On 08/05/2016 12:33 PM, Vikas Bandalli wrote:

Thank you Doug,

I am very sorry I hadn't explained our subject considered for analysis. The subjects group consists of Patients who are classified into a.Those with Psychosis - Psychotic-pt b. Those without Psychosis - Non-Psychotic-pt

We have Age and sex matched controls for every patient (case control study)

Now,the study was to see the structural variations using freesurfer and qdec, 4 individual 'qdec.dat' files were created such that two groups are considered at a time and analyzed namely

*qdec_1.dat :- Normal vs Patients(Psychotic + Non-Psychotic-pt)* *qdec_2.dat :- Normal vs Psychotic-pt (age and sex matched)* *qdec_3.dat :- Normal vs Non-Psychotic-pt (age and sex matched)* *qdec_4.dat :- Psychotic-pt vs Non-Psychotic-pt

The design was done as with Diagnosis/Psychosis as fixed factors and Age as covariate and ICV as nuisance factor for each 'pair' of analysis

Is the analysis-design appropriately constructed? or is this an inappropriate method.

so you created a new thickness file stack where you subtracted on an individual basis the paired subjects? That is fine.

The qdec GUI crashed with an error 'matrix m1=null' when the ICV values were used as it is (demeaning not done)

ICV definitely needs to be demeaned (and maybe rescaled by the stddev).

1.So,we can retain the negative numbered values for the above design. Also,I would be very interested to know why demeaning is necessary? My understanding is that since the ICV values are huge comapared to other factors

Correct

2.With the above mentioned study design,I wanted to know if the structural changes observed is due to the disease per se and not due to their Lithium intake (Some of the patients {both Psychotic-pt and Non-Psychotic-pt} were put on lithium). How would I have to design the analysis then? Is it possible to analyze this with qdecGUI?

To clarify, two diagnosis groups Psychotic Patients and Non-Psychotic Patients. Some of the patients are on lithium and some are not. Is that right? If so, then you have a two factor design where each factor has two levels, so you can do this analysis in QDEC.

Thank You,

Regards, Vikas Bandalli Raju

On Fri, Aug 5, 2016 at 9:28 PM, Douglas N Greve <greve@nmr.mgh.harvard.edu mailto:greve@nmr.mgh.harvard.edu> wrote:

1. It is not a problem that some of the demeaned ICV values are
negative; this is what happens when you demean a set of numbers;
you must leave the negative numbers there. Are you doing a paired
analysis or do you just have two groups? I don't think you can do
a paired analysis in QDEC.

2. would this give you 3 groups (controls, patients with lithium,
patients without)? If so, you cannot use QDEC for this.

If you can't use QDEC, you'll have to use the command line stream
and either set up your own FSGD file or your own design matrix


On 08/04/2016 01:36 PM, Vikas Bandalli wrote:

    Dear Freesurfers,

    Sorry if the question sounds too basic.

    1. I am running an analysis on qdec with paired groups -
    normal and patients,
    To control for IntraCranialVolume(ICV),I am adding it as an
    'nuisance factor',Is it correct?
    The suggested mode to add ICV to the tabled was to _demean_
    the ICV.
    But when I subtract the mean values from individual subject
    ICV values,I noticed that some of the subjects have a negative
    demeaned ICV value.
    My question is - How should I proceed to include the
    data,ignore the negative sign(because I thought a negative ICV
    makes no sense) and just keep the values or retain the numbers
    with the sign.
    Please correct me if I am going wrong in any of the steps.
    However,I ran two analysis - one with the sign retained and
    other with sign removed ,I noticed that both have very
    different results.

    2. If I want to see the effect due to lithium within the
    patient group (Between patients with and without psychosis)
    how should I design the matrix in GUI.
    I designed the following format
    Fixed factor :- Diagnosis - Psychotic-pt
                                          Non-Psychotic-pt
    Co-variate :- Lithium use :- 1 - lithium user
                                            0 - non user
    Nuisance factor :- ICV
    Is this design correct?
    If not how should I construct it ?
    Also what query would contain the result for anticipated question
    My intention is to control for Diagnosis and then see
    effect/change due to lithium use.

    Thank you very much




    -- 
    Vikas B. R.
    Medical student
    Bangalore Medical College and Research Institute (BMCRI),
    Bangalore,India,
    Ph : +918904286825


-- 
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: https://gate.nmr.mgh.harvard.edu/filedrop2
<https://gate.nmr.mgh.harvard.edu/filedrop2>
www.nmr.mgh.harvard.edu/facility/filedrop/index.html
<http://www.nmr.mgh.harvard.edu/facility/filedrop/index.html>
Outgoing:
ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/
<ftp://surfer.nmr.mgh.harvard.edu/transfer/outgoing/flat/greve/>



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
<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.

-- Vikas B. R. Medical student Bangalore Medical College and Research Institute (BMCRI), Bangalore,India, Ph : +918904286825