Cognitive archaeology: identifying and measuring the presymptomatic phase of dementia

The onset of Alzheimer’s disease and other common causes of dementia is an insidious process, which typically does not become apparent to a patient, their relatives, or medical practitioners, until considerable brain tissue loss has taken place. This is due to the fact that the brain possesses a reserve capacity, which can be used to contnue normal levels of functioning. It is believed that dementia only becomes apparent when this ‘cognitive reserve’ has been exhausted. People’s have different cognitive reserve capacities, which probably explains why the prognosis, and speed of progression is so variable between individuals with established dementia. Understanding the factors that give rise to a large cognitive reserve capacity represents an important way of trying to control the expected population increase in dementia over the next 25 years. By looking at language records laid down over decades prior to the emergence of dementia, we hope to be able to trace patterns of change that indicate when, and over what time-course, the earliest indications of intellectual impairment occurred. By correlating differences in these measurements with physical, social and demographic information, should reveal important, and in some cases, potentially modifiable, factors associated with the size of the cognitive reserve, and therefore resilience to the effects of Alzheimer’s disease and other later-life neurodegenerative dementias.

Variations in lead-time between the onset of the earliest symptoms of cognitive inefficiency (or ‘mild cognitive impairment’ (MCI)) and clinical dementia have offered insights into the factors associated with the brain s capacity to maintain normal levels of cognitive function in the face of degenerative change ( cognitive reserve ). However, the diagnosis of MCI is one of expectation, phenotypically heterogeneous, and liable to a degree of self-selection bias. An alternative would be to use a retrospective approach, looking back at the earliest signs of cognitive change in patients who have already developed dementia. Language abilities are known to decline in early Alzheimer’s disease, and many individuals have left records in which evidence of change can be identified, measured and dated. That such records contain signs of developing dysfunction has been clearly shown by the findings of the Nun Study, and more recently the Iris Murdoch text analysis project. Measurement of preclinical language change at regular intervals in a larger population of patients will lead to robust quantification of the length of this phase, and allow examination of the factors associated with a long or short duration. This project will subject archived material, in both spoken and written formats, to a range of data-driven textual analyses that have been validated in the fields of textual classification and scholarship, with the aim of characterizing the nature and timing of the earliest changes in language processing abilities. Subjects will include members of the OPTIMA cohort, a group consisting of patients with dementia and elderly volunteers in whom physical and cognitive indices, including language, are assessed twice yearly until death. Samples of spontaneous speech from first entry into the cohort, and written texts in the form of letters, notebooks and diaries dating back over two to three decades, will be digitally encoded, and classified using a range of automated techniques.