Browsing by Author "Egner, Tobias"

"Cognitive control" describes endogenous guidance of behavior in situations where routine stimulus-response associations are suboptimal for achieving a desired goal. The computational and neural mechanisms underlying this capacity remain poorly understood. The present dissertation examines recent advances stemming from the application of a statistical, Bayesian learner perspective on control processes. An important limitation in current models consists of a lack of a plausible mechanism for the flexible adjustment of control over variable environments. I propose that flexible cognitive control can be achieved by a Bayesian model with a self-adapting, volatility-driven learning scheme, which modulates dynamically the relative dependence on recent (short-term) and remote (long-term) experiences in its prediction of future control demand. Using simulation data, human behavioral data and human brain imaging data, I demonstrate that this Bayesian model does not only account for several classic behavioral phenomena observed from the cognitive control literature, but also facilitates a principled, model-guided investigation of the neural substrates underlying the flexible adjustment of cognitive control. Based on the results, I conclude that the proposed Bayesian model provides a feasible solution for modeling the flexible adjustment of cognitive control.

In the present study, we examined the predictions of an attentional-associative model (Schmajuk, Lam, & Gray Journal of Experimental Psychology: Animal Behavior Processes, 22, 321-349, 1996) regarding the effect of attentional manipulations on both discrete and contextual conditioned inhibitors.

The SLG model assumes that non-reinforced presentations of an inhibitory conditioned stimulus (CS) do not decrease its inhibitory associations. However, the model predicts that extended presentations will decrease attention to the inhibitor, thereby, decreasing both the expression of its inhibitory power in a summation test and the rate of acquisition in a retardation test. The model also predicts that subsequent presentations of the inhibitory CS with a novel CS will increase both its inhibitory power in a summation test and the rate of acquisition in a retardation test. Using a predictive learning design in humans, Experiment 1 examined the predictions involving the summation tests, whereas Experiments 2 and 3 examined the predictions involving the retardation tests. Experimental results were in agreement with the predictions of the model.

The SLG model also predicts that a salient extinction context (CX) becomes inhibitory and prevents extinction of the excitatory CS-unconditioned stimulus (US) association. Although some data seem to contradict that prediction (e.g., Bouton and King, 1983, Bouton and Swartzentruber, 1986, 1989), Larrauri and Schmajuk (2008) indicated that the CX might not appear inhibitory in a summation test because attention to the CX decreases with many but not few extinction trials. In a human predictive learning experiment, we confirmed the model's predictions that the inhibitory power of the extinction CX can be detected after a few extinction trials when attention to the CX is still high, but not after many extinction trials once attention to the CX has decreased (Experiment 4), and even after many extinction trials by presenting novel CSs to increase attention to the unattended CX (Experiment 5). Furthermore, using an eye-tracker, we confirmed the model's explanation of Experiment 4 results by showing decreased overt attention to the CX after many but not after few extinction trials (Experiment 6).

Importantly, the view that the extinction CX becomes inhibitory allows the model to explain spontaneous recovery (because attention to the excitatory CS increases before attention to the inhibitory CX), renewal (because the inhibition provided by the extinction CX disappears), and reinstatement (the inhibitory CX becomes neutral or excitatory), as well as a very large number of other experimental results related to extinction. Based on the prediction of the SLG, model the implications of our results for the treatments of anxiety disorders were discussed.

Of the core faculties underlying higher order human cognition, two of the most fundamental are the ability to focus on particular task while avoiding distraction (cognitive stability) and the ability to switch to new tasks in light of changing circumstances (cognitive flexibility). Research into the regulation of stability and flexibility has revealed that they often display an inverse relationship: prioritizing task focus (stability) is associated with an impaired ability to switch to new tasks (flexibility), and vice versa. Such findings have led to the common conception that stability and flexibility are endpoints of a one-dimensional stability-flexibility continuum, a perspective that requires them to reciprocate in all cases. However, many empirical findings seemingly contradict a one-dimensional account, such that stability and flexibility might better be described by two separate dimensions. The relationship between stability and flexibility therefore remains unclear. In the current dissertation, I present three studies that investigated the behavioral and computational mechanisms underlying concurrent stability and flexibility regulation. In the first study, I test the assumption of an obligatory stability-flexibility tradeoff, demonstrating across three behavioral experiments that stability and flexibility can be regulated independently. Next, I explore the putative mechanisms explaining tradeoffs, revealing that tradeoffs can be attributed to cost-of-control calculations seeking to reduce cognitive exertion in contexts where it is not incentivized. Third, I conduct a formal model comparison between a one- and a two-dimensional model of stability and flexibility, demonstrating that only the two-dimensional account can accurately reproduce empirical data patterns. The better-fitting two-dimensional model also reveals the dissociable influences stability and flexibility have on the putative decision process underlying task shielding and switching. Finally, I summarize the results and explore the implications of these studies for translational research.

Much of our real-life decision making is bounded by uncertain information, limitations in cognitive resources, and a lack of time to allocate to the decision process. To mitigate these pressures, people satisfice, foregoing a full evaluation of all available evidence to focus on a subset of cues that allow for fast and “good-enough” decisions. Although this form of decision-making likely mediates many of our everyday choices, very little is known about the manner in which satisficing is spontaneously triggered and accomplished. The aim of this dissertation, therefore, is to characterize cognitive and neural mechanisms underlying human satisficing behavior via tasks that closely model real-life challenges in decision making. Specifically, the empirical studies presented here examine how people solve a novel multi-cue probabilistic classification task under various external and internal pressures, using a set of strategy analyses based on variational Bayesian inference, which can track and quantify shifts in strategies. Results from these behavioral and computational approaches are then applied to model human functional magnetic resonance imaging (fMRI) data to investigate neural correlates of satisficing. The findings indicate that the human cognitive apparatus copes with uncertainty and various pressures by adaptively employing a “Drop-the-Worst” heuristic that minimizes cognitive time and effort investment while preserving the consideration of the most diagnostic cue information.

The stability-flexibility dilemma describes the challenge of balancing the antagonistic goals of focusing on the current task-set (cognitive stability) and updating that task-set in response to changes in the environment (cognitive flexibility). Dynamic adjustments of cognitive flexibility are observed in cued task-switching paradigms, wherein switch costs, or the performance costs associated with switching between tasks, have been shown to decrease as the proportion of switch trials within a block increase. This effect is referred to as the list-wide proportion switch (LWPS) effect, and presents evidence of meta-flexibility, or people’s ability to find an optimal level of flexibility based on contextual demands. While context-sensitive control adaptations have been extensively researched in the conflict literature, fewer studies have been dedicated to investigating such flexibility adaptations. Consequently, the underlying mechanisms of meta-flexibility remain unknown. Across four behavioral experiments, Chapter 2 teases apart the different levels of learning that may contribute to list-wide flexibility adaptations, controlling for stimulus- and task-level associations. Chapter 3 investigates the EEG neural signatures of meta-flexibility. In Chapter 4, we test for whether learned flexibility adjustments benefit from memory consolidation, like other forms of associative learning. Lastly, Chapter 5 utilized three novel behavioral paradigms to investigate different conditions under which flexibility learning transfers or fails to transfer. Collectively, the results in this dissertation suggest that flexibility adjustments to contextual demands occurs rapidly and transfers reliably across novel stimuli, such that, in high switch frequency blocks, participants could perform more rapid task switches even on trials involving items never seen before. However, both behavioral and neural evidence suggest that flexibility learning is also task specific, such that switch performance boosts to tasks that appear more often as switch trials do not generalize to other tasks that occur in the same temporal contexts.

For over 50 years, the Satisfaction of Search effect, and more recently known as the Subsequent Search Miss (SSM) effect, has plagued the field of radiology. Defined as a decrease in additional target accuracy after detecting a prior target in a visual search, SSM errors are known to underlie both real-world search errors (e.g., a radiologist is more likely to miss a tumor if a different tumor was previously detected) and more simplified, lab-based search errors (e.g., an observer is more likely to miss a target ‘T’ if a different target ‘T’ was previously detected). Unfortunately, little was known about this phenomenon’s cognitive underpinnings and SSM errors have proven difficult to eliminate. However, more recently, experimental research has provided evidence for three different theories of SSM errors: the Satisfaction account, the Perceptual Set account, and the Resource Depletion account. A series of studies examined performance in a multiple-target visual search and aimed to provide support for the Resource Depletion account—a first target consumes cognitive resources leaving less available to process additional targets.

To assess a potential mechanism underlying SSM errors, eye movements were recorded in a multiple-target visual search and were used to explore whether a first target may result in an immediate decrease in second-target accuracy, which is known as an attentional blink. To determine whether other known attentional distractions amplified the effects of finding a first target has on second-target detection, distractors within the immediate vicinity of the targets (i.e., clutter) were measured and compared to accuracy for a second target. To better understand which characteristics of attention were impacted by detecting a first target, individual differences within four characteristics of attention were compared to second-target misses in a multiple-target visual search.

The results demonstrated that an attentional blink underlies SSM errors with a decrease in second-target accuracy from 135ms-405ms after detection or re-fixating a first target. The effects of clutter were exacerbated after finding a first target causing a greater decrease in second-target accuracy as clutter increased around a second-target. The attentional characteristics of modulation and vigilance were correlated with second- target misses and suggest that worse attentional modulation and vigilance are predictive of more second-target misses. Taken together, these result are used as the foundation to support a new theory of SSM errors, the Flux Capacitor theory. The Flux Capacitor theory predicts that once a target is found, it is maintained as an attentional template in working memory, which consumes attentional resources that could otherwise be used to detect additional targets. This theory not only proposes why attentional resources are consumed by a first target, but encompasses the research in support of all three SSM theories in an effort to establish a grand, unified theory of SSM errors.

Internal states, such as one’s emotions, goals, and motivations, have profound impacts on how people perceive the world, interact with their environment, and remember their experiences. Understanding how internal states influence learning behaviors and episodic memory remains a critical challenge for the field of psychology and neuroscience. In this dissertation, I present three empirical studies that approach this challenge by examining how shifts in internal states influence event segmentation and memory organization, and how sustained internal states shape choices during reward learning and memory formation. In the first study, I tested whether shifts in internal goal states, as defined by switching between different task rules, created event boundaries in memory. Across five behavioral experiments, I found that both cued and voluntary task switching led to temporal memory effects that characterize event segmentation, and these effects cannot be attributed to perceptual changes or changes in task difficulty. These results demonstrate that shifts in internal goal states, even when they are generated without any external cue as in the case of voluntary task switching, organize event memory. In the second study, I further explored the role of internal goal states in episodic memory and found that transient increases in attention do not impact memory for temporal relationships between items despite robust boost to recognition memory for individual items encoded during behaviorally relevant moments. These results show that fluctuations in attention without updating internal goal states are insufficient to influence temporal organization of memory. In the third study, I used computational modeling and neuroimaging methods to investigate how sustained motivational states that focus on proximal and distal goals respectively influence learning behaviors and memory outcomes in a reinforcement learning environment. I found that sustained motivational states shifted the balance between reward-maximizing exploitation behavior and uncertainty-directed exploration behavior during reward learning. Furthermore, successful memory formation was supported by distinct neural routes for participants motivated by an urgent performance goal and those motivated by a future learning goal. These results suggest that different sustained internal states under the same reward learning environment can lead to divergent behavioral outcomes and differential neural mechanisms for memory encoding. Collectively, the work presented in this dissertation demonstrates the importance of internal states in learning and memory processes and furthers our understanding of how goals and motivations shape the human experience.

Cognitive control refers to the use of internal goals to guide how we process stimuli and select responses, and control can be applied proactively (in anticipation of a stimulus) or reactively (once that stimulus has been presented). The application of control can be guided by memory (“control-learning”); for instance, people typically learn to adjust their level of attentional selectivity to changing task statistics, such as different frequencies of hard and easy trials on attention-demanding tasks. This type of control-learning is highly adaptive, but its boundary conditions are not well understood. The aim of this dissertation, therefore, is to examine three core principles thought to underlie control-learning: its context-sensitivity, reward-sensitivity, and implicit nature. Two chapters use standard control-learning paradigms that manipulate the proportion of (easy) Stroop congruent trials within blocks and for specific stimuli. Here, I show that people can learn to generalize learned control across related contexts and that reinforcement may selectively enhance the recruitment of control in some specific contexts, if at all. The last chapter deploys a precuing paradigm to test whether conscious cue perception and knowledge impacts the recruitment of control, complementing the previous chapters with a causal manipulation of explicit awareness and showing that people make control adjustments most easily when they can consciously perceive and are aware of upcoming control-demand. These results have important implications for experimental designs, potential psychiatric treatment, and theoretical accounts of the mechanisms underlying control-learning. My findings will broaden our understanding of the relationship between attention and memory and add insight into how people can learn to flexibly adapt their processing strategies to changing demands in a context-dependent manner.

Attention, the cognitive means by which we prioritize the processing of a subset of information, is necessary for operating efficiently and effectively in the world. Thus, a critical theoretical question is how information is selected. In the visual domain, working memory (WM)—which refers to the short-term maintenance and manipulation of information that is no longer accessible by the senses—has been highlighted as an important determinant of what is selected by visual attention. Furthermore, although WM and attention have traditionally been conceived as separate cognitive constructs, an abundance of behavioral and neural evidence indicates that these two domains are in fact intertwined and overlapping. The aim of this dissertation is to better understand the nature of WM and attention, primarily through the phenomenon of memory-based attentional guidance, whereby the active maintenance of items in visual WM reliably biases the deployment of attention to memory-matching items in the visual environment. The research presented here employs a combination of behavioral, functional imaging, and computational modeling techniques that address: (1) WM guidance effects with respect to the traditional dichotomy of top-down versus bottom-up attentional control; (2) under what circumstances the contents of WM impact visual attention; and (3) the broader hypothesis of a predictive and competitive interaction between WM and attention. Collectively, these empirical findings reveal the importance of WM as a distinct factor in attentional control and support current models of multiple-state WM, which may have broader implications for how we select and maintain information.

The overall goal of this dissertation was to investigate the interaction between memory processes and implementation of cognitive control, focusing on the bi-directional relationship between the two. Working memory can serve as a source of interference for the implementation of cognitive control and episodic memory can guide the successful implementation of cognitive control.

Chapter 2 investigated how cognitive control is implemented over working memory-guided attentional processes. While the contents of working memory can guide external attention to similar or matching items whether adaptive or not, cognitive control can act over this guidance in service of top-down goals, limiting its influence when maladaptive. While previous work had demonstrated that control can act over working memory, how cognitive control specifically acts over the attentional capture process driven by working memory remained unclear. Here I presented work from electroencephalography (EEG) in order to demonstrate that cognitive control acts over the post-orienting processing of visual stimuli, reactively reducing the need to marshal additional resources to limit the interference on attention exerted by the contents of working memory. This demonstrated how control acts over working memory-guided attention, relying on reactive implementation over the post-orientation processing of stimuli, not the attentional orienting processes itself, in order to reduce the influence of working memory contents on attention in service of top-down goals.

In Chapter 3, I further demonstrated that cognitive control can be implemented over complex, action-related representations in working memory (i.e. task-sets). While prior work had focused largely on the influence from simple representations in working memory, much of the contents of our internal thoughts are action-oriented. Further, previous work had characterized the influence of working memory action-representations on ongoing behavior as automatic. However, in this chapter I presented two sets of studies which demonstrated that cognitive control can be proactively implemented on the creation of action representations in working memory, as well as reactively implemented over their interference in ongoing behavior, according to current goals and task demands. This extends the role of cognitive control to action representations in working memory, expanding our knowledge of how cognitive control acts over complex internal thoughts.

Finally, in Chapter 4, I demonstrated how memory processes not only serve as sources of interference for the implementation of cognitive control, but can also guide the successful implementation of cognitive control. Previous work had proposed that successful implementation of cognitive control was driven through episodic learning processes. Here I first present a set of behavioral studies demonstrating for the first time that a control state can be mnemonically associated with a specific stimulus in episodic-supported event-files in memory given the one-shot pairing of a stimulus and control state. Further, this episodic memory-supported event-file and associated control state can then be retrieved later, triggered by the reappearance of a stimulus, in order to successfully implement cognitive control. Further, using pupillometry as a measure of attention, I presented work showing that the encoding of these stimulus-control associations requires active, engaged attention, but their retrieval is automatic once encoded. This work demonstrates the role of memory processes in supporting the successful implementation of cognitive control.

All of us are taxed with juggling our inner mental lives with immediate external task demands. For many years, the temporary maintenance of internal information was considered to be handled by a dedicated working memory (WM) system. It has recently become increasingly clear, however, that such short-term internal activation interacts with attention focused on external stimuli. It is unclear, however, exactly why these two interact, at what level of processing, and to what degree. Because our internal maintenance and external attention processes co-occur with one another, the manner of their interaction has vast implications for functioning in daily life. The work described here has employed original experimental paradigms combining WM and attention task elements, functional magnetic resonance imaging (fMRI) to illuminate the associated neural processes, and transcranial magnetic stimulation (TMS) to clarify the causal substrates of attentional brain function. These studies have examined a mechanism that might explain why (and when) the content of WM can involuntarily capture visual attention. They have, furthermore, tested whether fundamental attentional selection processes operate within WM, and whether they are reciprocal with attention. Finally, they have illuminated the neural consequences of competing attentional demands. The findings indicate that WM shares representations, operating principles, and cognitive resources with externally-oriented attention.