Very quick physics lesson on hydraulic conductivity. (RE pre infusion behaviour). This is based on what I know of similar processes which occur on a larger scale. I am not an expert on coffee extraction, but I am an expert on using leaching to extract metals, such as gold, nickel, uranium and especially copper, from ore.
Dry material is more resistant to water flow than wet, so water will go through a wet puck faster. (And its the simplest reason why if channelling starts its not going to get better). Water will move through the puck by conduction (pressure force from above) and through diffusion/capillary action (Capillary action being a pressure load based on relative saturation).
This is assuming that any swelling of coffee particles is small. It is worth thinking that yes swelling may fill in some of the flow paths around the coffee grounds but they will also open up other paths more. Particle size effects is for another lecture... As is reaction/diffusion kinetics. I could also talk about carbon dioxide release but would have to be careful that I am not talking rubbish.
I will say that diffusion rates are always some variation of something which is initially very fast and then tails off slowly. Depending on what is dissolving (and other factors) you can get a fairly linear response for a while after the initial peak.
If you preinfuse a puck you are putting enough water in to diffuse through the puck to change its flow characteristics. At the same time the water itself will be in contact with the coffee particles and will be extracting the soluble compounds. This extraction will be very quick - the intial peak- as all the easiest compounds to liberate will be available, but will be limited by the volume of water and the solubility of the compounds. It could become saturated which will slow extraction.
When you then start the main extraction the first thing that happens is that the new water pushes out the preinfusion moisture, but beyond that its flowing through an evenly saturated puck (nice even flow properties) with (I am guessing) many of the main soluble compounds on the 'linear' part of the dissolution curve. I.e. its all very controllable and predictable.
Lesson done. Debate welcome, even if/especially if you can show me to be wrong