This uses the R gifti package (thank you, John Muschelli!) to read the GIfTI images, and the plot3D package for the plotting. It is possible to plot interactive 3d images in R (see for example, the gifti package vignette), but static images are more useful for the application I have in mind: batch-produced knitr files of surface (afni) GLM results.
Here is the first part of the knitr file produced in this tutorial; the full pdf can be downloaded here:
The gifti-plotting function I wrote displays the medial and lateral views, with hot colors for positive values and cool colors for negative. The function takes the positive and/or negative color scaling limits, with values closer to zero than the minimum value not shown, but those above the maximum plotted in the maximum color. This behavior is conventional for neuroimaging; for example, this is the same image, plotted in the Connectome Workbench with the same color scaling.
Full code for the demo is below the jump (or downloaded here). To run the demo you will need the two gifti overlay images, which can be downloaded here (left) and here (right), as well as the HCP 1200 underlay anatomies (see above). If you don't want to run the code within knitr the relevant functions can be separated from the latex sections, but you will need to set the image sizing and resolution.
Please let me know if you find this code useful, particularly if you develop any improvements!