The x-ray theory doesn't fit for two reasons. The defective ink would have to be x-ray fluorescent (artifacts have higher developed density). And the x-ray would have to hit the spooled film after it was shot (artifacts align in post-consumer position).
Reminds me I have a theory to test. I believe it's environmental moisture absorbed into the emulsion when the film is in camera and which softens the emulsion slightly. Then when film is wound to the take-up side, the moisture escapes through the hydrophilic background paper where emulsion re-hardens while moisture is trapped by the hydrophobic ink where the emulsion remains locally soft. Then when developed, the softer emulsion takes on developer more readily developing sooner/to a greater density than the re-hardened adjacent emulsion.
To test my theory, I reeled some film onto stainless reels and put them in a stainless tank. They've been waiting to be developed since 20-DEC-2018. I bet they have reached equilibrium by now.