 |
 |
Dear Marko and Helmut, These are excellent points. I think thing I would add is that if normalization error is random and small, relative to the longitudinal signals of interest, then normalizing each session separately may not change the results much. But as you note, it is best to not have normalization error confounded with time. Best, Dan On Jun 29, 2014, at 11:58 PM, Marko Wilke <[log in to unmask]> wrote: > Dear Daniel, > > following up on Helmut's comments, I think it is essential to minimize the processing steps that introduce differences between sessions you want to compare. I believe it therefore to be critical to use the same spatial normalization parameters for both timepoints, and aim for maximum overlap in native space. As Helmut laid out, this is limited by differences in the data that you cannot account for using a rigid body approach, but it is still better (in my subjective assessment) than introducing definite further differences between timepoints (on top of those you are interested in). But as Helmut said, I think that an empirical validation is still pending :) > > Cheers, > Marko > > Mathalon, Daniel wrote: >> Dear Marko, >> >> What would be wrong with separately realigning within session and between session images from Time 1 to MNI space, and then doing the same with Time 2 images. The resulting data should be all be optimally normalized to MNI, and it seems like direct comparisons from Time 1 and Time 2 could be achieved by statistical modeling (at least at the second level) by using Time 1 and Time 2 normalized beta images or contrast images. >> >> Would be interested to hear why this might not be a viable approach. >> >> Thanks, >> Dan >> >> Daniel H. Mathalon, Ph.D., M.D. >> Professor of Psychiatry >> University of California, San Francisco >> >> Mail Address: >> Psychiatry Service 116d >> San Francisco VA Medical Center >> 4150 Clement St. >> San Francisco, CA 94121 >> >> Office phone: (415) 221-4810, ext. 3860 >> Fax: (415) 750 6622 >> e-mail: [log in to unmask] >> >> >> >> On Jun 27, 2014, at 8:20 AM, Marko Wilke <[log in to unmask]> wrote: >> >>> Hi Patricia, >>> >>>> We feel it is not reasonable to enter all 8 runs in the same realign >>>> with such a big distance. Or is it? >>> >>> You could try :) As mentioned by John in the past, between-session motion is handled differently from within-session motion, in that the algorithm will try to get all first (or mean) images from all sessions into one space in a first go. So it may actually work. >>> >>>> We tried to move the Time2 time series using the "reorient image" option >>>> in checkreg, so they would be in rough alignment with the time series at >>>> Time1 before the realign, but found out that as long as the images are >>>> not resliced, the realign ignores those changes. >>> >>> It should not. Have you actually applied the changes to all images? Also, you could try "standard realignment" for all sessions from each session and then coregistration of the first/mean image of the first session from Time2 to the first/mean image of the first session from Time1, *taking along all images* from all sessions from Time2. You can also use the anatomical images if you have one from each timepoint, as they will provide better details and may be less vulnerable to different dropouts etc. >>> >>>> Is it ok to reslice the >>>> images twice, once before and once after the realign? (we would like to >>>> use dartel to register to the MNI space). Or is there a better solution >>>> to our problem? >>> >>> Try the above. John has gone to great lengths to avoid instances where the data needs to be resliced; that alone tells us it is not a good idea :) >>> >>> Cheers, >>> Marko >>> -- >>> ____________________________________________________ >>> PD Dr. med. Marko Wilke >>> Facharzt für Kinder- und Jugendmedizin >>> Leiter, Experimentelle Pädiatrische Neurobildgebung >>> Universitäts-Kinderklinik >>> Abt. III (Neuropädiatrie) >>> >>> Marko Wilke, MD, PhD >>> Pediatrician >>> Head, Experimental Pediatric Neuroimaging >>> University Children's Hospital >>> Dept. III (Pediatric Neurology) >>> >>> Hoppe-Seyler-Str. 1 >>> D - 72076 Tübingen, Germany >>> Tel. +49 7071 29-83416 >>> Fax +49 7071 29-5473 >>> [log in to unmask] >>> >>> http://www.medizin.uni-tuebingen.de/kinder/epn/ >>> ____________________________________________________ >> > > -- > ____________________________________________________ > PD Dr. med. Marko Wilke > Facharzt für Kinder- und Jugendmedizin > Leiter, Experimentelle Pädiatrische Neurobildgebung > Universitäts-Kinderklinik > Abt. III (Neuropädiatrie) > > Marko Wilke, MD, PhD > Pediatrician > Head, Experimental Pediatric Neuroimaging > University Children's Hospital > Dept. III (Pediatric Neurology) > > Hoppe-Seyler-Str. 1 > D - 72076 Tübingen, Germany > Tel. +49 7071 29-83416 > Fax +49 7071 29-5473 > [log in to unmask] > > http://www.medizin.uni-tuebingen.de/kinder/epn/ > ____________________________________________________ >