Dear SPM Experts,
I use SPM's cluster-level FWE correction to interrogate my results.
How do I estimate the critical extent threshold when the image I am
interrogating is one I computed myself, e.g., a minimum T map formed
from multiple T maps resulting from different second-level one-sample
t-tests. I believe the relevant function is spm_uc_RF. However, this
requires an estimate of the RESEL count for the defining search volume
(spm_resels_vol), which in turn requires an estimate of the smoothness
of the model residuals (spm_est_smoothness). Given that a minimum
statistic map is formed from multiple models, is it appropriate to
simply average the FWHM estimates from each model (i.e., the values in
SPM.xVol.FWHM) to use when estimating the RESEL count for the minimum
statistic image?
Thanks in advance for any tips.
Cheers,
Bob
|