Is it possible, with existing software, to extract signal changes at vertices, voxels, or ROIs for single trials, rather than across trials?  This would allow an examination of  intra-subject variability in BOLD response, habituation effects, etc...  Something like this was done by: Yamaguchi S, Hale LA, D'Esposito M, Knight RT (2004): Rapid prefrontal-hippocampal habituation to novel events. J Neurosci 24:5356-5363.