Still looking for an answer to this question:

  So in doing a more complicated version of the analysis with paradigm weighting, I had another experiment where I wanted to test out two different models, 1) that BOLD activity was related to my behavioral measure alone or 2) that BOLD activity was related to my behavioral activity plus a factor that represented the number of distractors present (which is constant across set size) (had a high and low distractor condition, so that this factor was effectively 0 for the low distractor condition and something like 1 for the high distractor condition).  I modelled the second my simply adding the behavior measure and the distractor factor together (which means that in practicality, nothing changed for the low distractor conditions between model 1 and model 2, but that the weight was increased by a constant factor  in the high distractor conditions between model 1 and model 2).

When I ran this analysis, what I got (shown in the attached picture in a group average) was an increase in the amount of activity and area of activity for model 2 (targets & distractors) compared with model 1 (targets only).  Am I right in interpreting this as model 2 being a better fit?  Or is can these maps be compared to show areas that are a better fit for model 1 and other areas that are a better fit for model 2?

I have ROI data on the unweighted analyses that goes along with the idea of model 2 being better, but I wanted to make sure I was intrepreting the  activity in this weighted analysis correctly.

Thanks,

Katie


On Tue, Feb 22, 2011 at 1:35 PM, Douglas N Greve <greve@nmr.mgh.harvard.edu> wrote:
Yes, your understanding is correct. People might choose Version 1 because they strongly feel that it accurately represents the underlying mechanism. It saves one DoF, but that's not going to change much. I think the problem with Version 1 is that, for most designs, it forces you to extrapolate outside of your data range. Few things are truly linear, but they are often well modeled by straight lines over a limited range. For your data, you might really expect the response to be 0 at x=0, but it become a very non-linear curve in that region.

doug

Katie Bettencourt wrote:
Ok, so just so that I understand (I realize I'm being a bit dense, and I appreciate you walking me through this).  If I am weighting my paradigm by my behavioral measure, then the activity I see with this weighted analysis, is areas where we are seeing a linear relationship between BOLD activity and my behavioral measure.  In version 1:1vs0, this assumes that my linear fit passes through xy=0.  However, in version 2:2vs0, we put an offset (or a sort of y intercept?) into the model so that the data doesn't have to pass through xy=0.  And you are saying that the larger amount of activated area that I am getting in Version 1:1vs0, is because I'm seeing areas activated for the fact that there's an increase from baseline but not exactly the linear fit between BOLD and behavioral measures, but for Version 2:2vs0, I see only the areas that show this linear correlation?  Does that sound about right?

And if I am right, why would anyone do Verison1?  Is it just because if your fit does pass through xy=0, you get more power this way compared with Version 2?

Katie

On Tue, Feb 22, 2011 at 1:00 PM, Douglas N Greve <greve@nmr.mgh.harvard.edu <mailto:greve@nmr.mgh.harvard.edu>> wrote:



   Katie Bettencourt wrote:


          You should also look at Version2:1v0. I bet a lot of the areas
          from Version1:1v0 will also show up. You can also create a
          Version3 in which you divide your presentations into a
       low-weight
          and a high-weight (but set the weight=1). Then create
       contrasts of
          low+high and high-low. The low+high should look like
       Version2:1v0
          and the high-low should look like Version2:2v0.



       So, what would this say?  That the extra activity I'm getting
       in Version 1 is just the slope/offset, and isn't actually load
       related data that I am trying to get at (ie. isn't activity
       that is increasing as the number of items the subjects is
       remembering increases, but is just baseline visual stimulation
       activity or something?)

   Yes. It's easier to explain if I can draw it out. But imagine an
   xy plot with your weight on the x axis and the fMRI response on
   the y axis. If there is an offset but no change in fMRI with
   weight, then the data points will be on a horizontal line. Now
   what happens if you try to fit that data with a line that is
   forced to pass through xy=0? You'll get some positive slope, but
   overall it won't fit very well.


       And What exactly to do you mean by divide my presentations
       into low and high weight?  What exactly would this be
       comparing?  Sorry if I'm being dense.

   I mean create a new paradigm file with two non-null conditions.
   Condition 1 would be all presentations with weight < 0.5 (low
   weight); Condition 2 would be all those with weight>0.5. You
   should pick your own weight threshold of course.

   doug


       Katie

       
          doug




          Katie Bettencourt wrote:

              Yes, those are the maps I"ve been comparing.  I've been
              comparing it to BV sort of, but that analysis is not
       surface
              based and I"m not used to it, so I can't quite tell
       which is
              more accurate, though Version 2 gives a much smaller
       area of
              activity, which fits with the description of what I've been
              given about what to expect in BV.  Attached is two
       pictures of
              the difference I get for Version 1:1v0 (labeled with
       "single"
              in the image name) and Version 2:2v0 (labeled with
       "double" in
              the name).  As you can see, Version1 activates a much
       larger
              area than Version 2.

              I guess part of my problem is that I'm having trouble
              understanding exactly what these two versions are
       telling me
              about the data and what the differences is.  Can you try to
              give me a sort of layman's explanation?

              Katie


              On Tue, Feb 22, 2011 at 11:45 AM, Douglas N Greve
              <greve@nmr.mgh.harvard.edu
       <mailto:greve@nmr.mgh.harvard.edu>
       <mailto:greve@nmr.mgh.harvard.edu
       <mailto:greve@nmr.mgh.harvard.edu>>
              <mailto:greve@nmr.mgh.harvard.edu
       <mailto:greve@nmr.mgh.harvard.edu>
              <mailto:greve@nmr.mgh.harvard.edu
       <mailto:greve@nmr.mgh.harvard.edu>>>> wrote:

                 I assume that you are comparing maps of Version1:1v0 and
                 Version2:2v0 ? I could imagine it going either way.
       If the true
                 slope is 0 but the offset is non-0, then Version1
       will give
              you an
                 artificially high slope (and Verion2 will give you
       the correct
                 slope at 0, and so no activation). Are you comparing
       this
              to a BV
                 analysis?

                 doug


                 Katie Bettencourt wrote:

                     So I created a weighted regression analysis to
       look at the
                     effect of memory load in a particular brain region.
              Basically,
                     I weighted the paradigms by a behavioral measure
       that
                     reflected the number of items actually
       remembered (as
              set size
                     was increased).  As far as Doug told me there are
              basically 2
                     ways to weight your paradigm files.

                     Version 1:
                     Have 2 conditions, baseline (condition 0) and
       all the set
                     sizes (condition 1).  Condition 1 would then be
       weighted by
                     the behavioral measure.

                     Version 2:
                     Have 3 conditions, baseline (condition 0), and
       then I
                     represented each presentation as two different
              conditions, one
                     with a weight that is always 1 (condition 1),
       the other
                     weighted according to the behavioral measure
       (condition 2).


                     The difference, as far as I understand it, in
       version
              1, it is
                     assumed that the response amplitude is ) when the
              weight is 0
                     (ie. that when you are attending to 0 items, brain
              activity =
                     0).  Whereas, version 2, tests the slope of the HRF
              amplitude
                     vs weight without the assumption above.

                     However, I'm a bit confused as to the results I got.
               When I
                     looked at the data from both versions, version 1
       provided a
                     much higher amount of activation and more areas
              activated than
                     version 2.  However, I believe version 2 better fits
              with the
                     multiple regression analysis that is done in
       Brain Voyager.
                     Can anyone give me a better explanation of what the
              difference
                     between these analysis models is?

                     Katie




                 --     Douglas N. Greve, Ph.D.
                 MGH-NMR Center
                 greve@nmr.mgh.harvard.edu
       <mailto:greve@nmr.mgh.harvard.edu>
              <mailto:greve@nmr.mgh.harvard.edu
       <mailto:greve@nmr.mgh.harvard.edu>>
              <mailto:greve@nmr.mgh.harvard.edu
       <mailto:greve@nmr.mgh.harvard.edu>
              <mailto:greve@nmr.mgh.harvard.edu
       <mailto:greve@nmr.mgh.harvard.edu>>>

                 Phone Number: 617-724-2358 Fax: 617-726-7422

                 Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
       <http://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting>
              <http://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting>
                 <http://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting>

                 FileDrop:
              www.nmr.mgh.harvard.edu/facility/filedrop/index.html
       <http://www.nmr.mgh.harvard.edu/facility/filedrop/index.html>
                     <http://www.nmr.mgh.harvard.edu/facility/filedrop/index.html>
                        <http://www.nmr.mgh.harvard.edu/facility/filedrop/index.html>




                 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.



                     ------------------------------------------------------------------------


                     ------------------------------------------------------------------------


          --     Douglas N. Greve, Ph.D.
          MGH-NMR Center
          greve@nmr.mgh.harvard.edu
       <mailto:greve@nmr.mgh.harvard.edu>
       <mailto:greve@nmr.mgh.harvard.edu
       <mailto:greve@nmr.mgh.harvard.edu>>
          Phone Number: 617-724-2358 Fax: 617-726-7422

          Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
       <http://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting>
          <http://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting>
          FileDrop:
       www.nmr.mgh.harvard.edu/facility/filedrop/index.html
       <http://www.nmr.mgh.harvard.edu/facility/filedrop/index.html>
          <http://www.nmr.mgh.harvard.edu/facility/filedrop/index.html>



   --     Douglas N. Greve, Ph.D.
   MGH-NMR Center
   greve@nmr.mgh.harvard.edu <mailto:greve@nmr.mgh.harvard.edu>
   Phone Number: 617-724-2358 Fax: 617-726-7422

   Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
   <http://surfer.nmr.mgh.harvard.edu/fswiki/BugReporting>
   FileDrop: www.nmr.mgh.harvard.edu/facility/filedrop/index.html
   <http://www.nmr.mgh.harvard.edu/facility/filedrop/index.html>



--
Douglas N. Greve, Ph.D.
MGH-NMR Center
greve@nmr.mgh.harvard.edu
Phone Number: 617-724-2358 Fax: 617-726-7422

Bugs: surfer.nmr.mgh.harvard.edu/fswiki/BugReporting
FileDrop: www.nmr.mgh.harvard.edu/facility/filedrop/index.html