I'm not sure anyone has a detailed explanation of exactly what is happening when seating depth is used to tune in grouping. I've heard a lot of different theories over they years as to different things it might be...such as barrel timing or entry into the lands/bullet engravement. Certainly there are arguments that seating depth could affect both, as well as arguments that neither one is the sole factor affected by seating depth.
One thing to remember about the notion that seating depth affects barrel timing/harmonics is that the bullet is moving infinitely slow in the first few instants after the weapon is fired. Because of that, even the extremely small [relative to barrel overall length] increments we use to test seating depth may have a profound affect on barrel timing.
As a general rule, I have not placed a great deal of emphasis on trying to understand exactly what we're accomplishing with seating depth optimization for the purpose of trying to predict how to optimize it. It is far more easy [and reliable] to directly test seating depth and let the rifle tell us where it wants a certain bullet seated than to try and guess what might happen to seating depth optima as charge weight/velocity are altered. For example, changing seating depth for jumped bullets by as much as .010" to .015" in either direction will generally cause a change in velocity due to altering the internal case volume by an amount that cannot be reliably detected by most chronographs...i.e. the change in velocity will be less than the ES/SD. So changing charge weight by even 0.1 to 0.2 gr will generally cause a more noticeable change in velocity than a large change in seating depth in that you can usually detect a significant change in velocity for that change in charge weight. Nonetheless, a change in seating depth of even .001" to .002" can have a huge impact on grouping. More often than not, I find the seating depth optimum for a given bullet doesn't change much if I adjust the charge weight very slightly in either direction, or the velocity is changed due to temperature differences on different days. However, I shoot F-Class, where it is not necessary to have a load that shoots 0.1 MOA (or less) to win. The people that can probably give you the best anecdotal information about this topic are the short range BR shooters, whom measure and change such parameters regularly, and record the results religiously. Perhaps a few will chime in on this topic.
Anyhow, if you are wanting to see any relationship between velocity and seating depth more clearly, try setting up a charge weight versus seating depth array, where you test increasing charge weight (vertical dimension) at several incremental seating depths each (horizontal dimension). That way, you can observe how both parameters affect grouping on the same target. For statistical reasons, you might actually want to shoot a couple identical targets.