Longitudinal Changes in the Language Function in Elderly People with Mild Cognitive Impairment

Mild cognitive impairment (MCI) describes a decline in memory and other cognitive domains, including language. It occurs with age and is often viewed as a transitional stage between healthy ageing and dementia, however, the conversion to dementia does not necessarily happen. Lack of longitudinal studies of MCI in the language function and the promising prospects offered by them motivated us to employ such design. We investigated longitudinal changes in the language function in elderly people and the demographic and cognitive factors that may predict an improvement or decrease in the language function over time. All patients (n = 97, 58 to 89 years old) were native Russian speakers from the Memory Clinic in Moscow. They were asked to undertake three screening assessments (MoCA, MMSE, and HADS), to fill in a questionnaire about subjective language and memory complaints, and to perform two language tasks (naming-by-definition and language comprehension). The language tasks were repeated one year later, to test the longitudinal changes in the language function. We applied Pearson's correlation tests and the linear mixed-effects models to the data. As measurement of language performance we chose naming accuracy, response time for the naming task, sentence comprehension accuracy and mean reading time for the sentence comprehension task. We also had to eliminate subjective complaints as a potential predictor due to the lack of data. For the analysis we calculated the percentage of change of language performance measures. The only statistically significant correlation was between the percentage of change for naming accuracy and MoCA (Est. = 0.04, SE = 0.02 , z = 2.36, p = 0.02). The results of this study can be explained by several limitations it has. To assess language performance we used non-standardized tasks. Moreover, potentially the most important predictor – subjective cognitive complaints – was excluded from the analysis. We also assume that it is not reasonable to draw any conclusions based on the only statistically significant correlation. These limitations should be taken into account in the future studies.