You should use a repeated-measures ANOVA to determine differences in
activity between conditions (or GLM Flex)
If you want to know where a specific condition is different than 0,
then you need to use a one-sample t-test (or GLM Flex).
The subtraction method is prone to both false negative (both
conditions different than 0, but also different from each other) and
false positives (one condition is barely significant and the other
condition is barely below threshold).
Best Regards, Donald McLaren
=================
D.G. McLaren, Ph.D.
Research Fellow, Department of Neurology, Massachusetts General Hospital and
Harvard Medical School
Postdoctoral Research Fellow, GRECC, Bedford VA
Website: http://www.martinos.org/~mclaren
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 Fri, Jul 20, 2012 at 10:01 AM, Kailyn Bradley
<[log in to unmask]> wrote:
> Hi all,
>
> I apologize for the rudimentary nature of this question. I feel like I
> should know this by know, but I could use a little help in explaining what
> the most correct or most accepted standard is for publishable fMRI
> statistics. In my current study, I have three conditions of interest I am
> looking at (novel words, group1 words, and groups2 words). I want to compare
> these three conditions to one another. At the first level, I created
> contrasts for each of these conditions for each of my subjects. Then at the
> second level, I did a one way ANOVA. I was also told that I could
> alternatively just do ttests at the second level to compare my three
> conditions to each other. Then I could do masking analyses where I mask out
> one condition from another instead of doing a subtraction type contrast
> through the ANOVA. My question is, which of these analyses is the most
> correct and accepted way to look at comparisons of this sort? Is it
> acceptable to use masking to look at differences in activation between two
> different ttests, or is this poor practice? Is the ANOVA the more acceptable
> way to go? Lastly, if it is ok to do these types of masking comparisons,
> what should the uncorrected mask p-value be set at? Is the .05 default fine,
> or should I use an uncorrected mask p-value of .001 in addition to p-value
> control threshold of .001 uncorrected?
>
> Any help explaining these differences to me would be greatly appreciated. I
> want to make sure the analyses I am doing are in line with publishable
> standards.
>
> Thank you,
> Kailyn
>
|