Kathryn -
In you want to model your conditions as epochs (blocks), rather
than runs of separate events, then, yes, you need only specify
the onset of each epoch (not of each event within that epoch).
You got the right end of the stick.
In this case, the SOA refers to the time between epoch onsets
(not the time between events). You could choose "Fixed" SOA,
in which case you will be prompted for the first onset of
each epoch-type, and the repetition time (in scans) between
each successive epoch of that type. However, given that you
only have one epoch of each type per session, you might as
well enter "Variable" SOA, and simply enter one onset time
for each epoch. (If you did have replications of conditions, you
would often choose Variable, since people normally randomise/
counterbalance the order of conditions, producing a variable SOA).
Finally, your units of scans are correct, starting at 0.
Rik
PS Note that one situation where you might want to model the
trials within each block as separate events is when the time
between trials is quite long (>4-5s), and a "boxcar" regressor
for the epoch-effect may not be an accurate model - ie periodic
signal changes around the mean effect will leave some structure
in the residuals. You may be able to ignore this in your case,
with reasonably rapid trials.
> Dear Drs. Karl, Rik, Geraint , Andrew, John and fellow SPM people,
>
> I remain confused over the distinction between fixed and variable SOA
> when specifying an fMRI model/ design matrix.
> Mine is a block design (not e-r) with five conditions (task blocks)
> that occur only once in each session (i.e. they are not replicated
> within session), with each block the same length of about 18 scans
> long. Stimuli were presented on average every 4 secs but with a pseudo
> random stimulus onset asynchrony of + 0.2 secs (so ranging from 3.8 to
> 4.2s), while the TR was 3.694 secs.
>
> To clarify, is each task block mentioned above considered an 'epoch'?
> >From the spm archives I get the impression that for block designs
> the SOA refers to the time (in scans) between different blocks
> (epochs) and not the time between each stimulus....or have I got the
> wrong end of the stick? (no pun intended).
>
> I understand that SOA (and condition onset times) are specified in
> units of scans (my TR was 3.694 secs). If I specified fixed (4 secs)
> my SOA would be 1.08 scans, and my condition onset times would be in
> units of scans (e.g., condition 1= 0 scans; condition 2= 18 scans
> etc.). Alternatively, should I specify my model using a variable SOA
> (as I have a variable SOA) in which case I would need to specify a
> vector of stimulus onsets - accounting for the actual variability in
> SOA in my stimulus presentation?
>
> I am really unclear about which is the appropriate model and seek some
> assistance from the honourable Drs. and fellow SPM people.
>
> Cheers, Kath
> --
> Kathryn Moores, Ph.D. Student
> Cognitive Neuroscience Laboratory
> School of Psychology
> Flinders University
> Adelaide, South Australia
> Phone 08 8201 2547
> Fax 08 8201 3877
> Email: [log in to unmask]
>
--
---------------------------8-{)}-------------------------
DR R HENSON
Institute of Cognitive Neuroscience &
Wellcome Department of Cognitive Neurology
17 Queen Square
London, WC1N 3AR
England
EMAIL: [log in to unmask]
URL: http://www.fil.ion.ucl.ac.uk/~rhenson
TEL1 +44 (0)20 7679 1131
TEL2 +44 (0)20 7833 7472
FAX +44 (0)20 7813 1420
---------------------------------------------------------
--
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|