A theory of cognitive control, aging cognition, and neuromodulation

Abstract

A theory is described which links cognitive changes observed in normal aging to an underlying decline in the function of the dopamine (DA) system projection to prefrontal cortex (PFC). The theory postulates that this neural mechanism is integral to the representation, maintenance and updating of context information, and as such impacts cognitive control across a wide range of cognitive domains, including working memory, attention, and inhibition. Behavioral and brain imaging data in support of the theory are discussed, which demonstrate selective impairments in context processing among healthy older adults associated with abnormal PFC activation. These findings highlight the utility of a computational approach to cognitive aging. Current directions for further refinement and validation of the model are outlined.

Introduction

An important goal in the study of cognitive aging is to discover potentially unifying underlying mechanisms that contribute to the diversity of cognitive changes occurring with advancing age. However, this endeavor is a difficult one because it is not usually clear how to link the observable behavioral phenomena typically found in older adults (e.g. slowing response time, poorer performance on laboratory tasks) with latent psychological and/or neurobiological changes. Computational models provide an important tool that can be used to demonstrate how particular changes in underlying cognitive and/or neurobiological mechanisms can lead to a wide range of observable behavior changes. As such, the application of computational models to the study of cognitive aging has the potential to advance our understanding of the core mechanisms that lead to cognitive changes in healthy aging [24]. In our work, we have focused on the mechanisms of cognitive control, how these mechanisms influence behavior across a wide range of cognitive domains, and how they are supported by specific neural systems (i.e. prefrontal cortex (PFC) and the dopamine system). Recently, we have applied this approach towards understanding age related changes in cognitive control, and how such changes might be linked to age-related changes in PFC and dopamine function. Here we present a brief overview of our theoretical model of cognitive control, how the model accounts for the nature of aging cognition, and the ability of this model to capture specific patterns of behavioral performance among older adults on particular cognitive tasks.

As discussed in the other articles in this issue, a large literature on cognitive function in healthy aging suggests that older adults display deficits in multiple different cognitive domains, including episodic memory, working memory, prospective memory, inhibition, attention, and ‘executive’ function [3], [16], [21], [28], [35], [43]. In our own work, we have suggested that there is a common element to all of these cognitive domains. Namely, that they place a heavy load on cognitive control. More specifically, we propose that successful performance in a wide variety of cognitive situations centrally depends upon the internal representation, maintenance, and updating of context information in the service of exerting control over thoughts and behavior [10], [12], [15]. We define context as any task-relevant information that is internally represented in such a form that it can bias processing in the pathways responsible for task performance. Goal representations are one form of such information, which have their influence on planning and overt behavior. However, we use the more general term context to include representations that may have their effect earlier in the processing stream, on interpretive or attentional processes. For example, in the Stroop task (which involves naming the ink color of color-named words, e.g. the word RED in green ink), the context provided by the task instructions must be actively represented and maintained to bias attentional allocation and response selection towards the ink color dimension rather than the word name. Thus, context representations may include a specific prior stimulus, or the result of processing a sequence of stimuli, as well as task instructions or a particular intended action. Representations of context are particularly important for situations in which there is strong competition for response selection. These situations may arise when the appropriate response is one that is relatively infrequent, or when the inappropriate response is dominant and must be inhibited (such as the word name in the Stroop task). Because context representations are maintained on-line, in an active state, they are continually accessible and available to influence processing. Consequently, context can be thought of as one component of working memory. Specifically, context can be viewed as the subset of representations within working memory which govern how other representations are used. In this manner, context representations simultaneously subserve both mnemonic and control functions. This aspect of the model differentiates it from standard models of working memory [2], which postulate a strict separation of representations for storage versus control.

We further suggest that context processing and cognitive control are subserved by a specific set of underlying neural mechanisms. In particular, we postulate that representations of context information are housed within the dorsolateral portion of the prefrontal cortex (DL-PFC) and actively maintained there when task demands require such active maintenance [30]. The dopamine (DA) projections to DL-PFC are postulated to regulate the access to such context information, insulating this information from the interfering effects of noise over intervals in which the information must be sustained, while at the same time allowing for the appropriate updating of such context information when needed [10]. These assertions are consistent with the neuroscience literature, in which active maintenance in the service of control is a commonly ascribed function to PFC [18], [19], [27], and the DA system is widely held to modulate the active maintenance properties of PFC [25], [37], [44]. In our model, the context processing functions of cognitive control critically depend upon DL-PFC and DA system interactions. As a consequence, the model predicts that individuals and populations with impairments in either or both DL-PFC or the DA system should demonstrate specific patterns of impaired cognitive control related to the processing of context.

A growing literature on the neurobiology of healthy aging suggests the PFC and DA systems are among the most strongly affected by increasing age [1], [13], [31], [32], [42]. We have explored the use of our theoretical model of cognitive control as a tool for understanding the nature of aging cognition, and to bridge the gap between neurobiological and cognitive findings of age effects. Next we describe in greater detail our model of cognitive control and how it might be used to simulate the cognitive and neurobiological consequences of healthy aging.