table of contents
LIPSIA     Multiple comparison correction using double thresholding
vmulticomp
This program performs multiple comparison correction using a combination of single voxel probability thresholding on the one hand, and cluster-size and cluster-z-value thresholding on the other hand. Optionally, hemispheric symmetru can be included as a third feature.

Example 1:

vmulticomp -in zmap.v -z 2.576 -fwhm 8.0 -p 0.05 -out thresholds.v

The parameter '-z' defines the initial cluster threshold of a randomly generated map of z-values. The output file contains thresholds for cluster sizes and peak z-values that are subsequently applied by the program 'vdomulticomp' to obtain an activation map that is corrected for multiple comparisons at the level specified by the parameter '-p' (whose default value is 0.05). The input image 'zmap.v' serves as a mask for the Monte-Carlo simulations. It should therefore have the same geometrical properties as the actual zmap to be analyzed, i.e. approximately the same number of voxels, the same voxel size, spatial extent, etc. The Monte-Carlo simulation fills the space of non-zero voxels in this zmap-mask with randomly generated values and counts the number of false positives.

Exploiting hemispheric symmetry:

The program 'vmulticomp' allows to incorporate hemispheric symmetry as an additional anatomically plausible feature. If hemispheric symmetry is present, then there is a chance of increased sensitivity. Note that if no hemispheric symmetry is present, then this option will not produce additional false positives. However, there will be no increase in sensitivity. To include hemispheric symmetry, the paramter '-symmetry' should be set to 'true' (see example 2). By default, it is set to 'false'.

Example 2:

vmulticomp -in zmap.v -z 2.576 -fwhm 8.0 -symmetry true -out thresholds.v

Spatial smoothness

The parameter '-fwhm' should correspond to the spatial smoothness of the data. This value depends both on the size of the spatial filter used when preprocessing the data plus the intrinsic smoothness of the data prior to spatial filtering. In 'vcolorglm' and 'v2ndlevel', and all second-level programs this value is automatically estimated, and written into the header of the zmap. It can be read using the command:

> grep -a smoothness zmap.v

Note, that there no smoothness estimation implemented in 'vwhiteglm'. Please use 'vcolorglm' to perform a smoothness estimation.

Note further, that the smoothness estimation can be counter-checked by visual comparison with the output image generated by 'valphasim'. Please refer to 'valphasim' for a more detailed explanation of this procedure.
Example
The output of 'vmulticomp' is a file in lipsia-format. It contains thresholds that specify features that a cluster must have in order to qualify as being significant. These thresholds can be applied to a zmap using the program 'vdomulticomp'. Example:

vdomulticomp -in zmap.v -out corrected_zmap.v -file thresholds.v

This call thresholds the original zmap so that only significant clusters remain. The significance is now corrected for multiple comparisons at the level specified by the parameter '-p' in 'vmulticomp'.
Parameters of 'vmulticomp':
-help
Prints usage information.
-in
Input file. Default: (none)
-out
Output file containing thresholds. Default: (none)
-z
initial threshold for defining clusters. Default: 2.576
-p
corrected p-threshold. Default: 0.05
-fwhm
fwhm of spatial smoothness in mm. Default: 5
-seed
Seed value for random number generator. Default: 555
-symmetry [ true | false ]
Whether to include hemispheric symmetry as an additional feature. Default: false
-iter
Number of iterations. Default: 1000
Parameters of 'vdomulticomp':
-help
Prints usage information.
-in
Input file containing a zmap. Default: (none)
-out
Output file containing thresholded and corrected zmap. Default: (none)
-file
File containing result of 'vmulticomp'.
Literature:

G. Lohmann, J. Neumann, K. Mueller, J. Lepsien, R. Turner (2008): The multiple comparison problem in fMRI - a new method based on anatomical priors. 11th Int. Conf. on Medical Image Computing and Computer-Assisted Intervention (MICCAI), Workshop on Analysis of Functional Medical Images, New York, Sept.10, 2008.



Max Planck Institute for Human Cognitive and Brain Sciences. Further Information: lipsia@cbs.mpg.de
Copyright © 2007 Max Planck Institute for Human Cognitive and Brain Sciences. All rights reserved.