Value-driven attentional capture enhances distractor representations in early visual cortex

Reward-modulated attention deployment is driven by suppression, not attentional capture

One driving factor for attention deployment towards a stimulus is its associated value due to previous experience and learning history. Previous visual search studies found that when looking for a target, distractors associated with higher reward produce more interference (e.g., longer response times). The present study investigated the neural mechanism of such value-driven attention deployment. Specifically, we were interested in which of the three attention sub-processes are responsible for the interference that was repeatedly observed behaviorally: enhancement of relevant information, attentional capture by irrelevant information, or suppression of irrelevant information. We replicated earlier findings showing longer response times and lower accuracy when a target competed with a high-reward compared to a low-reward distractor. We also found a spatial gradient of interference: behavioral performance dropped with increasing proximity to the target. This gradient was steeper for high- than low-reward distractors. Event-related potentials of the EEG signal showed the reason for the reward-induced attentional bias: High-reward distractors required more suppression than low-reward distractors as evident in larger Pd components. This effect was only found for distractors near targets, showing the additional filtering needs required for competing stimuli in close proximity. As a result, fewer attentional resources can be distributed to the target when it competes with a high-reward distractor, as evident in a smaller target-N2pc amplitude. The distractor-N2pc, indicative of attentional capture, was neither affected by distance nor reward, showing that attentional capture alone cannot explain interference by stimuli of high value. In sum our results show that the higher need for suppression of high-value stimuli contributes to reward-modulated attention deployment and increased suppression can prevent attentional capture of high-value stimuli.

From learned value to sustained bias: how reward conditioning changes attentional priority

Introduction

Attentional bias to reward-associated stimuli can occur even when it interferes with goal-driven behavior. One theory posits that dopaminergic signaling in the striatum during reward conditioning leads to changes in visual cortical and parietal representations of the stimulus used, and this, in turn, sustains attentional bias even when reward is discontinued. However, only a few studies have examined neural activity during both rewarded and unrewarded task phases.

Methods

In the current study, participants first completed a reward-conditioning phase, during which responses to certain stimuli were associated with monetary reward. These stimuli were then included as non-predictive cues in a spatial cueing task. Participants underwent functional brain imaging during both task phases.

Results

The results show that striatal activity during the learning phase predicted increased visual cortical and parietal activity and decreased ventro-medial prefrontal cortex activity in response to conditioned stimuli during the test. Striatal activity was also associated with anterior cingulate cortex activation when the reward-conditioned stimulus directed attention away from the target.

Discussion

Our findings suggest that striatal activity during reward conditioning predicts the degree to which reward history biases attention through learning-induced changes in visual and parietal activities.

Trichotomy revisited: A monolithic theory of attentional control

The control of attention was long held to reflect the influence of two competing mechanisms of assigning priority, one goal-directed and the other stimulus-driven. Learning-dependent influences on the control of attention that could not be attributed to either of those two established mechanisms of control gave rise to the concept of selection history and a corresponding third mechanism of attentional control. The trichotomy framework that ensued has come to dominate theories of attentional control over the past decade, replacing the historical dichotomy. In this theoretical review, I readily affirm that distinctions between the influence of goals, salience, and selection history are substantive and meaningful, and that abandoning the dichotomy between goal-directed and stimulus-driven mechanisms of control was appropriate. I do, however, question whether a theoretical trichotomy is the right answer to the problem posed by selection history. If we reframe the influence of goals and selection history as different flavors of memory-dependent modulations of attentional priority and if we characterize the influence of salience as a consequence of insufficient competition from such memory-dependent sources of priority, it is possible to account for a wide range of attention-related phenomena with only one mechanism of control. The monolithic framework for the control of attention that I propose offers several concrete advantages over a trichotomy framework, which I explore here.

Distinct Patterns of Connectivity between Brain Regions Underlie the Intra-Modal and Cross-Modal Value-Driven Modulations of the Visual Cortex

Past reward associations may be signaled from different sensory modalities; however, it remains unclear how different types of reward-associated stimuli modulate sensory perception. In this human fMRI study (female and male participants), a visual target was simultaneously presented with either an intra- (visual) or a cross-modal (auditory) cue that was previously associated with rewards. We hypothesized that, depending on the sensory modality of the cues, distinct neural mechanisms underlie the value-driven modulation of visual processing. Using a multivariate approach, we confirmed that reward-associated cues enhanced the target representation in early visual areas and identified the brain valuation regions. Then, using an effective connectivity analysis, we tested three possible patterns of connectivity that could underlie the modulation of the visual cortex: a direct pathway from the frontal valuation areas to the visual areas, a mediated pathway through the attention-related areas, and a mediated pathway that additionally involved sensory association areas. We found evidence for the third model demonstrating that the reward-related information in both sensory modalities is communicated across the valuation and attention-related brain regions. Additionally, the superior temporal areas were recruited when reward was cued cross-modally. The strongest dissociation between the intra- and cross-modal reward-driven effects was observed at the level of the feedforward and feedback connections of the visual cortex estimated from the winning model. These results suggest that, in the presence of previously rewarded stimuli from different sensory modalities, a combination of domain-general and domain-specific mechanisms are recruited across the brain to adjust the visual perception.SIGNIFICANCE STATEMENT Reward has a profound effect on perception, but it is not known whether shared or disparate mechanisms underlie the reward-driven effects across sensory modalities. In this human fMRI study, we examined the reward-driven modulation of the visual cortex by visual (intra-modal) and auditory (cross-modal) reward-associated cues. Using a model-based approach to identify the most plausible pattern of inter-regional effective connectivity, we found that higher-order areas involved in the valuation and attentional processing were recruited by both types of rewards. However, the pattern of connectivity between these areas and the early visual cortex was distinct between the intra- and cross-modal rewards. This evidence suggests that, to effectively adapt to the environment, reward signals may recruit both domain-general and domain-specific mechanisms.

Suppressive Control of Incentive Salience in Real-World Human Vision

Reward-related activity in the dopaminergic midbrain is thought to guide animal behavior, in part by boosting the perceptual and attentional processing of reward-predictive environmental stimuli. In line with this incentive salience hypothesis, studies of human visual search have shown that simple synthetic stimuli, such as lines, shapes, or Gabor patches, capture attention to their location when they are characterized by reward-associated visual features, such as color. In the real world, however, we commonly search for members of a category of visually heterogeneous objects, such as people, cars, or trees, where category examples do not share low-level features. Is attention captured to examples of a reward-associated real-world object category? Here, we have human participants search for targets in photographs of city and landscapes that contain task-irrelevant examples of a reward-associated category. We use the temporal precision of EEG machine learning and ERPs to show that these distractors acquire incentive salience and draw attention, but do not capture it. Instead, we find evidence of rapid, stimulus-triggered attentional suppression, such that the neural encoding of these objects is degraded relative to neutral objects. Humans appear able to suppress the incentive salience of reward-associated objects when they know these objects will be irrelevant, supporting the rapid deployment of attention to other objects that might be more useful. Incentive salience is thought to underlie key behaviors in eating disorders and addiction, among other conditions, and the kind of suppression identified here likely plays a role in mediating the attentional biases that emerge in these circumstances.Significance Statement Like other animals, humans are prone to notice and interact with environmental objects that have proven rewarding in earlier experience. However, it is common that such objects have no immediate strategic use and are therefore distracting. Do these reward-associated real-world objects capture our attention, despite our strategic efforts otherwise? Or are we able to strategically control the impulse to notice them? Here we use machine learning classification of human electrical brain activity to show that we can establish strategic control over the salience of naturalistic reward-associated objects. These objects draw our attention, but do not necessarily capture it, and this kind of control may play an important role in mediating conditions like eating disorder and addiction.

Neural correlates of value-driven spatial orienting

Reward learning has been shown to habitually guide overt spatial attention to specific regions of a scene. However, the neural mechanisms that support this bias are unknown. In the present study, participants learned to orient themselves to a particular quadrant of a scene (a high-value quadrant) to maximize monetary gains. This learning was scene-specific, with the high-value quadrant varying across different scenes. During a subsequent test phase, participants were faster at identifying a target if it appeared in the high-value quadrant (valid), and initial saccades were more likely to be made to the high-value quadrant. fMRI analyses during the test phase revealed learning-dependent priority signals in the caudate tail, superior colliculus, frontal eye field, anterior cingulate cortex, and insula, paralleling findings concerning feature-based, value-driven attention. In addition, ventral regions typically associated with scene selection and spatial information processing, including the hippocampus, parahippocampal gyrus, and temporo-occipital cortex, were also implicated. Taken together, our findings offer new insights into the neural architecture subserving value-driven attention, both extending our understanding of nodes in the attention network previously implicated in feature-based, value-driven attention and identifying a ventral network of brain regions implicated in reward's influence on scene-dependent spatial orienting.

Differential effects of intra-modal and cross-modal reward value on perception: ERP evidence

In natural environments objects comprise multiple features from the same or different sensory modalities but it is not known how perception of an object is affected by the value associations of its constituent parts. The present study compares intra- and cross-modal value-driven effects on behavioral and electrophysiological correlates of perception. Human participants first learned the reward associations of visual and auditory cues. Subsequently, they performed a visual discrimination task in the presence of previously rewarded, task-irrelevant visual or auditory cues (intra- and cross-modal cues, respectively). During the conditioning phase, when reward associations were learned and reward cues were the target of the task, high value stimuli of both modalities enhanced the electrophysiological correlates of sensory processing in posterior electrodes. During the post-conditioning phase, when reward delivery was halted and previously rewarded stimuli were task-irrelevant, cross-modal value significantly enhanced the behavioral measures of visual sensitivity, whereas intra-modal value produced only an insignificant decrement. Analysis of the simultaneously recorded event-related potentials (ERPs) of posterior electrodes revealed similar findings. We found an early (90-120 ms) suppression of ERPs evoked by high-value, intra-modal stimuli. Cross-modal stimuli led to a later value-driven modulation, with an enhancement of response positivity for high- compared to low-value stimuli starting at the N1 window (180-250 ms) and extending to the P3 (300-600 ms) responses. These results indicate that sensory processing of a compound stimulus comprising a visual target and task-irrelevant visual or auditory cues is modulated by the reward value of both sensory modalities, but such modulations rely on distinct underlying mechanisms.

Tasks and their role in visual neuroscience

Vision is widely used as a model system to gain insights into how sensory inputs are processed and interpreted by the brain. Historically, careful quantification and control of visual stimuli have served as the backbone of visual neuroscience. There has been less emphasis, however, on how an observer's task influences the processing of sensory inputs. Motivated by diverse observations of task-dependent activity in the visual system, we propose a framework for thinking about tasks, their role in sensory processing, and how we might formally incorporate tasks into our models of vision.

Value-driven modulation of visual perception by visual and auditory reward cues: The role of performance-contingent delivery of reward

Perception is modulated by reward value, an effect elicited not only by stimuli that are predictive of performance-contingent delivery of reward (PC) but also by stimuli that were previously rewarded (PR). PC and PR cues may engage different mechanisms relying on goal-driven versus stimulus-driven prioritization of high value stimuli, respectively. However, these two modes of reward modulation have not been systematically compared against each other. This study employed a behavioral paradigm where participants' visual orientation discrimination was tested in the presence of task-irrelevant visual or auditory reward cues. In the first phase (PC), correct performance led to a high or low monetary reward dependent on the identity of visual or auditory cues. In the subsequent phase (PR), visual or auditory cues were not followed by reward delivery anymore. We hypothesized that PC cues have a stronger modulatory effect on visual discrimination and pupil responses compared to PR cues. We found an overall larger task-evoked pupil dilation in PC compared to PR phase. Whereas PC and PR cues both increased the accuracy of visual discrimination, value-driven acceleration of reaction times (RTs) and pupillary responses only occurred for PC cues. The modulation of pupil size by high reward PC cues was strongly correlated with the modulation of a combined measure of speed and accuracy. These results indicate that although value-driven modulation of perception can occur even when reward delivery is halted, stronger goal-driven control elicited by PC reward cues additionally results in a more efficient balance between accuracy and speed of perceptual choices.

Mental control of uncertainty

Can you reduce uncertainty by thinking? Intuition suggests that this happens through the elusive process of attention: if we expend mental effort, we can increase the reliability of our sensory data. Models based on "rational inattention" formalize this idea in terms of a trade-off between the costs and benefits of attention. This paper surveys the origin of these models in economics, their connection to rate-distortion theory, and some of their recent applications to psychology and neuroscience. We also report new data from a numerosity judgment task in which we manipulate performance incentives. Consistent with rational inattention, people are able to improve performance on this task when incentivized, in part by increasing the reliability of their sensory data.

Guiding spatial attention by multimodal reward cues

Our attention is constantly captured and guided by visual and/or auditory inputs. One key contributor to selecting relevant information from the environment is reward prospect. Intriguingly, while both multimodal signal processing and reward effects on attention have been widely studied, research on multimodal reward signals is lacking. Here, we investigated this using a Posner task featuring peripheral cues of different modalities (audiovisual/visual/auditory), reward prospect (reward/no-reward), and cue-target stimulus-onset asynchronies (SOAs 100-1,300 ms). We found that audiovisual and visual reward cues (but not auditory ones) enhanced cue-validity effects, albeit with different time courses (Experiment 1). While the reward-modulated validity effect of visual cues was pronounced at short SOAs, the effect of audiovisual reward cues emerged at longer SOAs. Follow-up experiments exploring the effects of visual (Experiment 2) and auditory (Experiment 3) reward cues in isolation showed that reward modulated performance only in the visual condition. This suggests that the differential effect of visual and auditory reward cues in Experiment 1 is not merely a result of the mixed cue context, but confirms that visual reward cues have a stronger impact on attentional guidance in this paradigm. Taken together, it seems that adding an auditory reward cue to the inherently dominant visual one led to a shift/extension of the validity effect in time - instead of increasing its amplitude. While generally being in line with a multimodal cuing benefit, this specific pattern highlights that different reward signals are not simply combined in a linear fashion but lead to a qualitatively different process.

Value-Biased Competition in the Auditory System of the Brain

Attentional capture by previously reward-associated stimuli has predominantly been measured in the visual domain. Recently, behavioral studies of value-driven attention have demonstrated involuntary attentional capture by previously reward-associated sounds, emulating behavioral findings within the visual domain and suggesting a common mechanism of attentional capture by value across sensory modalities. However, the neural correlates of the modulatory role of learned value on the processing of auditory information has not been examined. Here, we conducted a neuroimaging study on human participants using a previously established behavioral paradigm that measures value-driven attention in an auditory target identification task. We replicate behavioral findings of both voluntary prioritization and involuntary attentional capture by previously reward-associated sounds. When task-relevant, the selective processing of high-value sounds is supported by reduced activation in the dorsal attention network of the visual system (FEF, intraparietal sulcus, right middle frontal gyrus), implicating cross-modal processes of biased competition. When task-irrelevant, in contrast, high-value sounds evoke elevated activation in posterior parietal cortex and are represented with greater fidelity in the auditory cortex. Our findings reveal two distinct mechanisms of prioritizing reward-related auditory signals, with voluntary and involuntary modes of orienting that are differently manifested in biased competition.

Motivational Salience Guides Attention to Valuable and Threatening Stimuli: Evidence from Behavior and Functional Magnetic Resonance Imaging

Rewarding and aversive outcomes have opposing effects on behavior, facilitating approach and avoidance, although we need to accurately anticipate each type of outcome to behave effectively. Attention is biased toward stimuli that have been learned to predict either type of outcome, and it remains an open question whether such orienting is driven by separate systems for value- and threat-based orienting or whether there exists a common underlying mechanism of attentional control driven by motivational salience. Here, we provide a direct comparison of the neural correlates of value- and threat-based attentional capture after associative learning. Across multiple measures of behavior and brain activation, our findings overwhelmingly support a motivational salience account of the control of attention. We conclude that there exists a core mechanism of experience-dependent attentional control driven by motivational salience and that prior characterizations of attention as being value driven or supporting threat monitoring need to be revisited.

The past, present, and future of selection history

The last ten years of attention research have witnessed a revolution, replacing a theoretical dichotomy (top-down vs. bottom-up control) with a trichotomy (biased by current goals, physical salience, and selection history). This third new mechanism of attentional control, selection history, is multifaceted. Some aspects of selection history must be learned over time whereas others reflect much more transient influences. A variety of different learning experiences can shape the attention system, including reward, aversive outcomes, past experience searching for a target, target‒non-target relations, and more. In this review, we provide an overview of the historical forces that led to the proposal of selection history as a distinct mechanism of attentional control. We then propose a formal definition of selection history, with concrete criteria, and identify different components of experience-driven attention that fit within this definition. The bulk of the review is devoted to exploring how these different components relate to one another. We conclude by proposing an integrative account of selection history centered on underlying themes that emerge from our review.

How Native Background Affects Human Performance in Real-World Visual Object Detection: An Event-Related Potential Study

Visual processing refers to the process of perceiving, analyzing, synthesizing, manipulating, transforming, and thinking of visual objects. It is modulated by both stimulus-driven and goal-directed factors and manifested in neural activities that extend from visual cortex to high-level cognitive areas. Extensive body of studies have investigated the neural mechanisms of visual object processing using synthetic or curated visual stimuli. However, synthetic or curated images generally do not accurately reflect the semantic links between objects and their backgrounds, and previous studies have not provided answers to the question of how the native background affects visual target detection. The current study bridged this gap by constructing a stimulus set of natural scenes with two levels of complexity and modulating participants' attention to actively or passively attend to the background contents. Behaviorally, the decision time was elongated when the background was complex or when the participants' attention was distracted from the detection task, and the object detection accuracy was decreased when the background was complex. The results of event-related potentials (ERP) analysis explicated the effects of scene complexity and attentional state on the brain responses in occipital and centro-parietal areas, which were suggested to be associated with varied attentional cueing and sensory evidence accumulation effects in different experimental conditions. Our results implied that efficient visual processing of real-world objects may involve a competition process between context and distractors that co-exist in the native background, and extensive attentional cues and fine-grained but semantically irrelevant scene information were perhaps detrimental to real-world object detection.

Stimulus-response learning and expected reward value enhance stimulus cognitive processing: An ERP study

Reward affects our attention to stimuli, prioritizing those that lead to high-value outcomes. Recently, it has been suggested that such reward-related cognitive prioritization might be associated with the process of learning new stimulus-response (S-R) associations, because both are acquired through extended reward training, and once established, they are hard to overcome. We used event-related potentials (ERP) to analyze the contribution of S-R links to the formation of reward-related cognitive prioritization during reinforcement learning. Reward-related cognitive prioritization was measured by comparing the ERP signals for stimuli predicting high-value and low-value outcomes. In addition, we compared a strong S-R link (same stimulus, same response), with a weak S-R link condition (same stimulus, two different responses). The participants' performance was more accurate and faster when the procedure allowed for establishing strong S-R links and for high-value outcomes. Furthermore, those stimuli associated with strong S-R links showed a larger P3 amplitude at parietal sites. Value effects (larger ERP activity for those stimuli predicting a high-value outcome) were obtained at parietal and occipital sites in the P3 time window. However, value effects did not benefit from strong S-R links in either the P1 or the P3 components. These results suggest that strong S-R learning is not necessary to develop reward-related modulations of ERP activity.

Combined influence of valence and statistical learning on the control of attention: Evidence for independent sources of bias

Selection history exerts a powerful influence on the control of attention. Stimuli signalling reward and punishment capture attention even when physically non-salient and task-irrelevant. Repeated presentation of a salient distractor at a particular location generates learned suppression, resulting in reduced attentional processing at that location. A debate in the field concerns whether different components of selection history influence attention via a common underlying mechanism of learning-dependent control or via distinct, independent mechanisms. We probed this question with a particular focus on reward/punishment history and learned suppression. Participants were trained to suppress a particular location (high probability distractor location) and associate colours with reward or no outcome (no-reward). In a subsequent task, reward and no-reward distractors appeared in all locations equally often. In a separate experiment, we replaced reward with electric shocks. Reward and shock distractors captured attention more strongly than no-reward and no-shock distractors irrespective of their location. Distractors appearing in the high probability location showed reduced capture irrespective of their type. The results imply that reward and punishment learning and learned suppression have independent influences on the attentional system.

Multiregional communication and the channel modulation hypothesis

Multiregional communication is important to understanding the brain mechanisms supporting complex behaviors. Work in animals and human subjects shows that multiregional communication plays significant roles in cognitive function and is associated with neurological and neuropsychiatric disorders of brain function. Recent experimental advances enable empirical tests of the mechanisms of multiregional communication. Recent mechanistic insights into brain network function also suggest new therapies to treat disordered brain networks. Here, we discuss how to use the concept of communication channel modulation can help define and constrain what we mean by multiregional communication. We discuss behavioral and neurophysiological evidence for multiregional channels modulation. We then consider the role of causal manipulations and their implications for developing novel therapies based on multiregional communication.

Selection history is relative

Visual attention can be tuned to specific features to aid in visual search. The way in which these search strategies are established and maintained is flexible, reflecting goal-directed attentional control, but can exert a persistent effect on selection that remains even when these strategies are no longer advantageous, reflecting an attentional bias driven by selection history. Apart from feature-specific search, recent studies have shown that attention can be tuned to target-nontarget relationships. Here we tested whether a relational search strategy continues to bias attention in a subsequent task, where the relationally better color and former target color both serve as distractors (Experiment 1) or as potential targets (Experiment 2). We demonstrate that a relational bias can persist in a subsequent task in which color serves as a task-irrelevant feature, both impairing and facilitating visual search performance. Our findings extend our understanding of the relational account of attentional control and the nature of selection history effects on attention.

Previously Reward-Associated Stimuli Capture Spatial Attention in the Absence of Changes in the Corresponding Sensory Representations as Measured with MEG

Studies of selective attention typically consider the role of task goals or physical salience, but attention can also be captured by previously reward-associated stimuli, even if they are currently task irrelevant. One theory underlying this value-driven attentional capture (VDAC) is that reward-associated stimulus representations undergo plasticity in sensory cortex, thereby automatically capturing attention during early processing. To test this, we used magnetoencephalography to probe whether stimulus location and identity representations in sensory cortex are modulated by reward learning. We furthermore investigated the time course of these neural effects, and their relationship to behavioral VDAC. Male and female human participants first learned stimulus-reward associations. Next, we measured VDAC in a separate task by presenting these stimuli in the absence of reward contingency and probing their effects on the processing of separate target stimuli presented at different time lags. Using time-resolved multivariate pattern analysis, we found that learned value modulated the spatial selection of previously rewarded stimuli in posterior visual and parietal cortex from ∼260 ms after stimulus onset. This value modulation was related to the strength of participants' behavioral VDAC effect and persisted into subsequent target processing. Importantly, learned value did not influence cortical signatures of early processing (i.e., earlier than ∼200 ms); nor did it influence the decodability of stimulus identity. Our results suggest that VDAC is underpinned by learned value signals that modulate spatial selection throughout posterior visual and parietal cortex. We further suggest that VDAC can occur in the absence of changes in early visual processing in cortex.SIGNIFICANCE STATEMENT Attention is our ability to focus on relevant information at the expense of irrelevant information. It can be affected by previously learned but currently irrelevant stimulus-reward associations, a phenomenon termed "value-driven attentional capture" (VDAC). The neural mechanisms underlying VDAC remain unclear. It has been speculated that reward learning induces visual cortical plasticity, which modulates early visual processing to capture attention. Although we find that learned value modulates spatial signals in visual cortical areas, an effect that correlates with VDAC, we find no relevant signatures of changes in early visual processing in cortex.

The Importance of Considering Model Choices When Interpreting Results in Computational Neuroimaging

Model-based analyses open exciting opportunities for understanding neural information processing. In a commentary published in eNeuro, Gardner and Liu (2019) discuss the role of model specification in interpreting results derived from complex models of neural data. As a case study, they suggest that one such analysis, the inverted encoding model (IEM), should not be used to assay properties of "stimulus representations" because the ability to apply linear transformations at various stages of the analysis procedure renders results "arbitrary." Here, we argue that the specification of all models is arbitrary to the extent that an experimenter makes choices based on current knowledge of the model system. However, the results derived from any given model, such as the reconstructed channel response profiles obtained from an IEM analysis, are uniquely defined and are arbitrary only in the sense that changes in the model can predictably change results. IEM-based channel response profiles should therefore not be considered arbitrary when the model is clearly specified and guided by our best understanding of neural population representations in the brain regions being analyzed. Intuitions derived from this case study are important to consider when interpreting results from all model-based analyses, which are similarly contingent upon the specification of the models used.