Hello again FSL experts,
I am wondering if someone can provide a bit of insight into which option
may make more sense for working with single session data/1 run in standard
space?
Below are a few options I think may work, but I am not certain which may
be best or if there are other complexities that I may be over looking.
1. applywarp to the First-level analysis stat images.
2. Run a Higher-level analysis even though my input is a single
session/run. Would this option reapply statistical constraints and/or
misrepresent the original stat values?
3. I have used renderhighres before, but this seems to only transform the
z-stats and I am not sure which transformation is being applied.
I have provided more details in my previous emails below, but please let
me know if you have any questions.
Thank you for the help.
Jacob
> Hello Stephen,
>
> Thank you for your reply and for the information. I think in an attempt to
> provide a few different solutions, my question may have gotten a bit lost.
>
> For all runs within the cohort, I have worked from the same .fsf file and
> scripted in the appropriate information. My concern about the registration
> steps, only refers to my attempt at manually applying the registration
> transformations to the subjects that have 1 run, and for this project, had
> not yet been run through a FEAT Higher-level analysis.
>
> For the subjects with 2 runs, these have been combined in a FEAT
> Higher-level analysis (Fixed Effects) and I have been able to extract the
> relevant data from their stats in standard space. For the subjects with
> only 1 run, I would like to work with their stats in standard space as
> well.
>
> I am wondering, if I only include the 1 set of first level cope images in
> a FEAT Higher-level analysis, will this reapply statistical constraints
> and/or misrepresent the original stat values?
>
> If so, would I be better off applying the registration matrices to the
> FEAT First-level stat images that are in native functional space?
>
> I am hopeful that this email and my question is a bit more clear and maybe
> now some of the possible solutions I mentioned previously will make a
> little more sense.
>
> Thank you again for the help!
>
> -Jacob
>
>
>
> Hi - all runs will have their stats transformed into standard space using
>> the same set of tools, if you setup FEAT registration in the same way in
> all runs - this is still true even if you have multiple runs per subject
> or one, with the standard way of doing things in FEAT. Even when you
> combine several runs together from a single subject (eg using a
>> higher-level fixed-effects FEAT), all those runs got transformed into
> standard space before that higher-level analysis.
>>
>> Cheers.
>>
>>
>>
>> On 2 Apr 2014, at 03:15, [log in to unmask]
>> <[log in to unmask]> wrote:
>>
>>> Hello FSL Experts,
>>> I have three hopefully simple registration questions/possible solutions
> that I would appreciate your insight into.
>>> I am working on an analysis that requires data from individual
> subjects'
>>> fMRI stats. For the majority of the cohort, subjects have multiple
> runs,
>>> which have been combined into higher-level analyses. For those subjects
> with only 1 run, and thus only a first-level analysis, I want to be
> certain that I am applying the same registration transformations as has
> been done in the higher-level analyses (FLIRT and FNIRT).
>>> Option 1. Considering the group level analysis I am working on will be
> constructed manually, I am wondering if I can simply make copies of the
> one run only .feat directories and then combine the same subjects'
> identical copes in a higher-level analysis? I am a bit skeptical of this
>>> approach because I am not sure if combining identical copes in a
> higher-level analysis may have an effect on the statistics, as in, I am
> not sure if the combining of copes and the generated stats factor in some
>>> form of weighting that would ultimately misrepresent the original
> stats.
>>> Option 2. Assuming that Option 1 not only expose my ignorance to how
> cope
>>> images are combined, but also introduces a weighting error to the
> generated stats, I am wondering if there is a way to modify
>>> renderhighres
>>> to apply the proper transformations matrices to all the stats?
>>> Option 3. Considering I am using both FLIRT and FNIRT, I am wondering
> if
>>> the example_func2standard.mat is the only transformation I need to
> apply
>>> and if I can use applywarp directly to the stat images from the
> first-level analyses?
>>> I have gone to the log files for the multi-run subjects to take a look
> at
>>> the command lines, and there seems to be a few applywarp commands used.
> This leads me to believe that I may need to apply several
>>> transformations
>>> in order to replicate what has been applied to multi-run subjects
> higher-level stats, but again, I am not certain.
>>> In summary, if you can please point me in the general direction of a
> solution I would be sincerely grateful.
>>> Thank you for the all of the help and please let me know if I may be
> missing or overlooking any details.
>>> Much appreciated,
>>> Jacob
>>> The information in this e-mail is intended only for the person to whom
> it is
>>> addressed. If you believe this e-mail was sent to you in error and the
> e-mail
>>> contains patient information, please contact the Partners Compliance
> HelpLine at
>>> http://www.partners.org/complianceline . If the e-mail was sent to you
> in error
>>> but does not contain patient information, please contact the sender and
> properly
>>> dispose of the e-mail.
>>
>>
>> ---------------------------------------------------------------------------
> Stephen M. Smith, Professor of Biomedical Engineering
>> Associate Director, Oxford University FMRIB Centre
>>
>> FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
>> +44 (0) 1865 222726 (fax 222717)
>> [log in to unmask] http://www.fmrib.ox.ac.uk/~steve
>> ---------------------------------------------------------------------------
>>
>> Stop the cultural destruction of Tibet
>>
>>
>>
>>
>>
>
>
>
|