Computational protocol: Postnatal Microstructural Developmental Trajectory of Corpus Callosum Subregions and Relationship to Clinical Factors in Very Preterm Infants

Similar protocols

Protocol publication

[…] DTI data were transferred to an offline workstation for further image processing and analysis. We pre-processed all images using FSL 5.0 software of FMRIB Software Library (Analysis Group, FMRIB, Oxford, UK) and DTIStudio 3.0.2 of MRI Studio (Johns Hopkins University, Baltimore, MD). Briefly, we converted raw DTI DICOM data into analyze format using DTI Studio and imported this into FSL to perform eddy current correction. To further minimize the effects of motion and artifacts, we performed Automatic Image Registration (AIR) and automatic outlier slice rejection in DTI Studio. Next, we performed, tensor estimation and generated scalar maps (FA, MD, radial diffusivity [RD], and axial diffusivity [AD; L1 – lamda1]). Figure  illustrates examples of FA color maps at each of the four postmenstrual age time points. Last, in FSL, we employed the Brain Extraction Tool to perform brain extraction and BEDPOSTX (Bayesian Estimation of Diffusion Parameters Obtained using Sampling Techniques) to run probabilistic tractography and model for crossing fibers. BEDPOSTX creates all the files for running probabilistic tractography by employing Markov Chain Monte Carlo sampling to build up distributions on diffusion parameters and model crossing fibers within each voxel of the brain. This software permits inclusion of datasets with up to 60 diffusion directions.Figure 5 [...] Using the “ROI Tool” in Analyze 12.0 software (Mayo Clinic), the CC bundle was segmented on the most clearly delineated midbrain slice in the sagittal orientation. The “Auto-Trace” tool was selected to trace out the CC and divide it into 30 vertical segments of the same width (Fig. ). This allowed us to divide the CC into six component parts, as described by Thompson et al.. Of the 30 segments, proceeding in an anterior to posterior direction, the first 5 segments were labeled as the genu (5/30), the next 5 as the rostral body (RB) (5/30), followed by anterior midbody (AMB) (5/30), posterior midbody (PMB) (5/30), isthmus (4/30), and splenium (6/30) (Fig. ). The rostrum (red segment adjacent to genu in Fig. ) was typically small or unidentifiable in many very preterm infants and therefore was included as part of the genu whenever it was visible. Last, we imported each of the six newly labeled subregions of the CC as a seed point mask into FSL’s probabilistic tracking with crossing fibers (PROBTRACKX) tool to perform probabilistic tractography (Figs  and ). With the exception of using 0.4 mm for the step length (due to the smaller infant brains), we used the default parameters suggested by FSL developers. To ensure good inter-rater reliability, two raters segmented a different set of cases prior to these study cases to establish consistency in methodology. The raters then independently segmented the CC subregions of a random sample of 15 cases to measure inter-rater reliability.Figure 6Figure 7 […]

Pipeline specifications

Software tools FSL, MRI Studio, BET, Probtrackx
Applications Magnetic resonance imaging, Diffusion magnetic resonance imaging analysis
Organisms Homo sapiens
Diseases Apnea, Bronchopulmonary Dysplasia
Chemicals Caffeine