Myself, I tend to agree that some connection or the IC inside the disk
was being effected by the cold enough to allow an otherwise weak signal
to be read. It is hard to understand how a mechanical function could be
effected by the cold that would give the same results. However, I am
always willing to learn new things.

Last time I opened a bad hard drive, there was an IC right on the flex
ribbon that goes to the heads. If that is the read / write amplifier, that's
one component you cannot swap out by changing the pcb. Maybe it was a
thermal problem in that IC. Since you have recovered the data, tear the
cover off that bad boy and take a look see.
Regards,
Tom