Hi Joy-Loi,

yes, you can do that using the free surfer tools (e.g. check if 'atrophy', i.e. change across the two methods, is different from zero, meaning if there is a bias).

Any of the statistical methods described here: http://surfer.nmr.mgh.harvard.edu/fswiki/LongitudinalStatistics can do that.

You will find that results differ, but you will not know if they differ because of the - different quality of the scans - difference in processing (averaging 2 scans) - or simply the noise one would always expect from the scanner (imaging / positioning etc).

It would be interesting to distinguish those sources of variability. But for that you'd need a second time point I think.

Best, Martin

On Mar 19, 2013, at 10:47 AM, Joy-Loi Chepkoech j.l.chepkoech@psykologi.uio.no wrote:

Hello Martin,

Thank you for your swift reply.

I failed to make myself clear, and will try describing my study again. First, an important point: the study is not longitudinal per se, but I am interested in applying the methods of longitudinal analysis. I only have one time point, all the scans have been collected within the same session.

I am studying a large group of subjects whom each underwent a scan session with 4 scans. Each of these scans were graded for quality on a scale from 1-4. An example of what I then did was to process each subject twice, once with the scan of quality ”1”, and a second time averaging the scans of quality ”1” and ”2”, thus resulting in two groups between which each subject is compared to itself.

Now I am attempting to set up a longitudinal analysis with this data. I have processed through the longitudinal stream, and I guess my question starts at the point of post-processing: How, if feasible, can I transfer the idea of “across group comparison between time point and time point 2” to “across group comparison between 1 scan and 2 scans” in setting up the QDEC table and continuing from there?

I hope this make sense, your help is greatly appreciated.

Best regards,

Joy-Loi

On 2013-03-18 19:13, Martin Reuter wrote:

...

Hi Joy-Loi, so you have 2 time points? the first with a single scan, the second with two within-session scans that you want to average? I don't understand exactly what you are trying to do. To test differences in reliability you'd have to scan a bunch of subjects twice in a session and then twice in another session. Then you can look at reliability when using only one scan (e.g. the first) or when averaging both scans in both time points. 'varying quality' with respect to motion also scares me a little. There can be individual outliers that severely affect your mean response (unless you have lots of subjects). Anyway for these things you would process your results through the longitudinal stream (this allows you also to be unbiased with respect to the time point). There are several design questions to consider (e.g. does it make sense to include all virtual time points into the same base, or run everything twice: one base for the averaged images , another for the single images and then compare etc. ) Either way this is not a standard analysis and you'd be pretty much on your own with how to set it up. Best, Martin On 03/18/2013 09:02 AM, Joy-Loi Chepkoech wrote:

...

Hello FreeSurfer experts, I am currently trying to systematically investigate the differences in cortical and subcortical estimates that occur when performing recon-all on one scan, or when averaging across scans of varying quality with regard to movement. I have a list of subjects that each have been processed (recon-all) twice (once with one scan and once with two scans), and would like to run some FreeSurfer statistical analysis on them. Is it feasible to set up a longitudinal analysis similar to the one described here: (http://surfer.nmr.mgh.harvard.edu/fswiki/FsTutorial/LongitudinalTutorial)? If it is, could someone assist me in setting this analysis up? (I have tried following the tutorial, where I on each subject set "0 years" for one scan, and "1 year" for two scans in the QDEC table, but the results from the QDEC analysis don't seem to agree with previous SPSS results that I have). Thank you, Joy-Loi _______________________________________________ Freesurfer mailing list Freesurfer@nmr.mgh.harvard.edu https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer

-- Dr. Martin Reuter Assistant in Neuroscience - Massachusetts General Hospital Instructor in Neurology - Harvard Medical School MGH / HMS / MIT A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129 Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu 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.

--------------------------------- Dr. Martin Reuter Assistant in Neuroscience - Massachusetts General Hospital Instructor in Neurology - Harvard Medical School MGH / HMS / MIT

A.A.Martinos Center for Biomedical Imaging 149 Thirteenth Street, Suite 2301 Charlestown, MA 02129

Phone: +1-617-724-5652 Email: mreuter@nmr.mgh.harvard.edu reuter@mit.edu Web : http://reuter.mit.edu