Section 2: Processing an fMRI Time Series

scanSTAT can use a linear systems method to calculate the paradigm vector file (see, Cohen, MS. "Parametric Analysis of fMRI Data Using Linear Systems Methods", NeuroImage 6:93-103, 1997, for more details.), which describes the expected behavior of an "activated" pixel. In this release, the program can calculate correlations of each pixel of the time series with this paradigm vector, thus indicating which pixels are part of an "activated" region.

Tutorial/Demonstration - Statistical Processing of MR Images

  1. If it is not already open, start scanSTAT by double-clicking the program icon
  2. Make sure that Process From File is checked, under the File menu, Mode submenu. This places the program into offline processing mode
  3. Select "File -> Open Time Data..." to choose an image time series.
  4. Using the standard File dialog which pops up, select an image time series. A good choice might be '1R0114095-01:Sample Data:MGH types:Small.bshort'. The program is capable of handling any time series, however.
  5. Select "File -> Define Paradigm... -> Use File..." to select a Paradigm File.
  6. Using the standard File dialog, select an appropriate paradigm. 'Small.C.p',in the same directory, describes the behavior (visual stimulation) associated with small.bshort. It was created using the Cproto application, which will be integrated into scanSTAT in a future release.
  7. Real-time display of the statistics is enabled by default, on the Display menu. While this is not necessary for processing a previously-collected time series (and in fact slows down the processing slightly), it will provide a simulation of the appearance of real-time processing and image updating.
  8. Make sure Global Normalization, under the Statistics menu, is off.
  9. Choose "File -> Start Calculation" to perform the statistical analysis needed to create colorized correlation maps.

The functional maps will evolve as each new image in the input data is analyzed. Assuming that you have used small.bshort as your time series, 75 images are analyzed to create the final functional map.

The functional map during processing

The functional map after processing is completed

If all has gone properly, you should be looking at a one-slice base image in grayscale with a statistical overlay in color. This time series was acquired during visual stimulation, the timecourse of which is described in the paradigm file small.C.p. As each sequential image is analyzed, you can observe how the statistical overlay changes.

In addition, as the processing progresses, the paradigm reference vector is displayed below the image. Images containing artifacts or motion above threshold are flagged with a colored marker on this display. The color corresponds to the type of artifact detected...see the walkthrough section on artifact detection for more information.

Go to the previous section

Go to the next section

Go back to the main page