While for data inspection the screen resolution (e.g. 1mm equals one pixel for the 3-slices view) is sufficient, paper-based publication (articles, posters, etc.) usually require much higher resolution images (e.g. at least 150 dpi, which means that for those pictures, a super-sampling of at least factor 6 is required to get 25.4 * 6 = 152.4 pixels into one inch!).

As the main UI does not allow that kind of high-resolution zoom, a different UI is available for this task.

The image montage UI is available via the Visualization → Create montage images menu item and looks like this:

This dialog has the following controls and properties:

• slicing direction (supports sagittal, coronal, and axial/transversal)
• flipping of slicing direction (non-flipped: min:step:max, flipped: max:-step:min)
• X-flip (for coronal/axial slicing, flip along dataset's X axes, orientation)
• X/Y/Z coordinate range (set minimum value and step to 'S' for single slice image)
• super-sampling factor (up to 8)
• montage packing layout (default: letter formatted output)
• re-setting background color (replacing black with a color, incl. smoothed border support)
• adding up to 10 pixels of additional border between images
• either show output or write to file (using imwrite, different formats supported)
• pre-defined volumes to cover the brain of different template spaces

This is a typical usage example. Make sure to configure the statistical map(s) with the correct threshold (stat, k), colors, etc!

• for the MNI brain, the drop-down can be set to MNI brain, which then sets
  • X, Y, and Z min/max to -78/78, -110/82, -64/92
  • using the configured step-size to determine number of slices
  • sets a new default layout (and updates the layout dropdown choices)
• enabling X-flip
• choosing a step-size of 3 sets the layout default option to 6 x 9
• setting the super-sampling factor to 4
• enabling anatomical transparency
• choosing a new background color R: 40, G: 100, B: 160 (middle-light blue)
• adding 6 pixels of border
• choosing Show in figure
• disabling (if necessary) Display during slicing option
• click on Create montage…

The output then looks like this (for a random covariate thresholded with r >= 0.275)

And a zoomed snippet (using Matlab's internal zoom/lens button) reveals the sub-millimeter resolution and the shaded transparency of the border: