And the winner is... BMC Research Notes!

At long last:

The Comment that IEEE-TPAMI did not want you to read!

https://www.biomedcentral.com/1756-0500/...


So to finally fill in all the blanks - about
a year ago, I saw a terribly invalid "registration" paper, in
IEEE-TPAMI of all places. Contacted Brian Avants; he had seen it too and
shared my concerns.

So we obtained the original data and
confirmed our suspicions that the proposed method was not doing
registration at all. To correct the record, we then wrote a Comment,
which we sent to PAMI.

Much to our surprise, our Comment was
rejected without review as "technically valid but insignificant." We
appealed but were turned down, with some rather interesting arguments.
Such as: "you are ignoring the assumptions made by the authors" (we were
using their own data!), and "the algorithm assumes intensity constancy -
it says in the title, 'intensity-based registration'."

Long
story short, we took the Editor in Chief's friendly advice, sent our
Comment elsewhere, and here it is. For your further entertainment and
education, you may also want to look up "Dunning-Kruger Effect" on
Wikipedia. :)