AFNI Notes


Notes regarding AFNI and its operation...

A tremendous presentation by Doug Ward about proper thresholding for your functional results, based on a function space method that allows you to infer causality, is linked to here:

Unix stuff:

Some causes of spatial correlation in the FMRI "noise":

  • non-square k-space trajectories (i.e., spiral)
  • image reconstruction onto a finer matrix than the image resolution (e.g., reconstruction 64x64 EPI data to 128x128 grid)
  • "noise" that is induced by respiration and/or cardiac effects, or by micro movements of the head
  • "noise" in the scanner operation from shot to shot (e.g., a wire in the shim coils moving slightly)

That's all I can think of offhand, but I'm sure there are other sources. bob cox

Helpful websites:

Home Site:

AFNI Educational Material

AFNI seminar notes

Hillary Schaefer has created a good introduction:

Analysis of fMRI data with AFNI

AFNI documentation on how to get started

info on k-space


Math explanations

AFNI publication reference:

Comput Biomed Res. 1996 Jun;29(3):162-73. AFNI: software for analysis and visualization of functional magnetic resonance neuroimages.

Cox RW.

Biophysics Research Institute, Medical College of Wisconsin, Milwaukee 53226-0509, USA.

A package of computer programs for analysis and visualization of three-dimensional human brain functional magnetic resonance imaging (FMRI) results is described. The software can color overlay neural activation maps onto higher resolution anatomical scans. Slices in each cardinal plane can be viewed simultaneously. Manual placement of markers on anatomical landmarks allows transformation of anatomical and functional scans into stereotaxic (Talairach-Tournoux) coordinates. The techniques for automatically generating transformed functional data sets from manually labeled anatomical data sets are described. Facilities are provided for several types of statistical analyses of multiple 3D functional data sets. The programs are written in ANSI C and Motif 1.2 to run on Unix workstations.

False Detection Rate stuff: How to use FDR manual

3DDeconvolve handout: - a pretty clear 3dDeconvolve explanation, using the general linear model as a starting point...