It's not clear to me if the dementia was a direct cause of death here. Some losses that could not be truly recovered, yes, depending on the extent of cell deat

It's not clear to me if the dementia was a direct cause of death here. Some losses that could not be truly recovered, yes, depending on the extent of cell death. But much of the symptoms in Alzheimer's disease seems to be due to loss of function in the systems that, roughly speaking, drive the management of memories rather than those that store the memories themselves (hence, for example, possibility for relief of the symptoms in early stages by increasing cholinergic transmission), and neural memory is generally distributed, with loss of neurons or synapses translating into degradation of fidelity or acuity of the memory, rather than downright destruction. It could well be possible in principle to restore resolution to them from what remains of the representation and some assumptions, though it's certainly beyond what we can do today. So in other words, I claim that some of what seems lost may simply be inaccessible, and some of what really is lost is neither completely nor irrecoverably so.