Hi Gang,

Thanks!

> 3dcalc -a stats.Imagination001_REML+orig'[4]'
> -expr 'fitt_t2p(a,27)/2' -datum float -prefix
> brick4p_one_sided

This will not do it. I guess it has to be something like:

if the t-statistic is negative, then the value is 1 - fitt_t2p(a,27)/2 and if the t-statistic is nonnegative, then it is fitt_t2p(a,27)/2.

So, I think it should be something like?

3dcalc -a stats.Imagination001_REML+orig'[4]' -expr 'ispositive(a) * fitt_t2p(a,27)/2 + isnegative(a)*(1 - fitt_t2p(a,27)/2)' -datum float -prefix brick4p_one_sided


Btw, it seems to me that it might be an useful value to report, as an option to the two-sided probability (for the t- and normal cases).

Is there a function which simply calculates the cdf to a probability? This would be equivalent to calculating the p-value.

> I still fail to see why you would explicitly need
> the p-values: the p-values for the popular
> statistics can be directly displayed on the AFNI
> viewer.

Yes, they can, but my objective is to use the values obtained to test methodology that we are developing.

Many thanks again!
Best wishes,
Ranjan



Edited 3 time(s). Last edit at 04/03/2015 10:37AM by maitra.