Cognitive Control and Memory Retrieval

Often stimuli in the world remind us of experiences or ideas with little effort, showing that retrieval can often proceed automatically. Though this automaticity is useful, there are times when it poses problems that require higher level cognitive control to resolve. Sometimes, we need to recall a particular experience, name, or fact, from memory, even though the cues available may remind us of other things that cause interference. Other times, cues may call to mind memories we would rather not think about. In each case, the automatic retrieval of a trace is unwanted because it is interfering, distracting or upsetting. Both situations trigger an attentional control response to handle excess activation on the undesired traces. What is the control process that enables us to manage such internal distraction? Much of the work in the Memory Control Lab focuses on the nature of these control processes, with a special focus given to the role of inhibitory control in helping us to regulate excess accessibility of activated memories. We are concerned with the functional properties of this inhibition mechanism, the consequences of recruiting it, and the manner in which it is implemented in the brain.

Inhibition refers to a hypothesized process that deactivates a trace, reducing the interference is causes in the present moment, and potentially affecting its later accessibility as well (Anderson, 2007). Inhibitory control is widely studied in cognitive neuroscience, and is viewed as a central component of our capacity for cognitive control in general (Anderson & Spellman, 1995, Anderson & Weaver, 2009). We study inhibitory control in the context of long-term memory, which also presents problems of selection and stopping that are isomorphic to those present in action and perception. Our unique approach focuses on careful documentation of the behavioural footprints of inhibition in the form of aftereffects on suppressed traces, and ties those aftereffects to their neural origins. We thus study the role of inhibitory control in memory through a multi-level approach using behavioural, fMRI, and EEG methods, and we ultimately, will use intracranial cellular recordings to understand this process at its most basic levels. These questions are pursued in a diversity of contexts, ranging from episodic memory for simple stimuli, autobiographical memory for real life events, spatial memory, and semantic memory. Moreover, we are in general interested in differences in inhibitory functions across populations defined by age, psychiatric status, and neuropsychological status. Studying such populations gives unique insights into the neural mechanisms underlying control, but also provides an opportunity to better understand those populations.

Control Processes in Selective Retrieval

When cues automatically activate traces other than the ones that we seek, it usually impedes retrieval. Because most stimuli have many associations, some of them very strong, this interference slows retrieval or may even lead to retrieval errors. A fundamental questions then is how do people succeed in selectively retrieving a desired memory in the face of such interference. Because retrieval is a ubiquitous cognitive function, if one could better understand how selective retrieval is accomplished, it would have implications for virtually every process that relies upon retrieval of knowledge or experience. Thus, learning about a general characteristic of retrieval might inform issues ranging from how we retrieve word meanings or sounds, to how we remember what we did last Thursday, to finding our way to where we parked today, as opposed to yesterday.

When one attempts to selectively retrieve a particular memory, fact, or idea, one is, in essence, trying to attend to a trace, despite distraction from other competing memories that may divert our attention or otherwise impede our focus on the target memory. This is a case of mnemonic selective attention. Theoretically, selective attention can be achieved by mechanisms that enhance or activate the target memory and also by suppressing competing traces by inhibitory control. In our work on selective retrieval, we have focused on the role of inhibitory control in achieving this function. At the behavioural level, our strategy builds on the observation that if inhibitory control is engaged to suppress distracting traces, one might expect to observe lingering aftereffects of suppression on the inhibited traces. Thus, by repeatedly retrieving some associates of a cue, one ought to be able to observe impaired retention of competing memories, a prediction referred to as retrieval-induced forgetting (Anderson, Bjork, & Bjork, 1994; Anderson & Spellman, 1995; Anderson,. 2003).

Over the last two decades, we have established that retrieval induced forgetting exists, and that it occurs in an incredibly diverse range of situations. Retrieving some members of a category impairs retention of other members; retrieving some facts about a topic, impairs retention of other facts associated to that same topic. Retrieving visual features of an object observed at one location disrupts retention of the color, shape, and location of others similar objects. Retrieving some aspects of an autobiographical memory impairs retention of other aspects. And retrieving words that you know in a foreign language while simply trying to speak impairs recall of the corresponding words in your native language. Retrieval-induced forgetting is thus a ubiquitous feature of our mental life that reflects the underlying mechanisms of how retrieval is accomplished. It also informs us about the mechanistic sources of our memory failures in a variety of contexts.

The work that we now do on retrieval induced forgetting tries to build on this strong behavioural foundation to understand how this process takes place in the brain. Current projects are concerned not only with the role of the lateral prefrontal cortex in helping to resolve competition during selective retrieval, but also in inducing disruption of competing memories. We are also actively seeking to establish the applicability of these mechanisms in helping to understand forgetting of real life experiences through the application of SenseCam technology.

Control Processes in Retrieval Stopping

Sometimes stimuli automatically elicit memories, even when we don't mean to be retrieving anything in particular at the time. If the unbidden memory is pleasant or neutral, and we don't have a particular need to focus on other tasks, such remindings are often welcome, and, in any case inevitable features of mental life. But when remindings are unpleasant or distracting, the efficiency and automaticity of memory retrieval works against our aims. We've all encountered reminders to things that we'd rather not think about, and when this happens, we make an effort to limit the duration of the reminding in awareness. When this happens, we are seeking to prevent or override retrieval, for some motivated purpose--either concentration or emotion regulation. A fundamental question is how we stop retrieval in this manner. What mechanisms are engaged? What are the consequences of engaging these mechanisms on the trace that has been excluded?

One important answer that we have explored over the years is that people engage inhibitory control processes of the sort that are widely studied in work on motor control to shut down the retrieval process. These processes, once engaged, suppress the excluded trace, rendering it less accessible that it otherwise would be. Thus, we see retrieval suppression as a prime case of inhibitory control applied to declarative memory (Anderson & Green, 2001). We have found that suppressing retrieval of a given memory repeatedly impairs one's later ability to recall it, and that this effect is produced, in part, by inhibitory control processes.

Much of the ongoing work in the laboratory concerns the mechanisms by which retrieval can be stopped. There are several reasons for this focus. First, the capacity to shut down retrieval is an extremely concrete and useful model system for studying the elementary computational function of stopping, and how stopping is achieved more broadly. Thus a full understanding of how the fundamental function of stopping is achieved in the brain is likely to come by comparing and contrasting several model cognitive and neural systems that implement stopping for various types of content, including stopping of motor actions, memories, and emotional responses. Second, the capacity to suppress retrieval is fundamentally an act directed at controlling conscious awareness of a memory. As such, this provides a means of studying consciousness and its voluntary regulation. Third, retrieval stopping s likely to be engaged during attempts to control awareness of unpleasant experiences and so provides a concrete means of studying a process central to affect regulation. The impact of retrieval stopping on the retention of suppressed traces moreover provides an intriguing model of motivated forgetting that may explain how people actively shape what we remember of our past to suit our needs. Finally, the stopping process itself reveals neural pathways by which frontal cortex can modulate hippocampal and neocortical activity, providing an important tool for study basic control operations of the brain. Thus, studying how retrieval stopping works is likely to inform a variety of issues of theoretical and applied significance (Anderson et al., 2004).

As such, much of the work on the subject focuses on trying to understand the foundational neurocognitive mechanisms underlying retrieval stopping, by making uses of cognitive analysis, fMRI, EEG, and soon, MEG. Our working hypothesis is that retrieval suppression is implemented in part by interactions between the lateral prefrontal cortex and the hippocampus, with the former serving to down-regulate mnemonic activity in the latter. Such down-regulation is thought to disrupt momentary recollection of the unwanted trace, and to impair its long term retention. However, our work also examines other mechanisms for achieving retrieval control, including modulation of neocortical activity, and processes relating to thought substitution. Ongoing collaborative work also seeks to relate what we learn about these fundamental mechanisms to understanding and possibly remediating disordered control over memory in post-traumatic stress disorder and depression.

Memory Control in Relation to General Attention Mechanisms

As the preceding section illustrates, we view the problems of memory control as special cases of broader problems of selection and stopping that occur not merely in memory but also in motor action and perception. As such, common mechanisms may mediate selection and stopping across these otherwise distinct areas of cognitive life. Alternatively, the problems may be solved by computationally isomorphic, but anatomically distinct processes. In any case, comparing the processes involved in various forms of stopping can be instructive about the shared and distinctive features of each.

To achieve greater understanding of memory control, we have studied the parallels between selection and stopping in memory and motor action. With Chelan Weaver, for instance, we have developed experimental procedures for isolating inhibitory aftereffects in motor action. Armed with these, we have further sought to identify the neural systems that mediate this motor inhibition with fMRI and to compare these to memory inhibition. With Ben Levy, Ean Huddleston, and Mike Posner at the University of Oregon, we have also compared the attentional orienting systems supporting perceptual attention with those observed in attention to memory representations. This work, supported by the National Science Foundation of the United States and by the Medical Research Council, seeks to integrate the control of memory representations within broader attentional frameworks.

Adaptive Forgetting of Experiences and Facts

One interesting implication of the foregoing view of memory control is that our attempts to interact with our memories often have inhibitory aftereffects on suppressed traces. If so, and if inhibition lingers, theoretically, it is possible that much of our forgetting derives from inhibition. This, is, forgetting may be a mere passive consequences of processes such as decay, interference or changes in context, but rather may derive from active suppression processes deployed for a functional purpose. In other words, forgetting may be functional, even adaptive (Bjork, 1988). This viewpoint starkly contrasts with the widely held feeling that forgetting is always a negative thing that happens to us passively.

An important objective of the lab is to assess the extent to which our forgetting may often reflect the action of such controlled mechanisms rather than merely being the result of passive processes. Several key publications have made the argument that we need to revise our view of how incidental forgetting occurs to reflect the manner in which memory control processes interact with traces to determine their longevity in memory (Anderson & Spellman, 1995; Levy & Anderson, 2002; Anderson, 2003) . Passive mechanisms surely do contribute to our experiences of forgetting, but the extent to which attentional control processes contribute to this appears to have been substantially unappreciated. We are actively involved in our own forgetting!

Motivated Forgetting in Clinical and Non-Clinical Populations

The capacity to control the retrieval process to exclude an unwanted memory from consciousness may provide a mechanism that could explain how people forget experiences that they do not wish to remember. If one must be forced to repeatedly confront reminders to an unpleasant experience, and one consistently takes action to suppress retrieval of the unwanted memory, the processes that suppress the trace may ultimately undermine its long-term retention, as we have demonstrated. A fundamental question is the extent to which these inhibitory aftereffects may generalize to the suppression of more realistic emotional memories of the sort that people wish to combat in the aftermath of a traumatic experience. Might retrieval suppression provide a good model of motivated forgetting?

To better understand this, we are taking a multi-pronged approach. First, we are interested in generalizing retrieval suppression to more naturalistic autobiographical experiences with emotional content. Second, are trying to determine whether the brain systems that people engage when they suppress their own personal traumatic experiences are related to the brain systems we have observed in more well controlled laboratory studies. Although a great deal of work remains to be done on this subject, substantial progress has already been made on building a case for retrieval suppression as a model of motivated forgetting (Anderson & Huddleston, 2011).