JiscMail Logo
Email discussion lists for the UK Education and Research communities

Help for FSL Archives


FSL Archives

FSL Archives


FSL@JISCMAIL.AC.UK


View:

Message:

[

First

|

Previous

|

Next

|

Last

]

By Topic:

[

First

|

Previous

|

Next

|

Last

]

By Author:

[

First

|

Previous

|

Next

|

Last

]

Font:

Proportional Font

LISTSERV Archives

LISTSERV Archives

FSL Home

FSL Home

FSL  2004

FSL 2004

Options

Subscribe or Unsubscribe

Subscribe or Unsubscribe

Log In

Log In

Get Password

Get Password

Subject:

Re: Concatenate run/session or not?

From:

X Liu <[log in to unmask]>

Reply-To:

FSL - FMRIB's Software Library <[log in to unmask]>

Date:

Tue, 5 Oct 2004 20:22:52 +0100

Content-Type:

text/plain

Parts/Attachments:

Parts/Attachments

text/plain (168 lines)

Another follow up of this issue. I analyzed the data again from scratch
without concatenating the runs/sessions. The most significant clusters are
quite similar, though there are slight differences in significance and
cluster pattern. Also it seems the new analysis is more "sensitive" in
general (though see below for a potential reason).

However, another issue in the group model setup troubles me. Basically, the
current design is similar in a way to both examples (Paired Group
Difference, Multi-Session & Multi-Subject) in the feat document. The first
(Paired Group Difference) requires only a second-level analysis to access
the mean group difference, which is basically the setup proposed below to
exam the main effect of IV2 (with ev3-ev7 at group level modeling out the
paired/repeated measure). The second (Multi-Session & Multi-Subject)
requires a second-level analysis to pool over runs/sessions and a third-
level analysis to exam group mean. In the setup below, I originally
proposed to use the first contrast to get the average across means of
subjects (ev3-ev7) to examine the main effect of IV1. With a second though
and reference to the Multi-Session & Multi-Subject example, I found that it
is problematic because it doesn't seem to treat the subjects as a random
variable but instead treating the runs/sessions as a random variable but
subjects as a fixed variable. Am I right? This might be the reason that the
results turned out to be more significant. Therefore, should I have a third-
level analysis to get the main effect of IV1 from the second-level cope
images? In the setup below, I didn't have a contrast/cope setup for each
subject's pooled runs/sessions effect, but the PE maps (pe3-pe7) should be
the same, right?

If it is indeed the case, which I have to set up different stages to
examine effects of different IVs, it seems not efficient and make the
analysis and interpretation of the results confusing. Does it imply that we
should NOT design the study this way in the first place, with one IV
manipulated within a run/session and another IV manipulated across-run?

Thanks again for the feedback.

On Sat, 18 Sep 2004 07:57:19 +0100, Stephen Smith wrote:

>Hi - this sounds ok, but your method of generating data for approach 2, by
>splitting up the results of approach 1, is probably not optimal - in
>practice you'll still probably have some effects in the temproal
>filtering, and varying amounts of smoothness induced by the unusual way of
>doing motion correction. better to compare the approaches, if that's what
>you want to do, byt doing each totally separately!
>
>Cheers.
>
>
>On Mon, 13 Sep 2004, Xun Liu wrote:
>
>> Just to follow up this thread, I finished the analysis of Approach 2
below.
>> Besides the conceptual complexity, Approach 2 also took much much longer
>> time (e.g., one cope#.feat took more than 10 hours with 12 subjects).
>> However, the results are not much different from those of Approach 1
(with
>> concatenated runs). There are some trivial differences but the major
>> clusters/patterns stay the same, especially after thresholding.
>> One thing should be noted though. To better compare the two approaches, I
>> took the pre-processed time series from Approach 1 (i.e.
>> filtered_func_data -- this is the data after all pre-processing steps but
>> before FILM prewhitening, right?) and split them into separate runs, in
>> order to keep the pre-processing part constant. Though there were some
>> subjects with larger across-run motion, I took care to motion correct
those
>> with two-stage motion correction (within-run and then across-run) already
>> when I went with Approach 1. The spatial filter and temporal filter
(high-
>> pass) should not be affected either. In terms of the signal intensity
>> differences across runs, I had two constant EVs to model the mean shifts
of
>> the second and third runs with regard to the first run in Approach 1.
>>
>> P.S. When it is doing the estimation, some times it pops up the "ndtri
>> domain error" message when the percentage numbers progress. But I think
the
>> results are not affected by this. Any idea of what the message means?
>>
>>
>> >> On Sat, 4 Sep 2004 09:21:50 +0100, Stephen Smith wrote:
>> >>
>> >> >Hi, yes that all makes sense. Approach 2 is what you want and you're
right
>> >> >to take out the constant-EV ev1 from your original email.
>> >> >
>> >> >Cheers.
>> >> >
>> >> >
>> >> >On Tue, 31 Aug 2004, Xun Liu wrote:
>> >> >
>> >> >> On Tue, 31 Aug 2004 21:17:33 +0100, Xun Liu wrote:
>> >> >>
>> >> >> >Concatenate run/session or not?
>> >> >> >
>> >> >> >We have a design with 2 independent variables (IV1 and IV2), each
with 3
>> >> >> >levels. IV1 is manipulated across blocks within a run and IV2 is
>> >> >> >manipulated across runs. There are two approaches I can think of
to anaylze
>> >> >> >the data for the main effects and interaction for this design.
What is the
>> >> >> >advantage and disadvantage of each approach, from the conceptual
and
>> >> >> >practical perspectives? Is one more valid than the other from the
>> >> >> >statistics point of view? Thanks very much.
>> >> >> >
>> >> >> >
>> >> >> >Approach 2:
>> >> >> >Analyze each run separately and model just the 3
conditions/levels of IV1
>> >> >> >(ev1, ev2, ev3). Then the main effects of IV1 can be set up as
below.
>> >> >> >
>> >> >> >Contrast ev1 ev2 ev3
>> >> >> >mean (1)  1   1   1
>> >> >> >IV1  (2)  1  -1   0
>> >> >> >     (3)  1   0  -1
>> >> >> >     (4)  0   1  -1
>> >> >> >
>> >> >> >And then proceed to the group analysis with the EVs/contrasts
setup as
>> >> >> >below for the three levels of IV2 (say for 5 subjects)
>> >> >> >
>> >> >> >Then the .gfeat folder should include 4 cope#.feat subfolders,
one for each
>> >> >> >of the contrasts from the first level. zstat2 to zstat4 of
cope1.feat will
>> >> >> >assess the main effect of IV2 (zstat1 is the overall grand mean
of both IVs
>> >> >> >again baseline). zstat1 of cope2.feat to cope4.feat will assess
the main
>> >> >> >effect of IV1. zstat2 to zstat4 of cope2.feat to cope4.feat will
assess the
>> >> >> >interactions.
>> >> >>
>> >> >> Group ev1 ev2 ev3 ev4 ev5 ev6 ev7
>> >> >> 1      1   1   1   0   0   0   0
>> >> >> 1      1   1   0   1   0   0   0
>> >> >> 1      1   1   0   0   1   0   0
>> >> >> 1      1   1   0   0   0   1   0
>> >> >> 1      1   1   0   0   0   0   1
>> >> >> 1     -1   0   1   0   0   0   0
>> >> >> 1     -1   0   0   1   0   0   0
>> >> >> 1     -1   0   0   0   1   0   0
>> >> >> 1     -1   0   0   0   0   1   0
>> >> >> 1     -1   0   0   0   0   0   1
>> >> >> 1      0  -1   1   0   0   0   0
>> >> >> 1      0  -1   0   1   0   0   0
>> >> >> 1      0  -1   0   0   1   0   0
>> >> >> 1      0  -1   0   0   0   1   0
>> >> >> 1      0  -1   0   0   0   0   1
>> >> >>
>> >> >> Contrast ev1 ev2 ev3 ev4 ev5 ev6 ev7
>> >> >> c1        0   0  0.2 0.2 0.2 0.2 0.2
>> >> >> c2        1   0   0   0   0   0   0
>> >> >> c3        0   1   0   0   0   0   0
>> >> >> c4        1  -1   0   0   0   0   0
>> >> >>
>> >> >
>> >> > Stephen M. Smith  DPhil
>> >> > Associate Director, FMRIB and Analysis Research Coordinator
>> >> >
>> >> > Oxford University Centre for Functional MRI of the Brain
>> >> > John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK
>> >> > +44 (0) 1865 222726  (fax 222717)
>> >> >
>> >> > [log in to unmask]  http://www.fmrib.ox.ac.uk/~steve
>> >>

Top of Message | Previous Page | Permalink

JiscMail Tools


RSS Feeds and Sharing


Advanced Options


Archives

October 2019
September 2019
August 2019
July 2019
June 2019
May 2019
April 2019
March 2019
February 2019
January 2019
December 2018
November 2018
October 2018
September 2018
August 2018
July 2018
June 2018
May 2018
April 2018
March 2018
February 2018
January 2018
December 2017
November 2017
October 2017
September 2017
August 2017
July 2017
June 2017
May 2017
April 2017
March 2017
February 2017
January 2017
December 2016
November 2016
October 2016
September 2016
August 2016
July 2016
June 2016
May 2016
April 2016
March 2016
February 2016
January 2016
December 2015
November 2015
October 2015
September 2015
August 2015
July 2015
June 2015
May 2015
April 2015
March 2015
February 2015
January 2015
December 2014
November 2014
October 2014
September 2014
August 2014
July 2014
June 2014
May 2014
April 2014
March 2014
February 2014
January 2014
December 2013
November 2013
October 2013
September 2013
August 2013
July 2013
June 2013
May 2013
April 2013
March 2013
February 2013
January 2013
December 2012
November 2012
October 2012
September 2012
August 2012
July 2012
June 2012
May 2012
April 2012
March 2012
February 2012
January 2012
December 2011
November 2011
October 2011
September 2011
August 2011
July 2011
June 2011
May 2011
April 2011
March 2011
February 2011
January 2011
December 2010
November 2010
October 2010
September 2010
August 2010
July 2010
June 2010
May 2010
April 2010
March 2010
February 2010
January 2010
December 2009
November 2009
October 2009
September 2009
August 2009
July 2009
June 2009
May 2009
April 2009
March 2009
February 2009
January 2009
December 2008
November 2008
October 2008
September 2008
August 2008
July 2008
June 2008
May 2008
April 2008
March 2008
February 2008
January 2008
December 2007
November 2007
October 2007
September 2007
August 2007
July 2007
June 2007
May 2007
April 2007
March 2007
February 2007
January 2007
2006
2005
2004
2003
2002
2001


JiscMail is a Jisc service.

View our service policies at https://www.jiscmail.ac.uk/policyandsecurity/ and Jisc's privacy policy at https://www.jisc.ac.uk/website/privacy-notice

Secured by F-Secure Anti-Virus CataList Email List Search Powered by the LISTSERV Email List Manager