preprocessing before TBR

[riviera lin]

Dear Aaron

When conducting TBR, do we need to do nuisance covariate removal and wavelet despiking (Patel) ? The original paper mentions that these preprocessing may not be needed. Does this make a big difference to the result ? 

Thanks

Chemin

"Another future direction for exploration is the degree to which more or less preprocessing would affect results. For example if there are known sources of nuisance variance such as CSF signal, these could be removed prior to the TBR mapping procedure. Since the TBR method can implicitly control for nuisance variance insofar as the template in question is a good representation independent of nuisance variance, it may also be the case that preprocessing such as motion regression and filtering may not be strictly necessary and would further reduce the assumptions and data processing required to utilize the TBR method."(Neuroimage. 2014 Nov 15; 102(0 2): 620–636.)