This is an issue with the Destrieux atlas and not MRIcroGL. I describe the issue and provide a corrected image here:
 https://github.com/NeuroVault/NeuroVault/issues/740
I am not sure if it is the NeuroVault's responsibility to curate all the context provided to them, so you may want to see if you can contact the correct people to get this resolved.