Hi Nicolas
The "common random effects design matrix" expression
means that all subjects in your prospective study are expected in
advance to have the same random effects design matrix (Zi). If you
assume a linear trajectory over time and a two-year study with repeated
measurements every six months then this matrix can be:
1 0
1 0.5
1 1
1 1.5
1 2
In this case you are considering both intercept and time as random effects.
ZiCol
is the column of the above matrix that you are interested in (usually column 2,
slope of change over time).
To
compute sample size you also need estimates of the random effects
covariance matrix (Dhat), within-subject measurement error, effect size,
and drop-out rate. If your study is about Alzheimer-related hippocampal
volume or cortical thickness variations you can use our sample ADNI
data to compute those values (ADNI791_Hipp_and_Entorh.mat). Eg.
lme_plannedSampleSize(Zi,2,total_hipp_vol_stats.Dhat,total_hipp_vol_stats.phisqhat,1,.2,.8,0.05)
Alternatively,
you can perform a pilot longitudinal study with a small sample or you
can use information from previous studies that are similar to the study
that you are planning to run, or you can even analyse some public
available longitudinal data with lme to compute those values.
We performed sample size estimations for the study described in our paper in
the wiki page:
http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels
It may be useful for you to take a look at it.
Best
-Jorge
>________________________________
> De: nicolas guizard <n.guiz...@gmail.com>
>Para: freesurfer@nmr.mgh.harvard.edu
>Enviado: Jueves 11 de abril de 2013 14:28
>Asunto: [Freesurfer] Fwd: Sample size estimation from LME models
>
>
>
>Hi,
>I am new to Freesurfer and the Linear Mixed Effects (LME) Models packages. And
>I would like to compute sample size from volumetric measures for 2 groups.
>
>
>I followed the longitudinal example provided in the wiki
>(http://surfer.nmr.mgh.harvard.edu/fswiki/LinearMixedEffectsModels#a.29Univariate),
> with the ADNI example. But I am not sure what to use for the "common random
>effect design matrix" Zi and "column" ZiCol with the
>lme_plannedSampleSize(Zi,ZiCol,Dhat,phisqhat,effsz,dr,pw,alpha,gr_pr) function:
>
>
>Here is what I do in matlab:
>
>
>> load('ADNI791_Hipp_and_Entorh.mat')
>
>> total_hipp_vol_stats = lme_fit_EM(X_Hipp,[1,2],Y(:,1)+Y(:,2),ni);
>
>> lme_plannedSampleSize(total_hipp_vol_stats.X,[1,2],total_hipp_vol_stats.Dhat,total_hipp_vol_stats.phisqhat,1,.2,.8,0.05)
>
>> ans =
>
>
> 1.0e+07 *
>
>
> 1.3584 0.0154
> 0.0154 0.0100
>
>
>Is that right?
>
>
>Regards,
>Nicolas
>_______________________________________________
>Freesurfer mailing list
>Freesurfer@nmr.mgh.harvard.edu
>https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
>
>
>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.
>
>
>_______________________________________________
Freesurfer mailing list
Freesurfer@nmr.mgh.harvard.edu
https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer
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.
