Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Subject: Re: SOA ?
References: <[log in to unmask]>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: [log in to unmask]
Subject: Re: SOA ?
Stefan -
> | there is confusing information to me about the SOA
> | parameter. Using 45 scans with condition A=stimulation
> | and condition B=no stimulation and 5 scans within one
> | epoch with time series BABABABAB, I wonder wether to
> | use two trails: trail1=rest (SOA 1 11 21 31 41) and
> | trail2=stimulation (SOA 6 16 26 36) or to use only one
> | trail1=stimulation (SOA 10).
In the case of a simple (unconvolved) boxcar, a [1 -1]
contrast in the first model (2 trails) should give very
similar answers to a [1] contrast in the second model
(1 trail), bar a small difference in the df's.
Following convolution with the hrf, the two regressors
in the first model will not be inversely related, and you
may get slightly different answers for the two contrasts
in the two models. You might then favour the first model
(two trails), given that you are not short on df's with 45
scans and only 2 conditions.
Note also that in SPM99, time starts at 0, so if your
first condition actually starts exactly with the start of
the first scan, you may need to enter the SOAs as
[0:10:40] and [5:10:35]. (However, given your reasonably
short epochs - assuming a "normal" TR - the precise
timing may be important, and if you haven't slice-time
corrected during pre-processing, you could minimise
timing errors by synchronising with the middle slice
acquired by entering the SOAs as [0.5:10:40],[5.5:10:35]
(or changing the default value of fMRI_T0 to 8; see
earlier postings)).
> I was advised at SPM Course
> | this year in Hamburg to use the first possibility but
> | with wrong result. Whereas second possibility showed a
> | meaningful result. In Hamburg, I was further advised to
> | use for my fixed modell the variable option for SOA and
> | not the fixed option, what is a little bit confusing
> | for me. Can you plaese help me clarifing this question?
It should make no difference (when entered correctly):
ie entering a vector of 'variable' SOAs [0:10:40] and
entering a single 'fixed' SOA of 10 (starting at 0), both
with durations of 5 scans, should give the same result.
Rik
--
---------------------------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
---------------------------------------------------------
--
|