This is a display of different types of voxelwise correlation analysis. The pixelwise correlation results from analysis of a population of n subjects / images with associated variables. The correlation at each voxel indicates whether a change in that voxel is correlated with the varibale of interest.
The background image, in grayscale, show the average tissue amount ay each voxel of the n subjects. Overlayed on the background are the strength of the correlation at voxels below the p-Value threshold, which can be adjusted.
The interpretation of the correlation depends on the type of Analysis selected.
There are three different analysis results:
- Intensity: Classical voxel based morphometry.
- Allocation: Indicates tissue loss or gain using optimal transport mophometric analysis.
- Transport: Indicates movement of tissue using optimal transport mophometric analysis.
This analysis directly correlates the intensity at a particulr voxel with the variable. Thus a postive correlation indicates that an increase in intensity is correlated with the variable. E.g. if the inout images are white matter masks a positive correlations that at that voxek an increase in white matter is correlated with the variable of interest. Similarly for CSF this would indicate a less clear fluid.
The interpretation of allocation is similar to intensity. Instead of directly correlating intensity at a voxel an intermediate step computes a value at each voxel for each subject that corresponds to whether, at that voxel, tissue was added or removed with respect to average. Without further constraints this would be equivivalent to the intensity interpretation. However, mass is only added if the subject has overall more mass with respect to teh average and vice versa only removed if the subject has less mass overall.
The mass for each subject is added or removed such that it is in a strict mathematical sense optimal. Namely such that if the subject and the average with mass added or removed can be matched with minimla effort, i.e. minimal transport of tissue.
The transport interpretation stems from the above mass allocation deifntion of matching a subject to the the mass equalized average. The transport image shows were mass was moved from (positive voxel sign) and where mass was moved to (negative voxel sign). Thus, a positive correlation indicates a large movement of mass from that location and a negative correlation a large movement of mass to that location.
- Slices: Select slice to view
- P-Value Threshold: Only correlations with p-Value below the threshold are overlaid on the background image
- Correlation Window Level: Adjust minimum and maximum of the colorscale for the corrleation display
- Transparency: Transparency of the correlation display overlay
Clicking on a location in any slice will move the slice selection to that location and display the corresponding anatomical label from the Talairach atlas. Selecting a label from the anaomtical hierarchy will highlight that region.