One of the most in-depth studies of dementia and its causes has been carried out in France by Paquid, a group of epidemiologists based in Bordeaux.
It has been studying 4,134 people aged 65 in the Gironde and Dordogne since 1988 in a piece of research due to end this year.
Paquid has found a powerful association between academic achievement and the risk of dementia, and it believes the higher incidence of dementia in women aged over 75 could be explained by a lack of academic attainment.
But researchers in 1999 found that even after academic levels were taken into account, the increased risk of dementia among elderly women remained significant. This, it concluded, depended on their social activities and hobbies.
Gardening, travelling and doing odd jobs around the house were associated with a lower risk - things that required planning and initiative - yet not other activities such as reading, parlour games or caring for grandchildren.
"Le travail, c'est la sante," the French saying goes. The notion that an active mind is good for health continues to exercise clinical and psychological researchers world-wide, particularly the US-based Alzheimer's Association.
Among the research it has funded is a piece by Dr Tammy Hopper of the University of Arizona on the theme of "Maximizing the function of individuals with Alzheimer's disease through cognitive stimulation".
Evidence exists to show that Alzheimer's does not cause the same decline in all thinking skills. While event-related memory loss is affected early on and progressively worsens, the knowledge of concepts or words may not suffer as soon or as severely.
Some research has shown that those knowledge and thinking skills which are less affected may be tapped to improve the quality of life for people suffering from dementia.
Dr Hopper has concentrated on a learning method called "spaced-retrieval training" in which an individual is asked to recall a particular piece of information after increasingly longer time intervals.
The method will also be tested for correcting impaired behaviour, such as learning to recognise a cue for using the toilet regularly. Data may help researchers develop strategies for helping people with dementia to learn behaviour and associations for an improved quality of life.
Meanwhile Dr Kenneth M Langa, of the University of Michigan, has been funded to identify the costs and consequences of dementia in America.
There, Alzheimer's is projected to triple in the next 50 years, which will throw an enormous burden "on the formal health-care system and on informal caregivers who typically provide most of the care".
The aim is to find out how the growing costs of dementia are distributed across public payers, private insurers and families, and across racial and socio-economic groups.
A further summary of evidence suggesting that frequent mental activity in old age is associated with reduced risk of Alzheimer's disease appeared last year in MedALERT, a clinical journal review service produced in New Zealand.
A study was made of 801 older Catholic nuns, priests and brothers without dementia, examined annually for up to seven years, and assessed for the time they spent on common activities that "involve information processing as a central component".
These included watching TV, listening to the radio, reading newspapers, magazines and books, and playing games such as cards, chequers, crosswords or other puzzles. Visiting museums was another activity that was used.
After three years, 111 developed Alzheimer's, yet those most mentally active at the outset were 47 per cent less likely to develop the disease than people with "infrequent activity".
Researchers recognised, however, that a limitation of the study was that those in religious orders differed in education, lifestyle and other ways from other older groups.
And the need to study more diverse cohorts was emphasised.
* Develop methods for identifying the annual direct medical costs and informal care-giving costs for the elderly.
* Determine the relationship between severity of cognitive impairment and annual direct medical and informal care-giving costs
* Develop models for estimating and predicting decline, length of time from diagnosis to institutionalisation and death, and the lifetime direct medical and informal care-giving costs associated with cognitive impairment.