Hullo.
Based on the spiking problems of 3T-1, 3dDespike was created as an adjunct (preparatory) treatment to bandpass filtering of 3d+time EPIRT datasets, wherein time-series output is transformed into deviations from some mean. Presumably linear and non-linear trends have been eliminated.
Does this treatment of the data make the customary z-transformation (needed for merging of data from several brains) no longer necessary? What will z-transformation with -1zscore actually do to a dataset that has already been bandpass-filtered?