Neuroscience Neuroimaging Center's Core for Morphology, Statistics and Visualization
description
The NNC Morphology Core hosted by the Penn
Center for Functional Neuroimaging provides a range of services to neuroscientists. Historically, these
services centered on helping researchers define and execute statistical
morphometric analyses, but more recently, the role of the core has been
expanded to include
- Normalization of inter-subject 3D imaging across populations
- Generation of subject-specific templates within the VoxBo software framework
- Visualization of 3D fMRI statistical maps
- Software and processes for brain structure segmentation
aims
Normalization and Template Generation
With the introduction of the PICSL's registration tools into the VoxBo software framework, we have provided CFN users with powerful new techniques for normalizing anatomical differences across subjects. PICSL's tools include software for dense diffeomorphic image registration, as well as deformable registration techniques specialized for aligning white matter structures in diffusion imaging data. These tools have been illustrated to align specific anatomical structures with greater accuracy than low-dimensional techniques typically used by neuroscientists to normalize data. Furthermore, our techniques allow users to generate anatomical templates that are specific to the particular population (e.g., neonates, the elderly, animal studies). As the result, the sensitivity and specificity of morphometic and functional studies can be boosted significantly.
Visualization
Functional neuroimaging studies generate large multi-dimensional datasets, such as statistical maps that assign a p-value to every voxel in a 3D image. Visualizing these results in a meaningful way is a challenge, especially when the regions of interest are located deep inside the cortex. To address this need, we are developing new visualization tools, as well as defining and documenting procedures that use existing tools to provide high quality visualizations. Specific visualization programs and processes allow
- fMRI activation maps to be projected on the cortical surface
- clusters of significant activation to be rendered inside of a 'glass' brain.
- high-quality rendering of 3D scalar fields to be generated using raytracing software