I assume this means that the pattern of moisture gain or loss is not always consistent and depends on the wood's starting point moisture level?
In essence, yes. Briefly, it depends on the whether the wood is taking up or losing moisture.
This is contrary to how most people perceive MC changes in wood actually work but there's good reason for this as you'll read below.
But if someone has a diagram or could otherwise explain this in more detail that would be very helpful.
Source: Some Aspects of Wood Moisture Relations, Arno P. Schniewind. 1956 (!)
This is the clearest diagram of it I could find; there are much more complex numerical tables out there with the same basic info but I think this makes what's called the hysteresis loop easiest to grasp.
The best summary of this I've read comes from opening page of the same document as the above table:
...it is clear that two of the factors which will influence E.M.C. are temperature and relative humidity. In practical application, these are usually assumed to be the only factors. Although this assumption is justified, there are nevertheless several other factors which can also influence equilibrium moisture content. One of these is that we are dealing with wood, a biological material, which has a great deal of inherent variation. We can never expect one individual piece of wood to behave exactly like any other piece. Another factor is the species of wood which we are dealing with. Although the influence of species is not very pronounced and may be overshadowed by the variation within one species, these differences nevertheless exist. And finally, a factor which cannot be entirely discounted is sorption hysteresis.
Now despite the above it's clear that a simple equilibrium table gives a good enough picture of the MC you can expect at any given temp and humidity, for all species. Additionally, most designs in wood allow for any reasonable (and sometimes exceptional) changes in the MC of the wood, so to a certain degree it then doesn't matter what the MC ends up being.... and regardless of whether it's higher or lower than expected.
So, much as I hate to say it you can pretty much ignore this. As most woodworkers (of all levels) are unaware this even exists, therefore don't do anything to attempt to compensate for it and yet problems specifically related to hysteresis don't crop up in real-world applications, it's a fairly safe bet that you too don't need to worry about it.
To read more on this and related issues see chapter 4 of Wood as an Engineering Material, a Forest Products Laboratory publication referenced many times in previous Answers.