Compensatory Brain Activity in Older Adults.

Compensatory brain activity in older adults is an important research area concerned with cognitive aging and the possibilities of cognitive enhancement in late adulthood. Most comprehensively these issues are addressed by the STAC model - Scaffolding Theory of Aging and Cognition. The present study will be concerned with empirical investigations into some of its theoretical assumptions. The innovative character of the research solutions included in the present project is based, on the one hand, on new approaches to the verification of the STAC assumptions and, on the other hand, on the verification of the assumptions that have not been subjected to systematic investigations so far. Such a research agenda has been motivated by gaps in existing knowledge regarding some of the phenomena described in the STAC model. Filling these gaps will be the contribution of the project to the development of research on cognitive aging.

In this context, the aims of the study are the following:

1. The verification of the STAC model assumptions about the cognitive processes triggered during compensatory brain activity in older adults based on EEG time-frequency analyses.
2. The verification of the STAC model assumption about the possible influence of cognitive training on the emergence and disappearance of compensatory brain activity in older adults, has not been investigated so far.

Research problems justifying the aims:

Re. 1. The cognitive interpretation of compensatory brain activity in older adults, both in the STAC model and in light of research findings collected until now, is based on incomplete premises. Compensatory brain engagement is the overactivation observed in certain brain regions in older adults which is linked to a better performance on cognitive tasks. Better performance on cognitive tasks is regarded here as evidence for ongoing compensation. Compensatory brain activity in older adults has been confirmed by numerous studies utilizing fMRI and PET scans, which are the two most dominant neuroimaging methods in this area. Models of compensatory brain activity in older adults, such as the STAC model, are based on these neuroimaging data. Theoretical interpretations in terms of cognitive processes are imposed on these observations only subsequently. Thus, the verification of the compensatory mechanisms described in these models is limited as the development and the verification of the model are based on the same methodology. Subsequent studies that found a positive relationship between the overactivation of specific brain regions and better performance on cognitive tasks confirmed the empirical relationship that led to the development of the theoretical model. However, the confirmation of the relationship alone does not suffice to verify the theoretical interpretation encapsulated in the model. The need to deepen the understanding of cognitive processes engaged in compensation was indicated by the very authors of the STAC model.

The use of the EEG methods in the project (event-related oscillations, ERO), instead of the fMRI or PET imaging, will contribute to this goal. The reason is that the EEG-based approach requires hypotheses based not only on the localization of compensatory brain activity but also on the cognitive processes theoretically linked to the activity of interest. Since fMRI and PET findings cannot be directly expressed in EEG terms, the following reasoning has been employed to predict the results: compensatory brain overactivation observed in fMRI and PET → cognitive interpretation of the overactivation → determination of the EEG indicators (event-related synchronization, ERS and event-related desynchronization, ERD) of the cognitive processes specified based on the interpretation → the verification of the existence of the indicators in EEG signal recorded in older adults when their cognitive performance does not diverge from performance observed in young adults (and their lack when the older adults perform worse than young subjects). In other words, the aim is to identify the neural markers of the processes that have been proposed by the STAC model to account for compensatory brain activity based on the activation of certain brain structures as registered with the fMRI and PET but with the use of another technique that relies on alternative indicators of cognitive processes' engagement. Thus, thanks to the use of EEG measurements in the study, it will be possible to verify the compensatory role of the cognitive processes and not only confirm the localization of the overactivation, which has been a significant limitation of the studies conducted so far.

Re. 2. It is unknown if it is possible to influence compensatory brain activity in older adults through cognitive training, which is assumed in the STAC model. The STAC model assumes that compensatory brain activity can be stimulated by various kinds of interventions, including cognitive training. The authors of the model proposed that such training in the case of an under-activation or a deterioration of a neural network engaged in the cognitive task should establish new scaffolds, whereas when the performance on the tasks already relies on compensatory overactivation of certain brain areas, the training should decrease the activation in these secondary "scaffolded" regions and improve efficiency in the primary networks - a kind of a restoration of the brain activity characteristic to younger individuals.

Assumptions about the aims of the training indicated by the authors of the STAC model have not been verified empirically so far. Research findings on the training-induced changes in neural activity have been only interpreted post facto in the context of the STAC model or narrower models of compensatory brain activity in older adults: CRUNCH - Compensation-Related Utilization of Neural Circuits Hypothesis, PASA - Posterior-Anterior Shift in Aging or HAROLD - Hemispheric Asymmetry Reduction in Older Adults.

To our best knowledge, only two publications have reported so far studies that were designed directly to verify the assumptions of one of the models of compensatory brain activity in older adults - CRUNCH. The study, however, was conducted with the use of fMRI imaging, not EEG, and without including young adults, who constitute a reference group fundamental to the assessment of the specificity of brain activity characteristic to old age.

Although some studies on the efficiency of cognitive training have been interested in the neuro-correlates of training, in most of them the sample consisted solely of young individuals, and even if older subjects were included the research agenda did not address hypotheses regarding compensatory scaffolding in old age, or they were to interpret in such context only partially.

In the present study, the assumptions of the STAC model regarding the effects of cognitive training on compensatory scaffolding in older adults will be verified experimentally with the use of working memory (WM) training. The task employed in the training will be based on the n-back paradigm, which is the most common type of training employed in contemporary research. Although the transfer effect of n-back training is questionable, there is abundant evidence for its effectiveness in improving performance on the trained task.

The project was conducted in two phases: preliminary and experimental. The preliminary phase has three main goals: (1) Empirical determination of three difficulty levels of the n-back task for the subsequent use during EEG recordings - the lowest level of difficulty should be easy for old adults (with no difference in accuracy compared to young adults) and two other levels should be increasingly (significantly) more difficult; (2) Selection of the task for the active control group. This task should not involve working memory but should seem to improve the same aspect of cognitive functioning as the n-back task, which will be verified by experts as well as old and young lay persons not belonging to the study sample; (3) Recruitment of volunteers for the study sample and the assignment of the selected participants to the experimental and control groups ensuring equivalence of these groups in terms of cognitive performance (according to the cognitive assessment in the preliminary phase). Moreover, the preliminary phase is devoted also to the programming of the computerized tasks that will be used in the study. Computerized tasks for cognitive assessment and training will be developed in the PsychoPy programming environment with the use of the Python programming language.

The sample comprised 3 groups of older adults: experimental (n = 25), passive control (n = 22), active control (n = 7), and young adults: experimental (n = 25), passive control (n = 25), and active control (n = 12). Pre-test and post-test measurements were conducted in all groups.

At pre-test and post-test participants in all groups performed a computerized n-back task at three levels of difficulty (determined beforehand in the preliminary phase) with simultaneous EEG signal recording. Moreover, in all groups, cognitive assessment was conducted, which will concern: working memory and short-term memory capacity, attentional focus and psychomotor speed, interference resistance, and inhibition ability. Only the n-back task will be performed with EEG registration.

The experimental factor was the intervention using the n-back training, namely older and younger participants assigned to the experimental groups participated in a computerized process-based working memory training with the use of the n-back task. The training intervention lasted 4 weeks and consisted of 12 training sessions, about 45 minutes each. The n-back tasks used for the training purposes were analogous to those used at pre-test and post-test measurement (the same interface, instructions, and actions necessary to complete the task). The difference was the difficulty level because, in the case of the training, it was not planned but raised adaptively, according to the current progress of the particular participant. The other two groups (of old and young adults) will be active control groups. In these groups, cognitive training intervention with the use of n-back task was replaced by multi-domain non-adaptive training. The training schedule in active control groups will be the same as in the experimental groups. There were also two passive control groups (of old and young adults), which during the training period were not contacted by researchers, and no cognitive interventions were administered.