Sort of. It's basically the same idea as before - using the range selector. We also have to change the clustering to sort based on cluster intensity. In the script below, I chose the output to be based on cluster mean, but you may want the cluster maximum instead, or to account for negative activations too. See if this short script is what you want.

#!/bin/tcsh
# clust_mean_max.csh
#
# extract only the cluster with the largest mean for a min vol and rmm
# usage:
# clust_mean_max.csh dset+tlrc thresholdmin
#
#
set dset = "$1"
set thresh = $2
set rmm = 1
set vmul = 20
set bname = `@GetAfniPrefix "$dset"`
set bview = `@GetAfniView "$dset"`

# remove any previous output if any
rm ${bname}_clmean${bview}.*
rm ${bname}_clmaxmean${bview}.*

# put mean value of cluster in each voxel of cluster output
# ...may want 1clust_max or 1clust_amax instead
3dmerge -dxyz=1 -1clust_mean $rmm $vmul -1thresh $thresh \
-prefix ${bname}_clmean "$dset"

# find maximum (most positive) value of cluster means
set maxmean = `3dBrickStat -max -slow ${bname}_clmean${bview}`

# extract just that value from dataset with small range for roundoff
set mm_min = `ccalc "${maxmean}-0.001"`
set mm_max = `ccalc "${maxmean}+0.001"`

3dcalc -a ${bname}_clmean${bview}"<${mm_min}..${mm_max}>" -expr a \
-prefix ${bname}_clmaxmean${bview}