Experiments | VBM Analysis

Experimental contrasts in structural papers have several VBM-specific steps to code. All of these meta-data fields are Structural only.

Tissue Contrast

Contrast refers to the type of tissue used to measure changes. Gray matter is the most commonly analyzed, but white matter and cerebrospinal fluid (CSF) can also be used.

External Assessment

The External Assessment field in VBM Analysis specifies which of the external assessments previously recorded on the Subjects section apply to this experiment.

Observed Changes

Observed Changes describes the results of the SPI created from the VBM analysis. There are two parts, describing the directionality of the results and the type of comparison.

The directionality field has the options of "Patients > Controls", "Controls > Patients" and "Neither". The comparison-type field has the options "Volume", "Atrophy" or "Other", and is only entered if the directionality is not "Neither".


Threshold (P-value) - the P-value used for this experiment. (e.g., P<0.001, <0.01, <0.05, etc.)

Thresholding - the type of method used in thresholding, e.g. voxel-wise or cluster-wise. If this information not provided, "Unknown" can be chosen. If cluster-wise thresholding was used and a "k" value (cluster size threshold) is provided, include it in the Description.

Corrected for Multiple Comparisons

Refers to whether a statistical correction was applied to the data in order to correct for multiple corrections. If so, the type of correction used should be indicated in the Description.


This indicates whether the authors used an “optimized” VBM protocol. If the paper states they used the protocol described in Good et al., 2001 or Asburner and Friston, 2000, then this is “optimized”. Also, if images were modulated by the Jacobian determinant, then this is also considered “optimized”.

Coordinate Type

The published x, y, z coordinates could describe either a peak with the highest regional value also called a local maxima, or the center of mass which is a weighted center that better represents the center of a region. The Coordinate Type distinguishes between the two options.


A short description of the experiment.

If relevant, the description should include the external assessment as with the Subjects.

If corrected for multiple comparisons, it should also include the method of correction: e.g. "Family-wise Error Correction”.

If a cluster-size threshold is mentioned in the publication, it should be included here as well.

Copyright © 2003 Research Imaging Institute. All rights reserved.