History of AFNI updates  

|
bob cox
July 26, 2003 05:21PM
Given a z-score, you can compute the p-value for that score in 1 of 2 ways, using the AFNI programs. The first way is manually, 1 score at a time, using the "cdf" program, as in

cdf -t2p fizt 2.739

which returns the output line "p = 0.00616264".

You can also do this for each voxel in the z-score dataset, using the function 'fizt_t2p(a)', where 'a' is the input dataset symbol. You should probably force the output dataset to be stored in float format, not as scaled shorts, since p-values take a large range of values. You do this by using the '-datum float' option on the 3dcalc command line. Once you have calculated this p-value dataset, you could then use 3dmaskdump to extract the values you want.

Finally, if you are very advanced, you could actually calculate the p-value dataset on the 3dmaskdump command line directly, as in

3dmaskdump -mask ext+tlrc '3calc( -a zsc+tlrc -datum float -expr fizt_t2p(a) )'

This will use 3dcalc to compute the p-value dataset into a temporary file, read that dataset in, delete the temporary file, then process it. In this way, you don't need to delete it yourself -- what fun!?

bob cox
Subject Author Posted

writing out z-scores

Lukas Pezawas July 25, 2003 12:09PM

Re: writing out z-scores

Christine Smith July 25, 2003 01:16PM

Re: writing out z-scores

Lukas Pezawas July 25, 2003 02:50PM

Re: writing out z-scores

bob cox July 25, 2003 03:37PM

Re: writing out z-scores

Lukas Pezawas July 25, 2003 08:03PM

Re: writing out z-scores

Lukas Pezawas July 25, 2003 08:17PM

Re: writing out z-scores

bob cox July 26, 2003 05:21PM

Re: writing out z-scores

Lukas Pezawas July 26, 2003 05:43PM

Re: writing out z-scores

Tom Ross July 25, 2003 04:48PM