Up in the brain, everything is pretty much working by upregulating connections that are receiving input and downregulating connections that are not. (As the great Canadian Donald O. Hebb taught us, neurons that fire together wire together .) If you jump way up into the auditory cortex, the brain is still maintaining frequency mapping from down in the cochlea. If a particular frequency area is not getting input, then the other greedy areas around it will start to encroach. How much input that particular frequency area is getting, and thus how well it can maintain its mapping, is going to depend on the level of loss at that frequency out on the periphery. If you trust my memory, there is some evidence using reversible lesions in cats that if you allow those areas to atrophy and then reverse the lesion causing the reduced input, original mapping will sort of come back, but not precisely back to what it was. It’s a bit more disorganized, for whatever that means for function. Now in those experiments there was no actual damage to the auditory periphery, so reversal of the lesion restored normal peripheral function. Restoration of activation of those areas with amplification is probably going to be imperfect.
Now that I look back, though, are you talking about listening to audio over loud headphones akin to, say, part-time hearing aid use? AKA, How much part-time use does of amplification does one need to stave off any atrophy? (Or, rather than say atrophy, something like ‘neural reorganization’ seems a bit less loaded as the brain is just doing what it does best.) As far as I know that research area is still wide open. If you look at visual research–monocular deprivation in cats–you only need a much smaller percentage of normal binocular exposure to stave off the effects of monocular deprivation. Off the top of my head, I think it was something like one hour binocular experience per eight hours of monocular experience. Nature+Nurture: We are what we experience, but the brain is set up to ‘expect’ a certain ‘normal’ type of input and will weight it more heavily. One might suppose that something similar would happen in the auditory system, though what the absolute numbers are are anyone’s guess. My gut feeling is that the auditory system is a bit more plastic than the visual system, and therefore might need a higher percentage of normal input to maintain things.
Does this approach the sort of answer you were looking for?