next up previous contents
Next: Perfusion segmentation example Up: Sequence images segmentation examples Previous: Sequence images segmentation examples   Contents

FMRI example

It is assumed that the FMRI sequence has already been loaded into the Sequence Tool. Ensure that the NMR Analysis Tool, the Imcalc Tool and two Tv tools (one installed under sequence the other under imcalc) are all open.

  1. Ensure that dynamic timing information is present.

  2. Check the Stim Params options. Ensure that the hwave option correctly specifies the length of the expected `on' and `off' periods. Change the value of offset to the range of phases to be searched, or zero if no phase shifting. Check the iphase parameter and ensure it represents the correct start phase.

  3. Click the SQR option on the Stim. Func list.

  4. If desired, open another Tv and install it under imcalc graph. By changing the params, and clicking SQR the stimulus function can be modified until the correct form of the function is visible.

  5. Select the desired correlation function from the Sequence correlation function list.

  6. Click compare to initiate the comparison. The resulting correlation image will be placed on the top of the image stack and will thus appear in the imcalc Tv window.

  7. The correlation image may then be thresholded to extract a binary mask of the `active' and `inactive' regions using the thres button in the Imcalc Tool.

  8. The mask-$> $connect mouse option can be used to identify those activations regions of interest, cleaning the mask.

  9. The activation mask can be reused to define the stimulus function by selecting Mask from the Stimulus options. In this way, the square-wave function can be used to `bootstrap' the process. Once a rough estimate of the location of activations is found, capturing examples using masks reduces the restrictions of comparison to specific mathematical functions and provides an estimate of the most significant response curve in the temporal data. This process can be interpretted as an principle component analysis using the power method.


next up previous contents
Next: Perfusion segmentation example Up: Sequence images segmentation examples Previous: Sequence images segmentation examples   Contents
root 2017-11-19