hi lars,
i just checked with doug, and he agreed that, since your stimulus is periodic, 'sfa-sess' can do the job provided that you specify that the analysis should be carried out using a block design.
so, for example, if the TR is 2 sec, and since each of your conditions lasts 20 TRs per epoch for a total of 10 repetitions (or 20 epochs) per run, you can define the block design with:
mkblockpar -parname horiz_vs_vert_block.par -TR 2 -nperepoch 20 \ -nepochs 20 -epochid 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2
then define the analysis with something like:
mkanalysis-sess \ -analysis visual_meridian_block -paradigm horiz_vs_vert_block.par \ -designtype blocked -TR 2 -nconditions 2 \ -funcstem fmcsm5 -inorm -mcextreg \ -polyfit 2 -gammafit 2.25 1.25 -taumax 26
(please consult the documentation to determine how to choose the options on those last two lines.)
the key here is that the design type should be specified as "blocked" (instead of "retinotopy", which implies phase encoding).
in your case you can use 'sfa-sess', and for this example the command would be:
sfa-sess -df sesspar -sf sessid -a visual_meridian_block
-jon
On Wed, 27 Jun 2007, Lars Kuchinke wrote:
thanks Jon,
i didn't know that a simple block design analysis is possible to compute the visual area boundaries... (which is the reason why I tried to fit it into the standard sfa-sess analysis)... could you explain me (shortly) what to do than with my data
regards, Lars.
-design info: 8hz flickering checkerbords, 20 scans at horizontal meridian, 20 scans vertical..x10 repetitions...
Jonathan Polimeni schrieb:
hi lars,
the FSFast retinotopy analysis stream is geared towards temporal phase encoded data analysis, which produces field sign maps that can be used to define area boundaries.
if you opt to use stationary wedge stimuli to activate the horizontal and vertical meridian representations in visual cortex rather than the traditional phase encoding stimulus paradigm, since the vertical and horizontal meridian representations also lie along area boundaries, you can identify them in your data through a standard block design paradigm analysis. this is straightforward but outside the retinotopy stream. instead you can use 'mkanalysis-sess' and 'selxavg3-sess' as described on the fswiki.
of course it is possible to interpret your periodic stimulus as temporally phase encoded and run the usual frequency analysis (i.e., 'sfa-sess'), but i think that the block design analysis is more appropriate for your data.
hope this helps!
-jon
Lars Kuchinke, PhD Freie Universität Berlin Allgemeine Psychologie Habelschwerdter Allee 45 14195 Berlin
kuchinke@zedat.fu-berlin.de www.fu-berlin.de/allgpsy Tel: 030 838 55776