This educational session will introduce and discuss challenges for brain imaging of pediatric subjects. A brief overview of key differences between the developing brain and adult brain will be discussed. The presentation will be address logistics of patient handling, image acquisition, post-processing techniques and analyses for improved characterization of the maturing brain. Translational studies will also be presented to highlight the importance of advancing pediatric brain imaging not only in research but also in clinical care, and further identify the area of needs to spur interests from the audience.
First, the issue of motion during scans will be discussed as it is the main challenge of performing pediatric neuroimaging. Physical methods of swaddling and immobilization of the head will be shown and optimal workflow will be discussed to minimize the time away from the neonatal ICU.
A brief overview of key differences between the developing brain and adult brain will be presented. In addition to the obvious size differences between the pediatric and adult brain, the anatomy and functions of the brains are also different. These differences require optimization and sometimes, redesign of imaging hardware and acquisition sequences. The rapid changes of the anatomical structures during the early brain development also create difficulties in analyzing the data acquire, especially for group comparisons.
Hardware including MRI-compatible incubators and imaging coils will be briefly presented with a flavor of the historical developments over the last two decades starting in the early 90s with the works of Drs. James Barkovich and Donna Ferriero at UCSF and end with state-of-the-art high channel phased array coils.
With higher signal-to-noise ratio (SNR) of the higher field scanners (>3T), better gradients, and improved electronics for acquisition and processing, we now can acquire sub-millimeter resolution anatomic (T1, T2, T2*-weighted) images in a few minutes. These high resolution anatomic images have a wealth of information when the proper analytics are done to extract crucial information. Morphometric analyses yielding cortical folding patterns, curvature, gray and white matter patterns, and regional and global structural variations can be utilized to characterize normal and abnormal brain development.
Diffusion imaging is quickly becoming one of the most population and informative method for investigating tissue microstructure. By highlighting diffusion effects using directional gradients, diffusion models can be build to characterize membranes, axons, and structures that inhibit water motion in specific directions. In cases of injuries (i.e. trauma, hypoxia ischemia, stroke), abnormal disruptions in the microstructures are well highlighted on diffusion images. With the introduction of tractography and network modeling, efforts are being made to study the brain as organized network with hubs and connections. Due to the rapid changes in the developing brain, atlas based cortical parcellation schemes for analyzing adult brains are not optimal; newer methods of atlas-free and temporal atlases are needed to improve the network construction. While the utility of network analysis is still being investigated, preliminary results in translational studies are presented to encourage future works.
Functional networks using blood oxygenation level dependence (BOLD) are being often considered at the same time as microstructural networks. Task and resting state fMRI are used to assess different aspects of the functional network. Due to the long acquisition times, the majority of investigations are for research purposes with very limited clinical applications. Highlights of potential areas with clinical utility will be presented. Metabolism is one of the fundamental processes of life.
MR spectroscopy (MRS) is a method to detect metabolites in vivo. The metabolic activity in a young brain is very different than that of an adult. The changes in concentration of specific metabolites can be used to diagnose abnormal tissue condition before gross changes in anatomy and microstructural changes can be detected. The most commonly available spectroscopy method is the single voxel spectroscopic acquisition (SVS), in which one volume of interest is selected and measured. This method is simple to implement and use due to its simple acquisition and processing needs. 2D and 3D MRS acquisitions are also available and fast acquisition techniques are being explored to provide metabolic information over a larger volume of interest while keeping the scan times short. Most of this talk will be focused on steady state proton spectroscopy with a short introduction into dynamic metabolic imaging using dynamic nuclear polarization of carbon-13.