 |
 |
You need to analyze the two groups separately since they have different slice thickness that could influence the measured signal. As for needing all subjects at each voxel measured. We are in the testing phase of a toolbox that will allow a variable number of subjects at each voxel, while maintaining the SPM framework. Briefly, we generate a different model for each set of different subjects. Then we compute the pooled variance based on all voxels with these subjects. Finally, we adjust the t-/f-statistics so that they have the same df as if they had all the subjects (simple t/f --> p-value --> t/f does the trick) Best Regards, Donald McLaren ================= D.G. McLaren, Ph.D. Postdoctoral Research Fellow, GRECC, Bedford VA Research Fellow, Department of Neurology, Massachusetts General Hospital and Harvard Medical School Office: (773) 406-2464 ===================== This e-mail contains CONFIDENTIAL INFORMATION which may contain PROTECTED HEALTHCARE INFORMATION and may also be LEGALLY PRIVILEGED and which is intended only for the use of the individual or entity named above. If the reader of the e-mail is not the intended recipient or the employee or agent responsible for delivering it to the intended recipient, you are hereby notified that you are in possession of confidential and privileged information. Any unauthorized use, disclosure, copying or the taking of any action in reliance on the contents of this information is strictly prohibited and may be unlawful. If you have received this e-mail unintentionally, please immediately notify the sender via telephone at (773) 406-2464 or email. On Tue, Apr 26, 2011 at 8:35 AM, Jonathan Peelle <[log in to unmask]> wrote: > Dear Jonas, > > > We initially used a FOV in our EPI sequence resulting in some (larger) > brains being cut off at the parietal and frontal lobules. We then increased > the slice thickness to account for this. The problem is when we do the > second level analysis; only voxels which are present in every subject seems > to be included, resulting in a loss of data mainly in the parietal lobule > which we are interested in. Is there any way to include voxels that are only > present in some subjects (treating the rest as missing data points) or do we > have to throw away some of our subjects? > > I can think of two options: > > 1) Discard the subjects with partial-brain coverage, and report all > analyses on the same group of subjects (that include parietal regions > you are interested in); > > 2) Perform two sets of analyses: one on the full group (for regions > that all subjects have data), and a second restricted to the subset of > subjects that have full brain coverage. > > I think (1) will be much more straightforward. It's appropriate that > SPM enforces data in all subjects for second level analyses, because > this ensures the correct number of degrees of freedom, as well as > having identical statistical models for all voxels. > > Also, it sounds like you may have changed your scanning sequence to > achieve a larger FOV—does (1) also mean that all subjects will have > received the identical scanning sequence? This would also probably > make interpretation slightly easier (as you woudn't have to be > concerned with any potential differences introduced by scanning > sequence). > > Hope this helps! > > Best regards, > Jonathan > > -- > Dr. Jonathan Peelle > Department of Neurology > University of Pennsylvania > 3 West Gates > 3400 Spruce Street > Philadelphia, PA 19104 > USA > http://jonathanpeelle.net/ >