 |
 |
Hi Darren and Will!
Now, i think that i got it!....many thanks for your notice and helps...mixing your helps were great
Regards\
Nima
----- Original Message -----
From: "Darren R. Gitelman" <[log in to unmask]>
Date: Fri, 21 Nov 2003 14:59:52 --0600
To: [log in to unmask]
Subject: Re: PPI ...confusion!
Re: Deaer Nima:
Re:
Re: Let me try to answer your questions.
Re:
Re: At 12:42 AM 11/20/2003, you wrote:
Re: >Hi
Re: >say we have 4 conditions :A,B, C, R
Re: >and we made a contrast : A-B
Re: >which gave us an MIP ..
Re: >
Re: >Next step is to produce VOI!
Re: >
Re: >and say that we chose a voxel in this MIP named X with radious r!
Re: >then spm GUI asks us "adjust data for"!
Re: >it gives us some possibilities:
Re: >a. don't adjust
Re: >b. effects of interest
Re: >c. A
Re: >d. B
Re: >e. C
Re: >f. R
Re:
Re: Adjusting the data means removing effects pertaining to the adjustment so
Re: Don't adjust your data.
Re:
Re:
Re: >say that the next step is PPI!
Re: >it asks for spm.mat and then the analysis type!
Re: >with 3 options:
Re: >a.simple deconvolution
Re: >b.psychophysiologic
Re: >c.physiophysiologic
Re: >
Re: >if we choose option b:
Re: >it ask for VOI images...
Re: >and spmget window will ask for "VOIs"
Re: >
Re: >second question: why we are not able to select more than one VOI?
Re: >and if we should choose one, why it asks for "VOIs" rather than "VOI"?
Re: >i can not understand why we should choose 1 VOI?!?!!?
Re:
Re: Hmm. That is a quirk of the interface. Actually you can only choose 1 VOI
Re: at a time, so it should really say Choose a VOI file.
Re:
Re:
Re: >the next step it ask us about "psychological variables"!
Re: >
Re: >third question: here is the point that i go exteremely confuesd! which one
Re: >we should select?
Re: >logically
, i think that we should choose the variable(s) that we did not
Re: >previously use them in creating our main contrast (A-B here)
Re: > and VOI as well!
Re: >otherwise, it will look like using a variable for doing double regression!
Re: >am i true?
Re:
Re: You should use whatever variables you want to make a statement about. If
Re: you chose A and B, for example and gave them contrast weights of 1 and -1
Re: then your inference would pertain to signal changes occurring in A more than B.
Re:
Re: >fourth question:
Re: >The most confusing part is results!
Re: >It gives us 3 tables:
Re: >1. hemodynamic and neuronal response (which is understandable)..
Re: >2. psychological variable (which again is understandable)...
Re: >3. PPI tabel..!!!!!...what does it mean? we compared "what" with "what"...
Re:
Re: I don't think there are 3 tables. Do you mean graphs?
Re: 1. The first graph
corresponds to the hemodynamic response = first
Re: eigenvariate that you extracted in the VOI. The neuronal response is the
Re: deconvolved first eigenvariate.
Re: 2. The psychological variable corresponds to the onsets of your events or
Re: your epochs.
Re: 3. The PPI graph shows the neuronal response multiplied by the
Re: psychological variable and the product convolved with the HRF. This is what
Re: the 1997 PPI paper showed. That paper used a block design and scaled the
Re: graph to a minimum of 0 instead of mean of 0 for illustration purposes.
Re: That paper also did not deconvolve the hrf which seems to matter less for
Re: epoch data.
Re:
Re: Now, all the PPI button does is set up the vectors for use in an analysis.
Re: It does not actually do the analysis. After the PPI process finishes you
Re: need to take 3 of the vectors, PPI.ppi, PPI.P, and PPI.Y and include that
Re: in an analysis.
Re:
Re: When
setting up the analysis choose 0 conditions. Then when it asks for
Re: user defined covariates answer 3 and enter the 3 vectors above. Choose the
Re: images, high-pass filter, etc, as usual.
Re:
Re: When the analysis finishes your contrasts will be [1 0 0] or [-1 0 0], for
Re: activations either positively or negatively related to your ppi vector. The
Re: resulting contrast images can be entered into a second level analysis just
Re: like any other contrast images.
Re: Good luck,
Re: Darren
Re:
Re: -------------------------------------------------------------------------
Re: Darren R. Gitelman, M.D.
Re: Cognitive NNeurology and Alzheimer¹s Disease Center
Re: Northwestern Univ., 320 E. Superior St., Searle 11-470, Chicago, IL 60611
Re: Voice: (312) 908-9023 Fax: (312) 908-8789
Re: -------------------------------------------------------------------------
___________________________________________________________
Sign-up for Ads Free at Mail.com
http://www.mail.com/?sr=signup