Hi, Rookie-

Ah, I see what you mean: you want the Thr value (of a statistic, say) at the peak Olay value (of an effect estimate, say) within each cluster, and the peak effect estimate is not guaranteed to have the peak stat value. Fair enough. Indeed, I do think the "whereami" approach is generally more useful than condensing things down to a single value---I'm not sure that the peak statistic is even very meaningful after thresholding? (e.g., Clustering is a hypothesis-testing-like procedure, and so once things pass threshold, the exact statistical value shouldn't really matter much anymore?)

Addendum to my previous comment: another way to get the column 13 (in onebase) values would be to use AFNI subbrick selectors, as follows to get the same column in a 1D file:

1dcat o.cluster_report.txt'[12]'

Note that AFNI uses zero-based counting (e.g., as in C and Python), hence the number in square brackets.

So, if you did want the above, though, I think the way I would do it would be:
+ set Olay = effect estimate volume
+ set Thr = stat volume
+ Clusterize, get data table
+ From the data table, get the *locations of each Olay peak within the cluster, which would be columns 14, 15 and 16 (in one-based counting, which awk uses):

     MI RL        : Coordinate of the Maximum Intensity value in the
                    Right-Left direction of the volume cluster
 
     MI AP        : Coordinate of the Maximum Intensity value in the
                    Anterior-Posterior direction of the volume cluster
 
     MI IS        : Coordinate of the Maximum Intensity value in the
                    Inferior-Superior direction of the volume cluster

... and you could extract those coordinates from the table with 1dcat, too, note the column numbers:

1dcat o.cluster_report.txt'[13..15]' > o.peak_coords.txt

Then, you could loop over those coordinates and extract the value from the desired dataset at that location.

For example, you extract the value for the first/zeroth coord as follows:

# get 3 coords (likely in RAI notation), and pass those into 3dUndump to make a temporary mask of hopefully 1 voxel
1dcat o.peak_coords.txt'{0}' | 3dUndump -master DSET_FOR_GRID -xyz -overwrite -prefix tmp_mask -
# use that temp mask to extract a value from a dset of interest, such as a stat volume
3dmaskave -quiet -mask tmp_mask+tlrc. DSET_OF_INTEREST

I think those papers would be useful to look at. The danger is that condensing a whole cluster down to one voxel value, esp. a stat, can lose a lot of valuable information.

--pt