Re: [Freesurfer] FIR analysis, time window question

2013-03-05 Thread Douglas N Greve
yes, my mistake doug On 03/05/2013 01:30 PM, Cesar Echavarria wrote: > Also, did you mean ' fir 4 26' in your previous response? since the > TimeWindow parameter should include the 'PreStimulus' time > (I just want to double check since this is crucial to my question) > > Cesar > > > On Tue, Mar

Re: [Freesurfer] FIR analysis, time window question

2013-03-05 Thread Douglas N Greve
Hi Cesar, you cannot really do the analysis you want to do, it is mathematically impossible. If you do not expect a response from one of the conditions, then you could code it as 0. Alternatively, you can consider both conditions as a single 20 sec event and get the FIR for that event out to 2

Re: [Freesurfer] FIR analysis, time window question

2013-03-05 Thread Cesar Echavarria
Also, did you mean ' fir 4 26' in your previous response? since the TimeWindow parameter should include the 'PreStimulus' time (I just want to double check since this is crucial to my question) Cesar On Tue, Mar 5, 2013 at 1:00 PM, Douglas N Greve wrote: > > If you presented the stimulus for 1

Re: [Freesurfer] FIR analysis, time window question

2013-03-05 Thread Cesar Echavarria
Does the post-stimulus period have to be as long as 12 seconds? With such a long time window I get the following error when running selxavg3-sess: XCond = 1.47995e+18 (normalized) ERROR: design is ill-conditioned To clarify: in our design we have 10 seconds of condition 1 followed immediately aft

Re: [Freesurfer] FIR analysis, time window question

2013-03-05 Thread Douglas N Greve
If you presented the stimulus for 10sec, then you should set the time window to include about 12 sec after stimulus offset. If your stimulus is 10 secand you want the time window to start 4 sec prior to onset and go 12 sec after onset, then you will need a time window of 4+10+12=26sec, so it w

Re: [Freesurfer] FIR analysis, time window question

2013-03-05 Thread Douglas N Greve
Hi Cesar, for your case, it would correspond to 4sec prior to stimulus onset (the TR that goes from -4 to -3 sec). What seems to be wrong? The last frame would be the TR that goes from +9sec to +10 sec. Not related to your question, but Is 10 sec post stimulus long enough? doug On 03/05/2013 1

[Freesurfer] FIR analysis, time window question

2013-03-05 Thread Cesar E Echavarria
Hi Freesurfers, I did an FIR analysis to get the BOLD response over time, the following way mkanalysis-sess -analysis Afterimage_FIR2.rh -surface self rh -native -fwhm 5 -paradigm ShapeAE_New.par -event-related -polyfit 1 -mcextreg -TR 1.000 -nconditions 8 -force -fsd bold -fir 4 14 I then ge