Table of Contents

Phiwave wavelet toolbox for SPM

Phiwave is a toolbox for wavelet analysis within SPM. The toolbox uses statistical designs from SPM to run analyses on images in wavelet space. Please see this Phiwave presentation from HBM 2001 for an introduction to the ideas behind Phiwave.

We have also presented the application of Phiwave to the FIAC dataset, at the Human Brain Mapping conference in Toronto, 2005. Here is the Phiwave FIAC presentation. Please see our Phiwave FIAC page for links to the final paper, and full scripts to reproduce the analysis down the to publication figures.


Aston, J.A., Brett, M., & Turkheimer, F. E. (2006) HBM functional imaging analysis contest: data analysis in wavelet space. Human Brain Mapping (in press).

Turkheimer, F. E., Brett, M., Aston, J. A. D., Leff, A. P., Sargent, P. A., Wise, R. J. S., Grasby, P. M., & Cunningham, V. J. (2000). Statistical Modeling of PET Images in Wavelet Space. Journal of Cerebral Blood Flow and Metabolism, 20, 1610-1618.

For other relevant references, see Federico's home page.


All Phiwave releases are available via the Phiwave project SourceForge page.

For those who need the latest code and edits, please see the Phiwave github page


First download and install the latest stable version of MarsBaR.

Next download Phiwave as above.

Unpack Phiwave into a directory somewhere. This will probably give you a new directory called something like phiwave-3.2 (where 3.2 is the current version number).

Put this directory on your matlab path.

Make sure that mex file compilation is properly set up for matlab.

In the matlab console, compile the Phiwave routines with

>> phiwave make
Now you can start phiwave with
>> phiwave

How to run an analysis in Phiwave

To get started:
  1. Choose Design - set design from file
  2. Select an SPM2 or SPM99 analysis SPM.mat or SPMcfg.mat file
  3. Run Design - Convert to unsmoothed, if your design used smoothed images, and you still have the unsmoothed versions of the image files on disk. This allows Phiwave to use all the resolution of the data.
  4. Choose Results - estimate. Phiwave will do a wavelet transformation on all the images in the design, and estimate the model in wavelet space, saving the results to the current directory. If you get messages about being unable to open image files, have a look at the "Can't open image" section of the MarsBaR FAQ.
  5. Select Results - Specify / denoise comparison; enter a contrast of interest and follow the prompts to write out a denoised contrast image.
  6. Use the Results - display slices and Results - display sections options to display the denoised contrasts


There is a moderate amount of documentation of the Phiwave code, and wwe export this as web pages using Guillaume Flandin's excellent m2html. This is available in the doc subdirectory in the main phiwave code directory. It is also available in the doc/latest directory of the website.

Batch examples

Phiwave is easy to batch. See the examples directory in the Phiwave distribution.


Please email matthew.brett who is at, if you have any questions, problems or suggestions.

The end

Until the next time, arrivederci.

The Phiwave team Logo

Last Refreshed: Thu Jan 12 12:51:59 GMT 2006