Processing underpinnings of executive functions across adulthood: The interplay of cognitive control and inhibition

Abstract

Introduction Executive Functions (EF) are a set of “higher-level” cognitive functions that allow quick shifts of mindset, necessary to adapt our behaviour to a wide range of life situations. There are several EF, each often involving several subsidiary processes. This EF system supports planning, coordinating, sequencing, and monitoring of other cognitive operations. Deficits in EF are usually found in the elderly population, playing a key role as mediators of age-related changes in non-executive functions (nEF), such as memory and language. These deficits in EF are apparently selective rather than generalized: EF relying on the recruitment of automatic processes are less affected by ageing than processes relying on higher levels of cognitive control. In order to examine the nature of EF and the ageing impact upon these functions, we conducted a series of studies at two levels of analysis: the structural level and the processing level. At the structural level, we addressed the problem of the existence of a unitary EF, and the alternative to this view, that different EF do interact systematically, but without relying on a central EF “regulator” of some sort. At the processing level, we analysed the processing steps involved in the recruitment and implementation of cognitive control in a conflict task. Methods EF structure and age-related changes: We used a neuropsychological assessment battery comprising nine tests that theoretically assess EF, nEF pertaining to verbal abilities (VA), and processing speed (PS). We collected data from ninety young adults and used confirmatory factor analysis to examine the internal structure of the battery. We also studied the relation between EF, VA and PS. To study possible age-related changes in EF and nEF, we administered a neuropsychological battery to twenty young adults and twenty older adults. A review of the current literature on EF and ageing, with an emphasis on neuropsychological assessment, guided the selection of the nine tests. EF processing and age-related changes: We conducted four experiments using a spatial Stroop task, (i) three of them scrutinizing the nature of the control processes deployed to resolve response-conflict, and one (ii) contrasting the implementation of the conflict-resolving control setup in younger and older adults. The spatial Stroop task employed in all studies was designed after a review of the numerous paradigms and methods used in inhibitory control studies. In this review, we focused on studies that used the high temporal resolution of the event-related potential (ERP) technique to identify the time-course of different processes involved in the implementation of inhibitory control. We found that some paradigms, like the Stroop paradigm, are more appropriate to study control processes than others (e.g., the Go-Nogo). These cognitive control paradigms could be expected to mostly recruit controlled processes, but in fact the latter co-occur and interact with automatic processes. In order to analyse this interplay, we studied cognitive control in two different, yet closely related, manifestations in a spatial Stroop task: the interference effect and the congruency sequence effects. In our task, participants responded to the left/right direction of an arrow, while ignoring its left/right position in a computer screen. The arrow’s direction and position could be congruent (C) or incongruent (IC). (i) To study the processing underpinnings of cognitive control, we conducted a set of three experiments, designed to contrast two theoretical views pertaining to the nature of those processes, the Prediction of Response-Outcome (PRO) theory and the Conflict Monitoring Theory (CMT). In Experiment 1, we collected data from thirty seven young adults to analyse the effect of the trial n-1 congruency type on an nth C trial. In Experiment 2, we collected data from thirty two young adults to analyse the effect of the trial n-1 congruency type on an nth position-only trial (PO; participants must respond to the left/right position of a circle). In Experiment 3, we collected data from thirty six young adults to analyse the effect of the trial n-1 congruency type on an nth IC trial. (ii) To investigate age-related modulations of the interplay between the controlled and automatic processes involved in cognitive control processing, performance in the spatial Stroop task was contrasted in younger and older adults. The interference effect in IC trials as well as the effect of trial n-1 congruency type on a nth C trial were analysed. The young and older adults participants in this study were the same that participated in the ageing study, mentioned in the EF Structure and age-related changes section. Results EF structure and age-related changes: We found that a three-correlated-factor model (EF, VA and PS) was the most suitable for our data. EF and PS were related but separable functions, whereas the EF and VA factors were unrelated. Concerning the cognitive ageing study we found that older adults’ performance was inferior to young adults’ performance in only twelve of the twenty six neuropsychological measures analysed. These age-related deficits are mainly explained by cognitive slowing and/or by inhibition deficits. There were no age-related deficits in the other fourteen measures that included both EF and nEF. EF processing and age-related changes: Our results showed that PRO can better predict our participants’ performance than CMT. According to PRO, cognitive control is recruited in the presence of conflict, when multiple responses are available. The action plans yielding responses with an unacceptable cost (e.g., high error probability) are suppressed, leaving only the most appropriate action plan available. This enables the selection of the appropriate response, leading to conflict resolution. Concerning age-related modulations, our results revealed a generalized slowing that interacted with conflict resolution, yielding a larger interference effect for older adults. Despite this generalized slowing, the accuracy rates were similar in young and older adults, suggesting that with ageing, the implementation of cognitive control to resolve conflict becomes slower, but remains effective. Concerning the congruency sequence effects, older adults were slower than young adults in all conditions. However, we did not find any differences between the age groups in respect to the pattern found for the congruency sequence effects. Conclusions Taken together, our results support the diverse nature of EF. At the structural level, some EF tasks rely more on nEF and PS than others. Also, for an adequate performance in EF tasks designed to single-out one specific EF, multiple EF are in fact recruited. EF tasks less reliant on PS or on inhibition seem to be less affected by ageing. At the processing level, top-down processes are recruited in the presence of conflict for the computation and valuation of the multiple expected outcomes. Then, automatic processes, including both inhibition and enhancement processes, seem to play a key role in the cognitive control implementation. These automatic processes, as manifest in congruency sequence effects, are minimally affected by ageing. As for the controlled processes that are responsible for the implementation of control, it remains as an open question to what extent their age-related decline does not merely reflect a generalized reduction in PS.