Here is an article that explores the issue:
http://www.audioholics.com/loudspeaker-design/speaker-break-in-fact-or-fiction First the article acknowledges and documents the changes that occur in a couple example speakers as they "break in". There are, in fact, changes that occur.
Then it models the resulting differences in performance that the broken in speakers exhibit when placed in the same enclosure. Note: this plot is of the two response curves; before and after.
This explains to me why speaker manufacturers inform us that their speakers do "break in" yet they don't seem concerned about our responding with questions regarding how "break in" causes deviation from the intended designs.
I remain curious about the subject, but I tend to think that, for my personal experiences, the effect of pyschoacoustics over powers anything quantifiable in a fact based domain.
best regards,
mike