 |
 |
Dear Aaron, > 1) First, how does one go about calculating Total Intracranial > Volume (TIV) in SPM8? I read that it is possible to use NewCalc_tivs.m > in SPM5, but this script is not in SPM8. Also to calculate TIV should > one use modulated warped segmented images, or should you use segmented > images in native space and what are the file denotations for the two > kinds of segmented images? How do I go about adding TIV as a nuisance > variable. The following should get you started: http://en.wikibooks.org/wiki/SPM/VBM#Obtaining_Covariates http://keck.ucsf.edu/~schenk/NewCalc_tivs.m http://www.cs.ucl.ac.uk/staff/G.Ridgway/vbm/get_totals.m > 2) In the authoritative VBM methods paper on VBM from 2000 > Ashburner and Friston assert that “voxel-based extent statistics should > not be used in VBM” because of the “highly nonstationary nature of the > underlying neuroanatomy.” Does that mean I should leave the cluster > extent at “0” when in the interactive statistical results GUI? I am > confused because I was told by other researchers to set this between 25 > and 50. No, that value is 'just' a threshold on the cluster extent (as defined by the number of voxels constituting a cluster). The correction will still assume stationarity (uniform smoothness). If you want to do non stationary cluster extent correction, have a look at: http://fmri.wfubmc.edu/cms/NS-General http://dbm.neuro.uni-jena.de/vbm/non-stationary-cluster-extent-correction/ and rule number 5 of Ged's "Ten simple rules for reporting VBM studies": http://dx.doi.org/10.1016/j.neuroimage.2008.01.003 > 3) To perform a two-sampled two sided t test (two-tailed t test) do > I just use the [-1 1] for my contrast and set the type to F? Sounds good to me. > 4) Also, I noticed that SPM8 does not have the option for False > Discovery Rate (FDR) correction for multiple comparisons, but this is > included in SPM5. Is my SPM8 missing a toolbox or a particular update or > is this not available when using an F contrast? FDR correction is still there in SPM8 but in its topological instantiation, see: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=SPM;pxeBbw;20090406192205%2B0100 If you want to use the voxel-wise FDR as implemented in SPM5, edit spm_defaults.m so that, l.42: defaults.stats.topoFDR = 0; All the best, Guillaume. -- Guillaume Flandin, PhD Wellcome Trust Centre for Neuroimaging University College London 12 Queen Square London WC1N 3BG