When playing recordings through a speaker, the voice coil in the driver does heat up and there is slight movement in driver itself so it is possible for the parameters of the loudspeaker driver (Thiele/Small) to change slightly when in use. This has been shown by using extreme measures such as playing a sine wave through a driver for specific periods of time. It is a extreme test however it really doesn't prove anything. I can put a 10hz sine wave through the driver and I can tell you it will definitely cause "break-in".
The two problems I have with the pseudoscience behind break-in is knowing the exact time period the driver parameters have changed and then claiming it to an improvement in sound quality (can't it degrade sound quality? no change at all?). We never hear about the latter claims.
Secondly, with a well designed driver when at rest for a period of time (and it is not blown or defective) and measure the driver lets say after 24 hours of rest (not just for 1 minute or an hour after use), the Thiele/Small parameters will revert back to their equilibrium state (or very close to). This final measurement is what all these break-in tests lack. In real-world conditions we might have 12-24 hour breaks or more between our listening sessions.
