Print

Print


Dear Dorothy,

> I am now doing the fMRI experiment with the following data,
> 
> TR: 3000ms
> TE: 60ms
> ON-OFF experiment: 12-12-12-12ˇKstart with off state
> Multi-phase: 62
> 
> Can anyone tell me the best design matrix that I can specify for the SPM
> analysis? Also, what are the effects of selecting "SOA-variable",
> "Convolve with HRF" and "Add temporal derivatives" for fmri model setup?

One possibility is to make a design matrix with one condition. Your SOA is
fixed (the time between the start of each ON block) and these onsets
represent EPOCHS that are 12 scans long. Do convolve with the HRF, don't add
temporal derivatives (at least at first), select the default high pass
filter and use hrf for low pass filtering with no modelling of the
autocorrelations. Use a contrast of [1] to visualise ON>OFF. That should do
the trick I hope!

To answer your questions:

SOA-variable is used when either the start of the ON blocks, or the timing
of invdividual events in an event-related design, is irregular.

'Convolve with hrf' convolves the regressors with a synthetic haemodynamic
response function. This is usually a good idea unless you have made up your
own user-specified regressors that are already pre-convolved.

'Add temporal derivatives' makes a second column in the design matrix for
each regressor that approximates the temporal derivative of that regressor.
The idea is that some (unknown) combination of these two columns,
appropriately weighted, can model a (small) temporal shift in the fit of the
original regressor. This can be used to improve the overall fit of the model
if precise timings are uncertain, or to formally test for differences in the
onset of the BOLD response for different conditions.

Hope this helps.

Geraint
-- 
Dr. Geraint Rees
Wellcome Advanced Fellow,                       Lecturer,
Division of Biology 139-74,                     Institute of Neurology,
California Institute of Technology,             University College London,
Pasadena CA 91125                               London WC1N 3BG

voice   626-395-2880                            020-7833-7472
fax     626-796-8876                            020-7813-1420
web     http://www.klab.caltech.edu/~geraint
--





%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%