Determinants and correlates of intra-individual variability in reaction time

Abstract

Traditionally, reaction time (RT) was conceived of as an average speed of a number of responses made by an individual, or mean RT. Increasingly, however, intraindividual variability in reaction time (RT IIV) – the consistency of responses by a single person across trials – is used as an additional or even alternative measure. RT IIV is often found to be elevated in a number of conditions that affect the central nervous system functioning, such as traumatic brain injury or neurodegenerative diseases. It can predict change in cognitive performance in ageing, progression from normal ageing to mild cognitive impairment, and even death. Therefore, RT IIV may be of great practical importance. However, RT IIV and mean RT are correlated; therefore it is often problematic to draw conclusions about unique associations between these and other variables. One objective of the work presented in this thesis was to investigate determinants and correlates of simple and choice RT IIV and to test which associations may be accounted for by the individual differences in mean RT. The first investigation was concerned with age differences in RT IIV. Following a systematic review of literature, a series of meta-analyses demonstrated that older individuals (aged 60 years and above) have greater RT IIV than young or middle-aged adults in simple and choice RT tasks. The effects were reduced but still significant when RT IIV was adjusted for mean RT. The next study was a cross-sectional investigation of the associations between age and RT IIV, as well as of sex differences in RT IIV, across the lifespan in participants ranging in age from 4 to 75. Non-linear effects of age were found for RT IIV measures, such that variability decreased with age in children and increased with age in older adults. A novel finding from this study was that sex differences in RT IIV were present among adults but not children, suggesting that there might be an age threshold at which sexes diverge in their RT IIV trajectories. The results also indicated that findings regarding RT IIV may differ depending on the variability measure used (that is, whether and how mean RT is controlled). The second study on the same sample investigated variability on a trial-by-trial basis. Specifically, it tested the hypothesis that sex differences in variability are due to females being disproportionately slower at the first trial which inflates their overall RT IIV. This hypothesis was not supported. Another investigation used longitudinal data from the West of Scotland Twenty-07 study. Three cohorts of individuals aged approximately 15, 35 and 55, were followed up for 20 years and had RT data collected at four occasions. Analyses confirmed non-linear effects of age on RT IIV found in the earlier cross-sectional investigation. The final study investigated the effect of high altitude on RT IIV. It found that altitude-related increase in RT IIV is fully accounted for by general slowing of RT at high altitude. The overall pattern of results obtained from the investigations suggests that RT IIV increases with age in adults and that not all of the increase is due to general slowing. Moreover, the results show that sex differences in RT IIV are not uniform across the lifespan. Finally, whereas associations of RT IIV with some variables, for example age, are relatively robust to controlling for mean RT, others are fully attenuated by such practice.