Thank you very much for your help Christian.
Yuri

-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]]On
Behalf Of Christian F. Beckmann
Sent: Monday, March 29, 2004 11:29 PM
To: [log in to unmask]
Subject: Re: [FSL] 3-column format analysis


Hi Yuri,

If you want to contrast imitation with reversal you'll necessarily need
more than one EV. One possibility is to use 4 different EVs: imitation
(correct), imitation (incorrect), reversal (correct) and reversal
(incorrect).

You probably don't really want to lump correct and incorrect responses
into a single EV (with +1 and -1 height as suggested in your email),
unless you really expect that an incorrect BOLD response has the same
amplitude as a correct response, just with a different orientation.

Also, you might find it easier to model this using the 1-column file
format, where - for each TR - you just specify a single number, e.g.
for the 'imitation (correct)' have a column vector of zeros and ones
where a one specifies a correct response in the imitation condition.

hope this helps
christian


On 30 Mar 2004, at 07:08, Yuri Rassovsky wrote:

> Hi all,
>
> I am trying to analyze fMRI data for each individual subject, while
> incorporating their behavioral responses. I understand that the correct
> way to do it in FSL would be by using the 3-column format and know what
> should go in each column. However, it is unclear to me how exactly to
> structure the timing of these columns.
>
> In my experiment, subjects see a stimulus on the left or right side of
> the
> monitor. There are two main conditions - imitation and reversal. In the
> imitation condition, subjects are asked to press the left button when
> the
> stimulus appears on the left and the right when it appears on the
> right.
> The reversal condition is simply the opposite - see left, press right
> etc.
> Scanning order is 8 imitation trials, 8 reversal trials, 4 rest. This
> is
> repeated 4 times. TR = 3.
>
> My goal at this point is to contrast imitation and reversal EVs. My
> understanding is that in the first column I need to put onset time, in
> the
> second column duration, and in the third subject's actual response
> (e.g.,
> 1=correct, -1=incorrect, 0=no response). As the onset time for each EV
> is
> supposed to start with 0, it is unclear to me what would be the correct
> way to structure the timing of these columns in a way that would enable
> contrast of imitation with reversal conditions (since the procedure
> starts
> with imitation and continues with reversal).
>
> Any suggestions and/or examples are appreciated.
> Thanks a lot,
>
> Yuri Rassovsky, Ph.D.
> Assistant Research Psychologist
> Dept. of Psychiatry & Biobehavioral Sciences
> UCLA Neuropsychiatric Institute & Hospital
> 760 Westwood Plaza (C8-747/NPI)
> Los Angeles, CA 90024-1759
> Phone: (310) 478-3711 ext. 43197
> Fax: (310) 268-4056
>
--
  Christian F. Beckmann
  Oxford University Centre for Functional
  Magnetic Resonance Imaging of the Brain,
  John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
  Email: [log in to unmask] - http://www.fmrib.ox.ac.uk/~beckmann/
  Phone: +44(0)1865 222782 Fax: +44(0)1865 222717