The main causes of dementia are probably Alzheimer's pathology (plaques and tangles), Lewy Body pathology (associated with Parkison's disease) and multiple strokes (atherosclerosis).
But much of the clinical pathology of dementia is very likely associated with delirium/ acute confusion added on top of the dementia: dementia greatly increases the susceptibility to delirium.
And delirium is, in principle, treatable, improveable.
Just think how often it is observed that demented patients get better from time to time (not completely better, but much improved), and have good days, or a good few hours.
This strongly suggests that these 'good times' are the clinical picture of dementia without the delirium - and that most of the time the patients are delirious.
In an old and degenerate brain, it takes but little to flip into delirium - drug treatment and interactions, an infection, dehydration, lack of sleep or disturbed sleep, trauma, underlying tumour, autoimmune disease, or subtle degrees of any disease - hormonal, of a major organ system, and so on..
If - and this may well be impossible - all these potential causes of delirium could be checked and treated, restored to normal - I suspect that much of the apparent 'dementia' would be relieved.
Yet perhaps simple things are possible.
Greater attention to encouraging regular and restorative sleep (in particular) and hydration, and much greater attention to the perils of drug treatment might help many specific individuals?
This matter of a super-added psychopathology in dementia is usually conceptualised in terms of 'depression' - yet this is misleading.
In the first place, mainstream psychiatry has no coherent concept of what it means by 'depression' - so this is not an explanation at all, but rather a suggestion for 'antidepressant' treatment.
If the problem is delirium, then many 'antidepressants' are likely to worsen the problem, due to their anticholinergic side effects - psychological drug side effects are much commoner and more significant (and harder to rule out) than commonly realized.
On the other hand, there are many records of specific demented patients (of various types) having benefitted very obviously from electroconvulsive/ electroshock therapy - and ECT very probably has a general anti-delirium effect:
In particular, ECT seems to help some patients with 'Lewy body'/ Parkinsonian dementia - where the fluctuating, delirious and psychotic component is especially marked.
The take-home message is that much of the clinical pathology associated with dementia is probably not caused directly by the irreversible dementia process but is secondary to the microscopic damage caused by the dementing process, and the result of potentially-reversible delirium - an acute confusional state which will sometimes clear away to reveal a much less-severely impaired person.
The measure of dementia pathology is the patient at their best, in a 'good spell' - all the rest is likely due to super-added delirium.