> when I compare grey matter between two groupe (control and patient)
> in contrasts result in corrected hight thershold if I choose
> uncorrected threshold
> the default value is .001 but change it to .01 after that I get
> aresult says that there is signficant differant between groupe ,in
> another way when I choose corrected threshold the default value is .05
> and leave it the same but I didn't find any signficant differant
> between two groupe,please whate is the differant between the corrected
> valu and uncorrected p value.
I would suggest reading (or skimming through)
http://www.fil.ion.ucl.ac.uk/spm/papers/SPM_5/SPM_5.pdf
For a simpler introduction, you could take a look at:
http://www.mrc-cbu.cam.ac.uk/Imaging/randomfields.html
The result of a t test relates to the probability of seeing your
data under the null hypothesis - which is usually that there is
no difference between the conditions. A large t or F value would
give a small p value, indicating that the probability of getting
that result by chance is very small. e.g. threshold of p<0.05
would mean that if there was no difference at that point, there was
only a 1 in 20 chance of getting the result. This is an uncorrected
result.
For the whole brain, there are many voxels. If results based on
the same threshold were used, then about 1 in 20 voxels would be
deemed significant by chance. Without a correction for multiple
comparisons, you would be reporting a lot of differences that
were not really significant. This is what a correction for
multiple comparisons is about.
If I asked you to pick a card from a pack, and I guessed what
that card was on the first guess, then you would be (mildly)
impressed (p=0.0192 - without jokers). If this was repeated
50 times, and I got it right on one of those occasions, then
you would not be at all impressed.
Best regards,
-John
--
Dr John Ashburner.
Functional Imaging Lab., 12 Queen Square, London WC1N 3BG, UK.
tel: +44 (0)20 78337491 or +44 (0)20 78373611 x4381
fax: +44 (0)20 78131420 http://www.fil.ion.ucl.ac.uk/~john
|