Synonyms

Executive control; Self-control

Cognitive control and self-regulation are key determinants of goal-related behavior and known to be highly susceptible to increasing age. This entry provides an overview about new insights into the geropsychology of cognitive control and self-regulation. In two sections, each concept is briefly defined in combination with existing knowledge about their neuronal underpinnings. Then, empirical evidence on age differences in cognitive control and self-regulatory abilities as well as on how these can be improved by cognitive interventions will be summarized.

Definition of Cognitive Control

The term cognitive control refers to a set of higher-order processes that regulate basic sensory, motor, and cognitive operations for planning, guiding, and coordinating goal-directed behavior in everyday life (Miller and Wallis 2009). As these higher-order cognitive processes are assumed to allow pursuing internal goals and flexibly adapting to external changes, cognitive control is regarded as being critical to intelligent behavior.

Evidence from cognitive neuroscientific studies suggests that the mechanism of cognitive control can be attributed to a distributed brain network with the prefrontal cortex (PFC) taking a critical role. Due to its extensive connectivity, the PFC seems to be strongly activated whenever multimodal information from sensory and motor systems and subcortical structures needs to be integrated and maintained in a highly accessible state (Miller and Wallis 2009). Specifically, this PFC activation ensures the maintenance of goal-related information against distraction and serves the top-down guidance of neuronal activation in other brain areas required for the execution of controlled behavior. Advancing age is characterized by a marked neuronal deterioration particularly affecting the integrity of the PFC (Raz 2005). This neuronal alteration in the healthy aging brain has been linked to a pronounced decline of performance in tasks assessing cognitive control processes.

Age Differences in Components of Cognitive Control

Researchers suggested three key components of cognitive control that are assumed to be interrelated but separable abilities: (1) the updating and monitoring of working memory (WM) representations, (2) the inhibition of predominant response tendencies, and (3) the flexible switching between cognitive tasks (Buitenweg et al. 2012). At the theoretical level, aging researchers investigating cognitive control aimed at determining whether age differences in these abilities reflect process-specific limitations or whether they can be explained by age differences in one single general factor reflecting age-related changes in speed of cognitive processing (Kray and Eppinger 2010). At the applied level, aging researchers become more and more interested in assessing the extent to which these abilities can be improved by cognitive interventions (Kray and Ferdinand 2014). In the following, evidence on both levels will be reviewed for each of the three key components of cognitive control.

Working Memory. There is now ample empirical evidence that WM is a crucial determinant of age differences in cognitive control. In its traditional conceptualization, WM includes domain-specific buffers for short-term storage of visuospatial and verbal information along with central executive processes that monitor and manipulate the storage contents in the service of controlled, goal-directed operations (Hale et al. 2007). The large number of existing WM tasks distinguishes tasks that measure storage capacity, generally termed simple-span tasks (Reuter-Lorenz and Jonides 2007), from tasks that assess both storage capacity and central executive processing, generally termed complex-span tasks. For instance, digit-span tasks can be classified as simple-span tasks. They are assumed to assess the individuals’ ability to actively store a number of visually or auditorily presented items and to recall them in the correct order. In contrast, complex-span tasks assess not only storage processing but also the manipulation of the stored information or the processing of a secondary task (Reuter-Lorenz and Jonides 2007). For instance, reading-span tasks require the processing of a sequence of sentences and deciding for each sentence whether it is meaningful or not. At the same time, individuals have to encode the final word of each sentence and to remember these until the end of the task (Reuter-Lorenz and Jonides 2007).

Researchers investigating age differences in simple-span tasks found a differential decline in the performance on visuospatial and verbal WM tasks in older relative to younger adults, indicating larger deficits in visuospatial relative to verbal WM tasks (Hale et al. 2007). Other studies show that age differences are even more pronounced in complex-span tasks, tapping both storage and executive processes, than age differences in storage measures per se. While some researchers tried to separately investigate storage and executive processes of WM in order to identify process-specific limitations in old age, others argued that this separation is artificial, as even storage tasks might require executive control processes, particularly in older adults (Reuter-Lorenz and Jonides 2007). Evidence for this view comes from studies using functional magnetic brain imaging (fMRI) data. Older adults recruit regions of the PFC associated with cognitive control even in simple storage tasks, while these brain regions are not significantly activated in younger adults. These findings suggest that the recruitment of the PFC serves as a compensatory mechanism to maintain good performance in older adults, as assumed by the Compensation-Related Utilization of Neural Circuits Hypothesis (CRUNCH) (Reuter-Lorenz and Jonides 2007). However, as task complexity increases, and thereby the demands on executive control, older adults perform poor on WM tasks because their control processes are already taken up by lower storage demands (Reuter-Lorenz and Jonides 2007).

It is now well known that age differences in WM tasks are particularly pronounced with increased cognitive control demands. A critical question from an applied perspective is whether WM can be improved by cognitive interventions even in elderly individuals. In the last decade, a variety of training studies aimed at improving WM performance by means of computerized training programs (Klingberg 2010). These programs often used adaptive training schedules to optimally adjust training demands to individuals’ performance levels. Results of these studies indicate that a variable amount of training on visuospatial and verbal WM tasks results in considerable training gains not only in the trained task but also in closely related but untrained WM tasks (Buitenweg et al. 2012; Klingberg 2010). Less consistent evidence is reported on the extent to which these training gains can be maintained over a longer period of time and can be generalized to other cognitive tasks (Buitenweg et al. 2012; Klingberg 2010). These inconsistencies might be due to differences in the duration of the training interventions as well as in the type of training. For instance, it has been shown that practice of explicit memory strategies leads to WM improvements in older adults, but these strategies are not easily transferred to other memory tasks. In contrast, adaptive, process-oriented WM training sometimes also leads to performance gains in other cognitive tasks (Buitenweg et al. 2012).

Inhibition. Pronounced age differences have also been demonstrated in cognitive control tasks requiring the ability to inhibit irrelevant responses and predominant actions (Buitenweg et al. 2012). One frequently used task to measure the efficiency of inhibition processes is the Stroop task (Kray and Eppinger 2010). This task consists of color words either printed in a compatible color (i.e., “red” printed in red ink) or in an incompatible color (i.e., “red” printed in blue ink) (Kray and Eppinger 2010). Subjects are usually instructed to perform the less-practiced color naming and to inhibit the more-automatized reading of the word meaning. Results typically show longer reaction times and larger error rates in cases in which the word reading interferes with the color naming, that is, on incompatible stimuli relative to compatible ones. This so-called Stroop interference effect is usually larger in elderly adults than in younger adults. Results from a meta-analysis suggest that the larger Stroop interference effect in older adults can be fully explained by age differences in general speed of processing as a general underlying factor (Kray and Eppinger 2010) and not by specific limitation in inhibitory processing. However, when the demands on controlled processing are increased, for instance, by manipulating the frequency of trials on which subjects have to inhibit automatic responses, older adults tend to show larger deficits in inhibitory control than younger adults. The greater need to flexibly recruit cognitive control on less frequent conflict trials, inducing higher demands on cognitive control, lead to pronounced impairments in inhibition tasks in the elderly, similar to the reported findings on age differences in WM tasks (Kray and Eppinger 2010).

Despite the existing age-related decline in measures of inhibition, studies that aim to enhance inhibitory control in old age are lacking (Buitenweg et al. 2012), although there is some evidence for practice-related improvement in the Stroop tasks in older adults. Whether these training gains also transfer to other inhibitory control tasks or even to improvements in other measures of cognitive control remains to be examined (Buitenweg et al. 2012). It has also been suggested that older adults show better inhibitory control in the morning than in the evening, based on circadian arousal patterns for inhibitory processes that predominantly peak in the morning in the elderly (Hasher et al. 2007). In addition, there are some first results indicating that physical fitness is linked to better inhibitory control that is explained by increased prefrontal oxygenation. In this regard, it has been shown that 8 weeks of moderate aerobic training can improve performance on an inhibition task in older adults (Berryman et al. 2014).

Task Switching. Age differences in cognitive control are also obtainable in tasks assessing the flexible switching between task rule representations. In these types of tasks, subjects are required to alternate between two or more simple categorization tasks such as deciding whether a stimulus belongs to the category of fruits or vegetables (picture task) or whether it is gray or colored (color task). Results of a meta-analysis on age differences in task switching show larger general (or global) switching costs in older than younger adults when two tasks are performed in an alternating order on a task-switching block relative to performance on a single task block (also termed mixing costs) (Kray and Ferdinand 2014). These age-related deficits seem to map on age differences in WM as a key determinant of cognitive control, as age differences in the implementation of a task switch within a task-switching block, termed specific (or local) switching costs, are less pronounced than age differences in general switching costs (Kray and Ferdinand 2014). Importantly, age effects in these costs remain reliable after controlling for age differences in processing speed, suggesting process-specific limitations in the ability to maintain and select task sets.

Further evidence for this view comes from aging studies measuring the neuronal activity during task preparation and response selection by means of fMRI and event-related potential (ERP) data. For instance, if the upcoming task in a task-switching paradigm is announced by a preceding task cue, changes in neuronal activity suggest that older relative to younger adults tend to update the appropriate task representations in WM after task-cue presentation all the time, even if not required, i.e., when the response rules are exactly repeated compared to the previous task. Moreover, older adults seem to recruit a larger proportion of the PFC than younger adults even in single task blocks in which no task switching is required. Similar to age differences in WM, this finding may reflect that older adults tend to compensate for difficulties in maintaining task rule representations by activating a larger network of prefrontal brain areas (Kray and Ferdinand 2014). Together, the results of task-switching studies favor process-specific limitations, as behavioral and psychophysiological measures suggest age differences in the representation and selection of task goals in WM that cannot be attributed to processing speed as a single underlying factor. In contrast, the switching process itself seems to be relatively preserved in old age (Kray and Ferdinand 2014).

Despite these age-related limitations, recent intervention studies revealed substantial plasticity in task-switching abilities among older adults (Buitenweg et al. 2012; Kray and Ferdinand 2014). Strategy-based interventions, for instance, employed labeling strategies such as verbalizing the next task to promote the planning and preparation of the upcoming task switch and thus to facilitate goal-directed behavior. Results on these kinds of interventions show a substantial benefit of verbal self-instructions on switching costs particularly in older adults, indicating language processes to offer a promising approach to support action control in old age (Kray and Ferdinand 2014). Process-based interventions aim to enhance cognitive control by the practice of the underlying cognitive control processes involved in task switching. Recent studies reported a considerable reduction in switching costs for younger as well as older adults, indicating substantial potentials to improve switching ability (Buitenweg et al. 2012). Both age groups also showed larger performance gains in an untrained but structurally similar switching task, and these gains were even more pronounced in the older than in the younger age group. Importantly, training gains also generalized to untrained cognitive control task (e.g., inhibition, WM (Kray and Ferdinand 2014)), suggesting the training of cognitive control processes, and in particular, the training of maintaining and selecting (biasing) of tasks as required in dual-task-like situations is a promising approach to induce broader transfer to other cognitive domains (Buitenweg et al. 2012; Kray and Ferdinand 2014).

A Theoretical Framework for Explaining Age Differences in Cognitive Control

In sum, age-related differences in cognitive control have been shown in WM, inhibition, and switching tasks. Recently, the dual mechanisms of control (DMC; Braver et al. 2007) theory proposes that age differences in all of these tasks can be explained by age differences in one common mechanism, namely, the ability to process context information. Context information is described as the internal representation of task-relevant information such as rules, goals, or instructions within WM that is maintained and updated to serve controlled, goal-related behavior. Within this framework, context processing relies on the interaction between the dorsolateral PFC (DL-PFC) and the midbrain dopamine (DA) system. More precisely, sustained neuronal activity of the DL-PFC provides the online maintenance of context information in order to bias the activity in posterior and subcortical brain regions responsible for goal-related behavior in accordance with the current context representation. At the same time, phasic DA projections toward the DL-PFC in response to new, salient, or reward-predicting context cues are proposed to act as a gating mechanism, i.e., ensuring the appropriate updating of context information in the DL-PFC (Braver et al. 2007). Hence, sustained activity within the DL-PFC ensures the stability of goal-directed behavior against distraction, whereas the DA-guided gating mechanism simultaneously allows for the flexible adaptation to changing task demands. Given the well-known age-related neurobiological changes observed in the PFC and the midbrain DA system, deficits are expected in both the active maintenance and the gating of new context information that in turn impairs performance on a variety of cognitive control tasks. For instance, the active maintenance of task-relevant context information serves to protect information against interference, and disturbances therein particularly affect WM capacity. In a similar vein, deficits in actively maintaining context representations may also impair the ability to inhibit predominant response tendencies, as the maintenance of a contemporary task rule is thought to enable the activation of a weaker, task-relevant against a stronger but task-irrelevant response. Finally, phasic DA responses to the DL-PFC indicating the need for updating context representations are particularly important in task switching. Therein, context information represents the currently relevant task rule, and deficits in the gating mechanism might impair the updating and flexible attention shifting between cognitive tasks. These examples outline that instead of separating age differences in cognitive control into a decline of subprocesses such as WM, inhibition, and attention shifting, the DMC theory considers age differences in the neurobiological basis of context processing to be fundamental to account for age deficits in subcomponents of cognitive control (Braver et al. 2007).

Recent behavioral and neuroscientific studies on testing the assumptions of the DMC theory show that changes in the interplay between the PFC and the DA system inherent to healthy aging predominantly affect the time course of updating context information. Younger adults exhibit an early, proactive manner of context updating by the time context information is presented and hence update context information to prepare for an upcoming task in advance. In contrast, older adults show a late, reactive manner of context updating, only when needed such as when interference is detected in a reactive fashion (Braver 2012). While the temporal shift of context updating in a pro- versus reactive manner with increasing age has been supported on the basis of fMRI and ERP data (Braver 2012; Schmitt et al. 2014), age differences in maintaining context information have revealed mixed results and seem to occur only under specific task conditions. However, the DMC theory is promising as temporal differences in context updating seem to account for age differences in inhibition, WM, and task shifting that are often regarded as separable components of cognitive control (Braver et al. 2007).

Age-related differences in context updating have been shown to be susceptible to different training regimes. In two training studies, the AX continuous performance task (AX-CPT, Braver 2012) was applied in order to measure an individual’s ability to process context cue information required for correct responding to a subsequent probe stimulus. It has been demonstrated that both extended practice and directed strategy training toward the use of cue-based, proactive control in the AX-CPT reduced context processing deficits in older adults, indicating process- and strategy-based interventions to benefit context updating in old age. Moreover, the behavioral improvements in the AX-CPT in older adults were accompanied by increased PFC activation to the presentation of contextual information (Braver 2012). These results correspond to a recent training study showing training-related alterations in PFC activity to underlie the transfer of training gains to untrained cognitive control tasks (Bamidis et al. 2014). In this study, older adults performed an adaptive multitasking video game training offering a large stimulus variability and continuous feedback. The multitasking approach in particular encompassed the need for resolving task interference in the dual-task situation. Training gains were larger after multitask training than after training both tasks in isolation, transferred to other untrained cognitive control tasks such as WM and attention, and remained stable at a follow-up measurement 6 months after the training. Moreover, robust correlations between multitasking ability and changes in activation patterns of the PFC predicted the transfer gains to the untrained cognitive control tasks. Hence, process-based training interventions, such as multitasking training, that aim at improving cognitive control can result in alterations of the neuronal recruitment of the PFC in elderly individuals that may also generalize to other cognitive tasks relying on cognitive control networks (Bamidis et al. 2014).

Definition of Self-Regulation

The concept of self-regulation refers to the individual control of own actions, thoughts, and emotions toward the achievement of desired outcomes and intentions (Bauer and Baumeister 2011). It is very loosely defined and considered as a conglomeration of abilities, consisting among others the capability to override automatic habits, basic affects, and impulses, to control and monitor performance, to achieve distal aims, and to resist short-term temptations to the benefit of long-time goals. Accordingly, failures of self-regulatory ability affect both flexible behavior and social adaptation that can be observed in a broad range of psychological phenomena such as gambling, addiction, eating disorders, underachievement, prejudice, aggression, and so on (Bauer and Baumeister 2011).

In general, self-regulatory processing is considered as a system of feedback loops in which individuals concurrently monitor the discrepancy between the actual behavioral outcomes and feedback and the individuals’ goal and intentions (Bauer and Baumeister 2011). Whenever there is a discrepancy, individuals automatically or consciously engage in self-regulatory abilities to minimize the discrepancy until the goal is achieved. Hence, similar to the concept of cognitive control, self-regulation is highly important to adaptive, goal-directed behavior (Bauer and Baumeister 2011). For instance, it has been shown that individuals with better cognitive control ability, such as higher WM capacity, also tend to show better self-regulatory skills, such as less mind-wandering or more resistance toward the temptation of eating candy (Hofmann et al. 2011).

Neuronal Underpinnings

Recent research has also identified subprocesses of cognitive control to play an important role in the mechanisms of self-regulation (Wagner and Heatherton 2011), in particular the self-control aspects of self-regulation. This view is supported by evidence that akin to the mechanisms of cognitive control, a broader range of self-regulatory abilities depend on a network of specialized prefrontal brain regions, including the lateral PFC (L-PFC), the ventromedial PFC (VM-PFC), and the anterior cingulate cortex (ACC) (Wagner and Heatherton 2011). The L-PFC is highly related to other prefrontal regions, especially to motor cortices, the VM-PFC, and the ACC, and is assumed to contribute to the mere self-control processes involved in self-regulation, such as inhibiting inappropriate behaviors, maintaining multiple goals in WM and flexibly selecting between them, dealing with distraction, and carefully planning the sequence of goal-directed actions (Wagner and Heatherton 2011). In contrast, the VM-PFC is highly connected to subcortical structures involved in affective processing (e.g., the amygdala, the hypothalamus, the insula, and the ventral striatum). Therefore, the VM-PFC is seen to be particularly important for regulating affective and appetitive processes and adapting to social norms (Wagner and Heatherton 2011). This assumption has been supported by case reports showing patients with damage to the VM-PFC to exhibit drastic personality changes such as aggressive, socially inhibited behavior and a particular inability to respect social norms (Wagner and Heatherton 2011). Despite their functional differences, both the L-PFC and the VM-PFC are interconnected with the ACC that shares many connections with subcortical (e.g., the ventral striatum) and motor regions. Patient studies show that due to its close connection to motor cortices and subcortical structures involved in reward processing, damage to the ACC may result in general apathy, loss of motivation or interest, and an inability to generate behavior (Wagner and Heatherton 2011). Moreover, given its anatomically strategic position, neuroscientific research regards the ACC as a neuronal correlate of a conflict detection mechanism, signaling the need for increased control toward the L-PFC whenever performance errors are detected. This role closely reflects the conceptualization of self-regulation as a system of feedback loops (Wagner and Heatherton 2011).

The strong anatomical and functional overlap between control processes (e.g., inhibition of temptations and automatic behaviors) attributable to both the concepts of cognitive control and self-regulation has led to systematic investigations of their interaction. In the Strength Model of Control and Depletion (Bauer and Baumeister 2011), it has been argued that these control processes depend on a limited, domain-general physiological resource that – once depleted – results in impaired performance on task relying on this resource. Dieters, for example, whose resource for self-regulatory control on eating behavior was stressed by inhibiting temptation from nearby food, showed impaired performance not only in a subsequent task on self-regulation (i.e., eating ice cream) but also on a cognitive control task (i.e., WM) relative to non-dieters and dieters whose self-regulation was not additionally depleted by tempting foods (Bauer and Baumeister 2011). Likewise, participants who were required to take part in a difficult cognitive control task (i.e., attentional control) showed impaired self-regulatory control (i.e., emotion regulation) compared to control participants who did not complete the cognitive control task (Wagner and Heatherton 2011).

Age Differences in Self-Regulation

Age-related deficits in self-regulation have been strongly associated with impairments in cognitive control and in particular with impairments in inhibitory control (Von Hippel and Henry 2011). On the one hand, due to a failure to inhibit and control automatically activated thoughts and temptations, older adults seem to express more social stereotypes (i.e., race-related prejudices), exhibit more socially inappropriate behavior (i.e., talking about private issues in public and generating gratuitous comments), and engage more in risky gambling (i.e., larger perseverance in the absence of reward) than younger adults (Von Hippel and Henry 2011). Interestingly, individual differences in inhibitory deficits (as measured with standard cognitive control tasks) seem to directly mediate the extent of self-regulatory failure in older adults (Von Hippel and Henry 2011). Moreover, these inhibitory deficits have been shown to be sensitive to circadian rhythms, with smaller deficits obtained in the morning than afternoon based on biological changes in the underlying neuronal resource. Therefore, older adults showed more risky gambling and socially inappropriate behavior when they were tested in the afternoon relative to when the experiment took place in the morning (Von Hippel and Henry 2011). On the other hand, evidence exists that older adults are able to manage self-regulatory deficits and inhibit expressing stereotypes or inappropriate behaviors when they are aware of it. For instance, if older adults are forewarned about an upcoming, irrelevant stereotypic situation or if they know beforehand that they have to suppress a socially inappropriate action later on (Von Hippel and Henry 2011), they do not differ in the appropriateness of their behavior relative to younger adults. This suggests that older adults may prepare for potential inhibitory control deficits, and hence, can exert conscious control over their self-regulatory abilities.

Furthermore, there are also findings suggesting increased self-regulatory skills in older adults than in younger adults (Von Hippel and Henry 2011). For instance, in the domain of emotion regulation, it has been shown that older adults focus more on positive than negative or neutral information in order to voluntarily enhance their emotional well-being. This phenomenon, known as the age-related positivity effect (Mather 2006), has been explained in the framework of the socio-emotional selectivity theory (Mather 2006). This theory posits that personal goals have to be regarded within future time constraints. In the case that individuals value future time horizons as enduring, they will focus on goals related to the future, such as gaining knowledge. In contrast, if individuals recognize future time as restricted, just as it occurs in older age, they will focus on immediate, meaningful goals, such as emotional regulation and gratification (Mather 2006). Accordingly, relative to younger adults, older adults expressed higher emotional stability and skills of emotion regulation, showed more effective social problem-solving, focused more strongly on positive relationships with others, and reported less self-conscious negative emotions (Mather 2006). Similar to the controlled compensation of inhibitory deficits, the age-related positivity effect seems to be more pronounced when older adults are forewarned and for older adults with better cognitive control ability. In contrast, when cognitive load is increased or cognitive control abilities are impaired as in pathological aging, older adults are less able to invest in controlled processing of emotional information and the positivity effect vanishes (Mather 2006).

Apart from the stable or even improved ability of emotion regulation in the elderly, there are only a few studies that have investigated whether self-regulatory skills can be improved by cognitive interventions. These studies show that already a limited amount of practice in self-regulatory control, for instance, by controlling eaten food or engaging in regular physical exercises, is able to translate into improvements in key aspects of self-control in laboratory tasks and also transfers to self-regulatory skills in everyday life (e.g., decreased consumption of cigarettes, alcohol, and unhealthy food) (Bauer and Baumeister 2011). So far these studies have mainly been conducted in younger or middle-aged individuals and largely neglected the effect of self-regulatory practice in older adults (Hofman et al. 2012). However, given existing evidence for self-regulatory failures in old age (e.g., problematic gambling), it might be especially important to create successful self-control trainings and to investigate any potential transfer to measures of self-regulation in this age group. In this respect, it is also interesting to note that improvements in self-control have been demonstrated via training of cognitive control functions. In a study on middle-aged problem drinkers, WM training transferred to self-regulatory improvements, i.e., reduced alcohol consumption for more than 1 month after training (Hofman et al. 2012). Given the functional relationship between cognitive control and self-regulation, these improvements may draw back on the underlying neuronal process resource common to cognitive control and self-regulation. Thus, as cognitive interventions have already demonstrated significant potential to improve different facets of cognitive control in old age (see previous section on cognitive control), future research studies in older adults might investigate whether and to what extent these interventions are able to transfer to self-regulatory skills. Finally, these studies might also turn to training-related changes in the underlying brain network of self-regulation and cognitive control.

Summary

There is now accumulated evidence for age-related limitations in cognitive and self-regulatory control that are associated with alterations in different underlying neuronal networks. Age differences in cognitive control occur in tasks requiring high demands on WM, inhibitory control, and task switching that can be explained by a recent neurobiological theory on context processing. Age differences in self-regulation primarily concern self-control, while emotion regulation is relatively preserved. Intervention studies have revealed considerable plasticity in cognitive control in elderly individuals, while the potential benefit of training in self-regulation is not known yet. Given that cognitive control and self-regulation partly rely on similar neuronal networks, it will be an important challenge for future research to determine whether training in either of these abilities will lead to improvements in the corresponding other one.

Cross-References