Hi Vasudev,
PALM can be used for this analysis. Since you already have the average GM for each subfield, assemble a table containing one subject per row and one subfield per column. Make sure there are only numbers in the table (no headers for rows or columns). Save as .csv, then supply it with the option "-i". The full call would be something as this:
palm -i table_subfields.csv -d design.mat -t design.con -n 5000 -approx tail -nouncorrected -corrcon -logp -o myresults
If you'd like to compare the subfields, then there are various ways to do this and which would lead to the same result. Perhaps the simplest is to subtract the columns for which you are interested on the differences, then use the design as a one-sample t-test (thus further including the option -ise when running).
Hope this helps!
All the best,
Anderson