Print

Print
Hi Almudena,

If I understand correctly, this comparison will in fact investigate the heritability of the fMRI response to the verbal fluency task. There are various methods and tools available, and I'd suggest that you begin reading about OpenMx (http://openmx.psyc.virginia.edu), SOLAR (http://solar.txbiomedgenetics.org) and SOLAR-Eclipse (https://www.nitrc.org/projects/se_linux).

If you had DZ pairs instead of unrelated individuals, a possibility would be to use Falconer's method to compute heritability (http://en.wikipedia.org/wiki/Falconer%27s_formula), or, as a poor-man's solution, compute the intra-pair differences (absolute values), then run a 2-sample t-test in randomise comparing the MZ versus DZ. However, without DZ as controls, only unrelated subjects (the matching by age, education, etc., don't help), I believe that the best would be to model the family structure (and the lack of, for the controls) in SOLAR, but in any rate, not something that could be done in FSL as far as I know.

All the best,

Anderson



Am 04.03.14 15:34, schrieb Almudena Perez Lara:
[log in to unmask]" type="cite">

Hi Anderson,


I am interested in studying activation in Broca´s area during a verbal fluency task in twins, to assess if there is a similar pattern (in distribution and size of the clusters) between the paired twins and a paired control group.


Thank you very much,


Almudena Pérez


Date: Mon, 3 Mar 2014 22:40:21 +0000
From: [log in to unmask]
Subject: Re: [FSL] Statistical set up
To: [log in to unmask]

Hi Almudena,

Could you provide more details on your study? In particular, what are your hypotheses, i.e., what you'd like to investigate. Thanks!

All the best,

Anderson


Am 02.03.14 18:24, schrieb Almudena Perez Lara:
[log in to unmask]">

Dear FSL experts,


I am working on a fMRI project to study language in twins. I have 10 couples of twins, and a control group, also of couples, of non-twins matched for age and educational background.


I am having difficulties in setting up the correct statistics.


I was thinking of running in FEAT a separate independent t-test for twins’ pairs and the controls’ pairs, and then do a second high-level analysis running another t-test between them.

 

I am not sure if conceptually this is correct. Would you recommend me any other different approach?


Thank you very much for your attention.


Almudena Pérez