Neurology Applications

I have written a number of applications during my time at the University of Iowa, working in the Department of Neurology (in the Laboratory of Computational Neurology) and before that at the Image Analysis Facility. Those efforts begat a couple of applications that are still being actively used to this day. Some, like Voxblast, mtrace and vtrace are being handled by others. Those tools are perhaps best known for creating images like this one:

science cover image

Damasio H, Grabowski TJ, Frank RJ, Galaburda AM, Damasio AR, "The Return of Phineas Gage: Clues About the Brain from the Skull of a Famous Patient," Science, Vol 264, 20 May 1994.

I still support and work on the Brainvox tool and its supporting infrastructure (the tal_programs and I/OWA/CVIO).

Solid Brain Models

I just got one of these made. It is a 3D solid model of a brain "printed" as a volume, in color.  You have to see/hold this to believe it!

Solid Brain Model

These can be made actual size. I generated the .PLY files for this one using Brainvox (to align the MRI and PET data) and the tal_ programs to compute the basic surfaces and sample the PET data. The printing was done by Tom Sattler (thanks Tom!) at Quantum Leap on a Z-Corporation Spectrum Z510. The hue in this model comes from the PET, which does not cover the entire brain.  The saturation comes from the MRI. I am considering making it possible to directly export the necessary .PLY files from Brainvox in a future version.

Brainvox

Probably the most published/referenced tool I have written. It was originally described in:

Damasio H, Frank RJ, "Three Dimensional Mapping of Brain Lesions in Humans," Archives of Neurology, 49:137-143, 1992

and later with PET enhancements in:

Grabowski TJ, Damasio H, Frank RJ, Hichwa RD, Boles Ponto LL, Watkins GL, "A New Technique for PET Slice Orientation and MRI-PET Coregistration," Human Brain Mapping, 3:123-133, 1995

and still later with fMRI enhancements in:

Frank RJ, Damasio H, Grabowski TJ, "Brainvox: An Interactive, Multimodal, Visualization and Analysis System for Neuroanatomical Imaging," NeuroImage, 5:13-30, 1997

The application is still in heavy use today.  Here is a screenshot taken from the most recent version running under OSX with the basic image windows open:

brainvoxscreenshot

The Brainvox tool is not yet available for general download. Today, we support Linux, Irix and OSX with a Windows port in the works. To get more information about the application and discuss getting a copy, I refer you to the Laboratory of Computational Neurology wiki site.  

The latest: Brainvox 3.13 has been released.  

I just added a scripting language to Brainvox.  Not complete yet, but enough to allow one to script the rendering operations.  I also added read/write support for Analyze/NifTi-1 format files.

This version now supports Windows and runs great on a 7900GT graphics card (very fast).  The depth clipping of embedded geometry works again and a number of fixes for complex pathnames have been added. The Windows version is better integrated with drag and drop support as well as clipboard support. The tal_programs were also updated with that release.


tal_programs

The brainvox suite includes a collection of standardized volumetric image processing tools designed to be scripted together to perform complex analysis. The tools include support for:

  • Change-distribution analysis on a volume

  • Compute weighted centroid of a volume

  • Cluster detection functions (flood filling)

  • Mapping polygonal regions of interest to voxel volumes
  • Compute volume or planar Euclidean distance maps

  • Compute the Euler characteristic of a volume

  • Compute the forward and reverse fourier transform of a volume

  • Perform 2d and 3d connected components labeling

  • List the status of network volume servers

  • Perform general purpose volumetric math operations

  • Perform volume isosurface detection

  • Apply an adaptive median smooth filter to a volume

  • Perform simple linear volume math between two volumes

  • Create a shared memory volumes

  • Perform slicewise normalization using a polynomial fitting technique

  • Normalizes a volume with respect to its globals mean

  • Compute optimal surfaces through 3D cost functions

  • Orthogonal planar resampling of datasets

  • Compute the principle components of a series of volumes

  • Probe a volume for local values around a location in the volume

  • Perform multiple volume, general linear regression

  • Compute a volumetric connected bounding mask

  • Perform batch image file renaming operations

  • Compute non-parametric t-statistics for a series of volumes

  • Resample a Brainvox calibrated volume into Talairach space

  • Perform volumetric cropping and rescaling/resampling

  • Destroy a shared memory volume

  • Convert a series of ROIs into a triangulated polygon mesh

  • Output voxel values within a region specified by a set of ROIs

  • Compute volume or planar skeletonization maps

  • Perform volumetric Gaussian smoothing

  • Filter a list of tal_cluster points for significance using spatial extent

  • Performs a non-parametric T-test on differences volumes including the SPM sphericity computation

  • Perform table based and statistical pixel mapping operations

  • General purpose volumetric statistics (global and local)

  • Compute surface area and volume of a binary volume

  • Perform vertex reduction on a polygon mesh file

  • Perform slice-wise contour detection in a volume

  • View binary images in various ways

  • Perform general purpose volumetric warping
And a lot more...

They are probably best described here. Again, the sources are not yet generally available, but we have released them as research tools for several users on request.

Eventually, I would like to be able to distribute all the sources from this web page.