FormatConvert  v1.6

Input formats (reads):        Analyze, Pittsburgh 1.0, AFNI, GE (I*.nnn files, E*.MR files), BrainVoyager (VMR, VTC, STC, VMP), VAPET
Output formats (writes):   Analyze, Pittsburgh 1.0, AFNI, BrainVoyager (VMR, VTC, STC, VMP)

Platforms: FormatConvert has been extensively tested on SGI/IRIX and Intel/RedHat Linux, tested on Sun/Solaris and MAC/OS-X, we are currently porting FormatConvert to Windows (cygwin)

Distribution: FormatConvert is open-source code, distributed with the fiswidgets package.  Binary executables for supported platforms will be available soon.

Contact: FormatConvert was developed by Eugene Tseytlin, Maureen McHugo, Benjamin Berry (University of Pittsburgh) for more information contact: Dan Cunningham (cunninghamd@msx.upmc.edu)

Invocation

java FormatConvert

FormatConvert window

INPUT

• Format

Format of the data to be translated from.
• Options...

Brings up the input options window, which allows for the modification of certain input data parameters.
• Input Dataset

First select the type of input. The input can be a directory, a single file, a set of files, or a filelist. Then use the browser to select the files you want. NOTE: When selecting multiple files you need to click the expand button "<<". When expanded, the browser allows you to select multiple files by selecting individual file and clicking "add". When you click "open" only the files in the list box will be opened.

OUTPUT

• Format

Format the data should be translated to.
• Options...

Brings up the output options window, which allows for the modification of certain output data parameters.
• Output Dataset

First select the type of output. The output can be a directory, a single file, or a set of files. Then use the browser to select the files you want. NOTE: When selecting multiple files you need to click the expand button "<<". When expanded, the browser allows you to select multiple files by selecting individual file and clicking "add". When you click "open" only the files in the list box will be opened.

Available Formats

• AFNI format - http://afni.nimh.nih.gov/afni/

This format was developed by Robert W. Cox and is used by the AFNI package. The format consists of a data file and a header file. The header file has a .HEAD extension and is an ASCII file containing lists of attributes describing the data. The 3dinfo utility that comes with AFNI should be used to view the description of the header file. The data is in a separate file with a .BRIK extension. Some important header information must be set by the auxiliary parameters when converting to AFNI format.
NOTE: FormatConvert does not support AFNI subbrick notation embedded into the filename. If you want to specify a particular subbrick from AFNI dataset, you need to use range of volumes option.
• AFNI options
• Input
• Default filename prefix
• Range of volumes
• Swap bytes on input
• Output
• Default filename prefix
• Blocking factor
• Scale factor
• Datatype
• Swap bytes on output
• Merge AFNI header
• Auxiliary header info
• Anatomical Parent
• AFNI dataset from which the current dataset was derived
• View Type
• Select whether the dataset is Original, AC-PC aligned, Talairach or Registered View.
• Anatomical Type
• Anatomical type in AFNI means a single image called a 3D dataset or a series of time dependant images called a 3D+time dataset.
• Functional Type
• Functional type in AFNI means a single or a collection of statistical maps, with some statistical parameters attached to them. Statistical parameters are entered in the fields below
• Scale Factor
• AFNI dataset can be scaled. The scaling factor is stored in the header and the values on the disk are scaled during run time.
• Orientation
• This field and the Geometry field below enable the user to specify the orientation of AFNI dataset.
• TR
• This is the Time of Repetition, the time it takes to acquire one volume. The units can be specified in the pull-down menu on the right of TR field. The units can be seconds, milliseconds or Hz. The next pull-down menu sets the slice order (the order individual  http://www.mayo.edu/bir/index_noscroll.html slices have been acquired).

• Analyze format

This format was originally developed by the Mayo Clinic and is used by Roger Woods in the AIR package, as well as in SPM, FSL and NIS packages. The format consists of binary image files, and each with a binary header. The header files describe all relevant information about each volume, such as dimensions, datatype, etc... and use a .hdr extension. Use a utility like AIR scanheader to see the description of header files supported by AIR package. Each image file is a 3D volume and has an .img extension. Datatypes recognized by the AIR package include 1-bit, 8 and 32-bit integers, 32 and 64-bit floats, and 64-bit double floating point.
• Analyze options
• Input
• Default filename prefix
• Range of volumes
• Swap bytes on input
• Output
• Default filename prefix
• Range of volumes
• Scale factor
• Datatype
• Swap bytes on output
• Merge ANZ header
• Auxiliary header info
• x_size - x voxel size (mm)
• y_size - y voxel size (mm)
• z_size - z voxel size (mm)
• t_size - t voxel size (msec) (TR)
• Global min/max for int16 datatypes
• unsigned int16 (glmin/glmax)  glmin = 0       , glmax = 65535 - Preset glmin/glmax setting for AIR Type 1
• signed int16 (glmin/glmax)      glmin = 0       , glmax = 32767 - Preset glmin/glmax setting for AIR Type 2 (default)
• signed int16 (glmin/glmax)  glmin = -32768 , glmax = 32767 - Preset glmin/glmax setting for AIR Type 3
• Custom (glmin/glmax)       - arbitrary  glmin /glmax values.

• GE format - The GE format consists of binary slice files. Each file contains a 7904 byte binary header with information about the slice, followed by the slice data. FormatConvert only accepts GE I.nnn and E.MR files as an input.
• Ge options
• Input
• Default filename prefix
• Range of slices
• Swap bytes on input

• Pittsburgh 1.0 format - http://www.stat.cmu.edu/~fiasco/fiasco_doc/mri_pgh.html

This format was originally developed by Marc Hahn (University of Pittsburgh) and Greg Hood (Pittsburgh Supercomputing Center)  and is used by the Fiasco and Ifis Analysis packages. The format consists of a data file and an ASCII header file. The header file usually has an .mri extension and contains key-value pairs describing the data inside of the format. The required fields are the x, y, z, and t dimensions, but other key-value pairs can be present as well. The data is usually in a separate file that has a .dat extension, however it can also be stored in an .mri file or a file with an arbitrary extension. This format supports different integer and float datatypes such as 8, 16, or 32-bit integer and 32 or 64-bit floats.
• Pittsburgh options
• Input
• Name of PGH chunk
• If this option is not specified, fconvert scans the PGH header and takes the first chunk it finds. However, if a PGH dataset contains multiple chunks, you can use this option to specify the chunks to be converted.

WARNING: If the chunk specified is incorrect, fconvert will terminate.
• Default filename prefix
• Range of volumes
• Swap bytes on input
• Output
• Name of PGH chunk
• This option allows users to specify the name of the chunk in the output PGH dataset. The default chunk name is images.
• Default filename prefix
• Blocking factor
• Scale factor
• Datatype
• Swap bytes on output
• Merge PGH header
• Auxiliary header info
• mmTR (x,y,z)
• mmBR (x,y,z)
• mmTL (x,y,z)
• mmTLZ (x,y,z)

• BrainVoyager Formats - http://www.brainvoyager.com
• STC format - STC files contain functional time course data from a single slice and have  the .stc extension. At the beginning of each file is a 4 byte binary header containing the x and y dimensions of the slice. The functional data is stored after this header as 16-bit integers. The number of time points stored in an STC file is stored in an associated FMR file as the number of volumes. The number of slices per volume (total number of STC files) is given in the FMR file as the number of slices. Values not included in the STC header can be specified using the auxiliary parameters on input. (When the number of files specified on output is less than the z dimension of the input dataset, an error is displayed.)
• STC options
• Input
• Default filename prefix
• FMR file - file containing information (such as the z and t dimensions) about the STC files in a dataset.
• Swap bytes on input
• Output
• Default filename prefix
• Scale factor
• Swap bytes on output
• Auxiliary parameters
• FMR file - The name of the FMR file that links to STC files.
• Number of timepoints - t dimension
• Inplane resolution X,Y - x and y voxel size of STC file in millimeters
• Slice thickness - z voxel size of STC file in millimeters
• TR - t size in milliseconds
• VMR format - VMR files contain a 3D anatomical volume and have the .vmr extension. At the beginning of each file is a 6 byte binary header containing the dimensions of the volume, followed by the data itself. The data elements are stored as bytes. Voxel dimensions of VMR dataset must be cubic with 1mm resolution.
• VMR options
• Input
• Default filename prefix
• Range of slices
• Swap bytes on input
• Output
• Default filename prefix
• Swap bytes on output
• VMP format - VMP files contain 3D statistical results computed from a VMR file and have the .vmp extension. At the beginning of each file is a variable length binary header containing information about the data. The statistical information is stored as 4 byte floating point numbers. Voxel dimensions of VMP dataset must be cubic with 1mm resolution.  Currently, only VMPs containing a single map are supported.
• VMR options
• Input
• Default filename prefix
• Swap bytes on input
• Output
• Default filename prefix
• Scale factor
• Swap bytes on output
• BrainVoyager Version - this field indicates which version of BrainVoyager VMP files are read. This will be more relevant with subsequent releases of BrainVoyager, such as QX. The only version currently supported is 1.
• Statistical Map Type - VMP files can contain any type of statistic. The options are t- or F-values, correlation values, cross-correlation values, or other types of statistics.
• Number of Lags - the number of correlations computed or the number of TRs per cycle. This field is used only when cross-correlation maps are used.
• Cluster Size - clusters of statistical values.
• Statistical Threshold (Critical Value)
• Statistical Threshold (Maximum Value)
• Degrees of Freedom (1,2) - Degrees of freedom 1 is the numerator or effect. Degrees of freedom 2 is the denominator or error.
• (X,Y,Z) Start -

X Start - beginning of relative position of VMP volume on the x-axis within a 256³ VMR volume.
Y Start - beginning of relative position of VMP volume on the y-axis within a 256³ VMR volume.Z Start - beginning of relative position of VMP volume on the z-axis within a 256³ VMR volume.
 
• VTC format - VTC files store 4D functional data and have .vtc extension. . At the beginning of each file is a variable length binary header containing information about the data, followed by the data itself. The data is stored as 16-bit integers. Voxel dimensions of VTC dataset must be cubic with 1mm or 3mm resolution.
• VTC options
• Input
• Default filename prefix
• Range of volumes
• Swap bytes on input
• Output
• Default filename prefix
• Blocking factor
• Scale factor
• Swap bytes on output
• Auxiliary parameters
• FMR file - The name of the FMR file that links to STC files that were transformed into the VTC file.
• PRT file - The name of the PRT file linked to the VTC file.
• Hemodynamic delay
• Hemodynamic function, delta parameter
• Hemodynamic function, tau parameter
• Segment size - used with time course separation
• Segment offset - used with time course separation
• XStart - beginning of relative position of VTC volume on the x-axis within a 256³ VMR volume.
• YStart - beginning of relative position of VTC volume on the y-axis within a 256³ VMR volume.
• ZStart - beginning of relative position of VTC volume on the z-axis within a 256³ VMR volume.
• VAPET format -  http://neurovia.umn.edu/papers/tech_reports/vapet_format.html

Each VAPET files can contain either a 3D data volume or a set of 3D statistical maps. The format consists of binary data file with arbitrary extension. At the begining of each file there is an ASCII header. Data offset is specified within the header, but it is usually 512 bytes. The header describes all relevant information about each volume, such as dimensions, datatype, etc... You can use statndard Unix head utility with -c512 flag to see the description of each vapet file.  Datatypes recognized by the VAPET format include 8,16 and 32-bit integers, 32 and 64-bit floats.
• VAPET options
• Input
• Default filename prefix
• Range of volumes
• Swap bytes on input

• Menubar
• Toolbar
• Output window