The control condition helps for a number of reasons:
(1) it allows you to ignore baseline shifts due to the drug (e.g. Drug
increases or decreases general neural activity;
(2) it makes it a within-subject comparison which almost always has
more power than between subjects(e.g. If A increases by 5 in one
subject and 2 in another, then you will have large variability;
however if B increases by 4 and 1, respectively, there is low
variability for A-B;
(3) without knowing your design, it's hard to interpret the implicit
baseline and whether it is tied to any cognitive state or tied to
statistics. The latter occurs when you have all your data occurring in
your two conditions. If this is the case, then the baseline is
completely arbitrary and would easily explain your results.
On Wednesday, February 23, 2011, Israr Ul Haq <[log in to unmask]> wrote:
> Dear Spm users,
> I am trying to see treatment effects on a patient, by putting his pre and post treatment sessions (all the runs separately) in one fixed effects analysis, and defining post – pre contrast (t) by giving positive one to the experimental condition in the post treatment runs and negative one in the pretreatment runs. This is all being done at the ‘specify first level’ option in spm8. It made sense and the few people I did ask seem to think this was okay too. However I came across something that has me confused and it would be great to hear an explanation. I had been including the control condition in the contrast too, so for the post treatment runs it was included as a negative one and for the pretreatment runs a positive one (just how it’s supposed to be contrasted out of the experiment condition at each session level, by giving it a weightage equal and opposite to the experiment condition).
> I thought I was getting a reasonable result from the contrast till I fortunately or unfortunately tried a contrast without the control condition, which is positive and negative ones only for the experimental condition, post and pretreatment respectively. This now is giving me FAR less activation then the contrast with the control condition, which is opposite to what i was expecting, since I though the whole point of including the control condition was to subtract activations unrelated to the process of interest, making the result purer so to speak. Since I have a medicine background and limited statistical knowledge, it would be great if you can point out whether this is possible and adding the control condition into each run somehow increases the t statistic when the analysis is done across sessions, or if I am doing something wrong. My t contrast for the fixed effects is:
> -1 1 -1 1 1 -1 1 -1 . where the sequence of runs put into the model specification part is pre tx run1, pre tx run2, post tx run1 and post tx run2, and each run has an experimental and a control condition. Will extremely appreciate help in this matter.
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.