While I cannot prove this, logic suggests it's probably true. I have long believed that neck tension, powder charge shape differences ( flat or toward front or back of case) and jump can both create pressure and velocity variations. I have seen similar results to yours. I think bullets from round to round move into the lands at different times.
There have been numerous cases of modern smokeless 45 cal muzzle loader barrels blowing up. Many of us feel the bullets vibrate off the powder column during handling. This creates a pipe bomb. I believe the same thing to be true with some loads in centerfire rifles and can get far worse with jump.
I believe that bullets, particularly heavy for caliber bullets, can jump out of the neck on ignition, then jam into the lands and stall. This gives the powder more time and space to burn and build pressure The pressure builds way past normal levels as it tries to re launch the bullet that is now .020" or more into the lands. For that round you are basically shooting a round that is the same as a jammed load.
this condition varies from round to round and in my opinion can get far worse as jump increases. In a perfect world we imagine a bullet launching like a missile. In this case the bullets stutters.
Ive chased similar moving pressure signs at lower than max loads too in a 50bmg. I trimmed several cases, turned the necks, used a bushing and crimped all the loads. Never saw another spike.
I don't know about your procedures but I am very fussy about inside neck wall prep. Every case is cleaned, sonic washed, I run a rotary brush with acetone inside my necks. My goal is to create exactly the same resistance and tension from round to round.