheric asymmetries as well as gender differences (Bishop and Wahlsten, 1997; Luders et al, 2003, 2010a,b; Gurd et al, 2012)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
