A great many explanations can be offered for what MAY be occurring when cable breakin effects are perceived, including those cited in the Cardas paper Mofi linked to, and in the IMO well stated comments that were presented by Mr. Rodman (which I don't see any basis to dispute).
The problem is that none of the explanations that tend to be offered by cable manufacturers or others ever seem to be backed up by analysis establishing in a QUANTITATIVE manner that the claimed effects have a reasonable chance of being audibly significant. In fairness, though, perhaps it is generally not practical or possible for that to be done.
Thus, IMO there is no way to tell how many reports of perceived cable breakin are due to any one or more of the various technical effects that are alleged to occur; or to the placebo effect (coupled with the power of suggestion of the many claims that can be found in manufacturer literature and on the 'net); or to the vagaries of human perception that Onemug alluded to; or, perhaps most significantly in many cases, to failure to recognize and control extraneous variables (for example, assuring identical states of equipment warmup during the different parts of the assessment, ruling out the possibility that changes in AC line voltage or noise conditions may have occurred, ruling out the possibility that ongoing aging or breakin of system components may be occurring, etc.).
My bottom line instinct is that cable breakin effects are real, to some degree, but their significance is exaggerated in the minds of many audiophiles by these other factors.
Regards,
-- Al