Bottom-up Salience Research Articles

Biased choice and incentive salience: Implications for addiction.

Before we can make any choice, we must gather information from the environment about what our options are. This information-gathering process is critically mediated by attention, and our attention is, in turn, shaped by our previous experiences with-and learning about-stimuli and their consequences. In this review, we highlight studies demonstrating a rapid and automatic influence of reward learning on attentional capture and argue that these findings provide a human analog of sign-tracking behavior observed in nonhuman animals-wherein signals of reward gain incentive salience and become attractive targets for attention (and overt behavior) in their own right. We then consider the implications of this idea for understanding the drivers of cue-controlled behavior, with focus on addiction as a case in which choices with regard to reward-related stimuli can become injurious to health. We argue that motivated behavior in general-and addiction in particular-can be understood within a "biased competition" framework: Different options and outcomes compete for attentional priority as a function of top-down goals, bottom-up salience, and prior experience, and the winner of this competition becomes the target for subsequent outcome-directed and flexible behavior. Finally, we outline the implications of the biased-competition framework for cognitive, behavioral, and socioeconomic interventions for addiction. (PsycInfo Database Record (c) 2024 APA, all rights reserved).

Evaluating theories of neural information integration during visual search.

The brain routes and integrates information from many sources during behavior. A number of models explain this phenomenon within the framework of mixed selectivity theory, yet it is difficult to compare their predictions to understand how neurons and circuits integrate information. In this work, we apply time-series partial information decomposition [PID] to compare models of integration on a dataset of superior colliculus [SC] recordings collected during a multi-target visual search task. On this task, SC must integrate target guidance, bottom-up salience, and previous fixation signals to drive attention. We find evidence that SC neurons integrate these factors in diverse ways, including decision-variable selectivity to expected value, functional specialization to previous fixation, and code-switching (to incorporate new visual input).

An interactive cortical architecture for perceptual organization by accentuation

Accentuation has been proposed as a general principle of perceptual organization. Here, we have developed a neurodynamic architecture to explain how accentuation affects boundary segmentation and shape perception. The model consists of bottom-up and top-down pathways. Bottom-up processing involves a set of feature maps that compute bottom-up salience of surfaces, boundaries, boundary completions, and junctions. Then, a feature-based winner-take-all network selects the most salient locations. Top-down processing includes an object-based attention stage that allows enhanced neural activity to propagate from the most salient locations to all connected locations, and a visual segmentation stage that employs inhibitory connections to segregate boundaries into distinct maps. The model was tested on a series of computer simulations showing how the position of the accent affects boundary segregation in the square-diamond and the pointing illusion. The model was also tested on a variety of texture segregation tasks, showing that its performance was comparable to that of human observers. The model suggests that there is an intermediate stage of visual processing between perceptual grouping and object recognition that helps the visual system choose between competing percepts of the ambiguous stimulus.

The role of salience in the suppression of distracting stimuli.

Researchers have long debated whether salient distractors have the power to automatically capture attention. Recent research has suggested a potential resolution, called the signal suppression hypothesis, whereby salient distractors produce a bottom-up salience signal, but can be suppressed to prevent visual distraction. This account, however, has been criticized on the grounds that previous studies may have used distractors that were only weakly salient. This claim has been difficult to empirically test because there are currently no well-established measures of salience. The current study addresses this by introducing a psychophysical technique to measure salience. First, we generated displays that aimed to manipulate the salience of two color singletons via color contrast. We then verified that this manipulation was successful using a psychophysical technique to determine the minimum exposure duration required to detect each color singleton. The key finding was that high-contrast singletons were detected at briefer exposure thresholds than low-contrast singletons, suggesting that high-contrast singletons were more salient. Next, we evaluated the participants' ability to ignore these singletons in a task in which they were task irrelevant. The results showed that, if anything, high-salience singletons were more strongly suppressed than low-salience singletons. These results generally support the signal suppression hypothesis and refute claims that highly salient singletons cannot be ignored.

What Do We Know About Threat-Related Perceptual Decision Making?

The ability to make rapid and precise decisions regarding the presence or absence of threats in the environment is critical for survival. Although threatening stimuli may be detected more accurately and faster because of to the bottom-up salience of their features, in the real world, these stimuli are often encountered in familiar environments in which top-down cues signal their arrival. There has been significant progress in understanding of the mechanisms by which people make perceptual decisions regarding relatively routine stimuli; however, the mechanisms of threat-related perceptual decision making remain unclear. In this review, we discuss the psychological, computational, and neural mechanisms by which information from threatening stimuli is integrated with prior knowledge from cues and surrounding contexts to guide perceptual decision making.

Interaction of the salience network, ventral attention network, dorsal attention network and default mode network in neonates and early development of the bottom-up attention system

The salience network (SN), ventral attention network (VAN), dorsal attention network (DAN) and default mode network (DMN) have shown significant interactions and overlapping functions in bottom-up and top-down mechanisms of attention. In the present study, we tested if the SN, VAN, DAN and DMN connectivity can infer the gestational age (GA) at birth in a study group of 88 healthy neonates, scanned at 40weeks of post-menstrual age, and with GA at birth ranging from 28 to 40weeks. We also ascertained whether the connectivity within each of the SN, VAN, DAN and DMN was able to infer the average functional connectivity of the others. The ability to infer GA at birth or another network's connectivity was evaluated using a multivariate data-driven framework. The VAN, DAN and the DMN inferred the GA at birth (p < 0.05). The SN, DMN and VAN were able to infer the average connectivity of the other networks (p < 0.05). Mediation analysis between VAN's and DAN's inference on GA at birth found reciprocal transmittance of change with GA at birth of VAN's and DAN's connectivity (p < 0.05). Our findings suggest that the VAN has a prominent role in bottom-up salience detection in early infancy and that the role of the VAN and the SN may overlap in the bottom-up control of attention.

Temporal preparation accelerates spatial selection by facilitating bottom-up processing

Temporal preparation facilitates spatial selection in visual search. This selection benefit has not only been observed for targets, but also for task-irrelevant, salient distractors. This result suggests that temporal preparation influences bottom-up salience in spatial selection. To test this assumption, we conducted an event-related-potential (ERP) study in which we measured the joint effect of temporal preparation and target salience on the N2pc as an index of spatial selection and the N1 as an index of perceptual discrimination. To manipulate target salience, we employed two different setsizes (i.e., a small or large number of homogeneous distractors). To manipulate temporal preparation, we presented a warning signal before the search display and we varied the length of the interval (foreperiod) between warning signal and search display in different blocks of trials (constant foreperiod paradigm). Replicating previous results, we observed that the N1 and the N2pc arose earlier in case of good temporal preparation. Importantly, the beneficial effect on the N2pc onset latency was stronger when the target salience was initially low (i.e., small setsize). This result provides evidence that temporal preparation influences bottom-up processing and, thereby, facilitates spatial selection.

Progress and remaining issues: A response to the commentaries on Luck et al. (2021)

ABSTRACT Luck et al. (2021) reviewed evidence that observers can learn to suppress attentional capture by salient distractors. Several commentaries were written in response to this review paper, many of which raised important and interesting issues. Here, we respond to these commentaries. Although there has been substantial progress in the attentional capture debate, there are still remaining issues that need to be addressed before the debate is completely resolved. Specifically, we summarize the need for an independent measure of bottom-up salience and better metrics of how attentional control unfolds over time. Ultimately, the field may need a more refined theoretical model of visual attention that distinguishes between attentional priority, attentional orienting, and attentional engagement.

Salience effects in information acquisition: No evidence for a top-down coherence influence

The Integrated Coherence-Based Decision and Search (iCodes) model proposed by Jekel et al. (Psychological Review, 125 (5), 744–768, 2018) formalizes both decision making and pre-decisional information search as coherence-maximization processes in an interactive network. Next to bottom-up attribute influences, the coherence of option information exerts a top-down influence on the search processes in this model, predicting the tendency to continue information search with the currently most attractive option. This hallmark “attraction search effect” (ASE) has been demonstrated in several studies. In three experiments with 250 participants altogether, a more subtle prediction of an extended version of iCodes including exogenous influence factors was tested: The salience of information is assumed to have both a direct (bottom-up) and an indirect (top-down) effect on search, the latter driven by the match between information valence and option attractiveness. The results of the experiments largely agree in (1) showing a strong ASE, (2) demonstrating a bottom-up salience effect on search, but (3) suggesting the absence of the hypothesized indirect top-down salience effect. Hence, only two of three model predictions were confirmed. Implications for various implementations of exogenous factors in the iCodes model are discussed.

Salience determines attentional orienting in visual selection.

Recently the signal-suppression account was proposed, positing that salient stimuli automatically produce a bottom-up salience signal that can be suppressed via top-down control processes. Evidence for this hybrid account came from a capture-probe paradigm that showed that while searching for a specific shape, observers suppressed the location of the irrelevant color singleton. Here we replicate these findings but also show that this occurs only for search arrays with 4 elements. For larger array sizes when both target and distractor singleton are salient, there is no evidence for suppression; instead and consistent with the stimulus-driven account, there is clear evidence that the salient distractor captured attention. The current study shows that the relative salience of items in the display is a crucial factor in attentional control. In displays with a few heterogeneous items, top-down suppression is possible. However, in larger displays in which both target and distractor singletons are salient, no top-down suppression is observed. We conclude that the signal-suppression account cannot resolve the long-standing debate regarding stimulus-driven and goal-driven attentional capture. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

Top-down control of saccades requires inhibition of suddenly appearing stimuli

Humans scan their visual environment using saccade eye movements. Where we look is influenced by bottom-up salience and top-down factors, like value. For reactive saccades in response to suddenly appearing stimuli, it has been shown that short-latency saccades are biased towards salience, and that top-down control increases with increasing latency. Here, we show, in a series of six experiments, that this transition towards top-down control is not determined by the time it takes to integrate value information into the saccade plan, but by the time it takes to inhibit suddenly appearing salient stimuli. Participants made consecutive saccades to three fixation crosses and a vertical bar consisting of a high-salient and a rewarded low-salient region. Endpoints on the bar were biased towards salience whenever it appeared or reappeared shortly before the last saccade was initiated. This was also true when the eye movement was already planned. When the location of the suddenly appearing salient region was predictable, saccades were aimed in the opposite direction to nullify this sudden onset effect. Successfully inhibiting salience, however, could only be achieved by previewing the target. These findings highlight the importance of inhibition for top-down eye-movement control.

Antisaccade Deficits in Schizophrenia Can Be Driven by Attentional Relevance of the Stimuli.

The antisaccade task is considered a test of cognitive control because it creates a conflict between the strong bottom-up signal produced by the cue and the top-down goal of shifting gaze to the opposite side of the display. Antisaccade deficits in schizophrenia are thought to reflect impaired top-down inhibition of the prepotent bottom-up response to the cue. However, the cue is also a highly task-relevant stimulus that must be covertly attended to determine where to shift gaze. We tested the hypothesis that difficulty in overcoming the attentional relevance of the cue, rather than its bottom-up salience, is key in producing impaired performance in people with schizophrenia (PSZ). We implemented 3 versions of the antisaccade task in which we varied the bottom-up salience of the cue while holding its attentional relevance constant. We found that difficulty in performing a given antisaccade task-relative to a prosaccade version using the same stimuli-was largely independent of the cue's bottom-up salience. The magnitude of impairment in PSZ relative to control subjects was also independent of bottom-up salience. The greatest impairment was observed in a version where the cue lacked bottom-up salience advantage over other locations. These results indicate that the antisaccade deficit in PSZ does not reflect an impairment in overcoming bottom-up salience of the cue, but PSZ are instead impaired at overcoming its attentional relevance. This deficit may still indicate an underlying inhibitory control impairment but could also reflect a hyperfocusing of attentional resources on the cue.

The link between auditory salience and emotion intensity

ABSTRACT To ensure that listeners pay attention and do not habituate, emotionally intense vocalizations may be under evolutionary pressure to exploit processing biases in the auditory system by maximising their bottom-up salience. This “salience code” hypothesis was tested using 128 human nonverbal vocalizations representing eight emotions: amusement, anger, disgust, effort, fear, pain, pleasure, and sadness. As expected, within each emotion category salience ratings derived from pairwise comparisons strongly correlated with perceived emotion intensity. For example, while laughs as a class were less salient than screams of fear, salience scores almost perfectly explained the perceived intensity of both amusement and fear considered separately. Validating self-rated salience evaluations, high- vs. low-salience sounds caused 25% more recall errors in a short-term memory task, whereas emotion intensity had no independent effect on recall errors. Furthermore, the acoustic characteristics of salient vocalizations were similar to those previously described for non-emotional sounds (greater duration and intensity, high pitch, bright timbre, rapid modulations, and variable spectral characteristics), confirming that vocalizations were not salient merely because of their emotional content. The acoustic code in nonverbal communication is thus aligned with sensory biases, offering a general explanation for some non-arbitrary properties of human and animal high-arousal vocalizations.

The mechanisms of sensory sensitivity: A response to commentaries on Ward (2019)

ABSTRACT A recent framework has been proposed for understanding individual differences in sensory sensitivity in terms of whether increased neural responsiveness to sensory stimuli reflects increased responses to sensory signals or increased neural noise. This has different implications for behavioral performance. In this commentary, we discuss some related concepts including responsivity and perceptual expertise. We consider in more detail the role of attention including exogenous factors (altered bottom-up salience) and endogenous ones (indiscriminately assigning higher attentional priority), and the role that emotion may play in this process.

Healthy women with severe early life trauma show altered neural facilitation of emotion inhibition under acute stress.

Across psychopathologies, trauma-exposed individuals suffer from difficulties in inhibiting emotions and regulating attention. In trauma-exposed individuals without psychopathology, only subtle alterations of neural activity involved in regulating emotions have been reported. It remains unclear how these neural systems react to demanding environments, when acute (non-traumatic but ordinary) stress serves to perturbate the system. Moreover, associations with subthreshold clinical symptoms are poorly understood. The present fMRI study investigated response inhibition of emotional faces before and after psychosocial stress situations. Specifically, it compared 25 women (mean age 31.5 ± 9.7 years) who had suffered severe early life trauma but who did not have a history of or current psychiatric disorder, with 25 age- and education-matched trauma-naïve women. Under stress, response inhibition related to fearful faces was reduced in both groups. Compared to controls, trauma-exposed women showed decreased left inferior frontal gyrus (IFG) activation under stress when inhibiting responses to fearful faces, while activation of the right anterior insula was slightly increased. Also, groups differed in brain-behaviour correlations. Whereas stress-induced false alarm rates on fearful stimuli negatively correlated with stress-induced IFG signal in controls, in trauma-exposed participants, they positively correlated with stress-induced insula activation. Neural facilitation of emotion inhibition during stress appears to be altered in trauma-exposed women, even without a history of or current psychopathology. Decreased activation of the IFG in concert with heightened bottom-up salience of fear related cues may increase vulnerability to stress-related diseases.

Failures in top-down control in schizophrenia revealed by patterns of saccadic eye movements.

Successful execution of many behavioral goals relies on well-organized patterns of saccadic eye movements, and in complex tasks, these patterns can reveal the component processes underlying task performance. The present study examined the pattern of eye movements in a visual search task to provide evidence of attentional control impairments in people with schizophrenia (PSZ). We tested PSZ(N = 38) and nonpsychiatric control subjects (NCS, N = 35) in a task that was designed to stress top-down control by pitting task goals against bottom-up salience. Participants searched for either a low-contrast (nonsalient) or a high-contrast (salient) target among low- and high-contrast distractors. By examining fixations of the low- and high-contrast items, we evaluated the ability of PSZ and NCS to focus on low-salience targets and filter out high-salience distractors (or vice versa). When participants searched for a salient target, both groups successfully focused on relevant, high-contrast stimuli and filtered out target-mismatched, low-contrast stimuli. However, when searching for a nonsalient target, PSZ were impaired at efficiently suppressing high-contrast (salient) distractors. Specifically, PSZ were more likely than NCS to fixate and revisit salient distractors, and they dwelled on these items longer than did NCS. The results provide direct evidence that PSZ are impaired in their ability to utilize top-down goals to overcome the prepotent tendency to focus attention on irrelevant but highly salient information. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

Using Visual Salience in Empirical Game Theory

“Bottom-up” visual perception is fast, universal, and does not depend on personal goals or experience. Where exactly bottom-up attention is directed in a visual image-- usually based on contrast, color, feature orientation, and centrality-- can be accurately predicted by machine-learned algorithms. We test whether this type of salience can help explain economic decisions in four experimental analyses (three are new). When people pick between sets of valued fruits, salience can lead to mistakes (lower-value sets are sometimes chosen because they are salient). Two studies show strong and weak influence on choices, respectively, of salience in stock price charts and salient payoffs in game matrices. The central analysis is evidence from games in which choices are locations in images. When players are trying to cooperatively match locations, concentration of salience is associated with the success of matching (r=.57). In competitive hider-seeker location games, all players choose salient locations more often. This fact creates a “seeker’s advantage” (9% wins compared to the 7% predicted by unique equilibrium). The location game data are consistent with cognitive hierarchy and level-k models in which predicted salience influences nonstrategic level-0 choices. Model estimates from hider-seeker games can predict behavior in matching games despite their opposite strategic structures. Looking forward, bottom-up salience could be included as a predictable type of costless involuntary “pre-attention” in expanded models of Bayesian-rational inattention. Novel applications involving prices and taxes, disability, ethnic markers, and visually-influenced beliefs are suggested.

The impact of reward on attention in schizophrenia

Traditionally, attention was thought to be directed by either top-down goals or bottom-up salience. Recent studies have shown that the reward history of a stimulus feature also acts as a powerful attentional cue. This is particularly relevant in schizophrenia, which is characterized by motivational and attentional deficits. Here, we examine the impact of reward on selective attention.Forty-eight people with schizophrenia (PSZ) and 34 non-psychiatric control subject (NCS) discriminated the location of a target dot appearing inside a left circle or right circle. The circles were different colors, one of which was associated with reward via pre-training. In the first 2 blocks, targets were equally likely to appear in the left or right circle. In the last 4 blocks, the target was 75% likely on one side, thus allowing us to separately examine how attention was impacted by reward (color) and probability (location).PSZ had slower overall reaction times (RTs) than NCS. Both groups showed robust effects of spatial probability and reward history, with faster RTs for the rewarded color and for the more probable location. These effects were similar in PSZ and NCS. Negative symptom severity correlated with overall RT slowing, but there were no correlations between symptoms and reward-associated biasing of attention.PSZ demonstrated RT slowing but normal reward history and spatial probability-driven RT facilitation. These results are conceptually similar to prior findings showing intact implicit reward effects on response bias, and suggest that implicit processing of reward and probability is intact in PSZ.

A Generative Model of Cognitive State from Task and Eye Movements.

The early eye tracking studies of Yarbus provided descriptive evidence that an observer's task influences patterns of eye movements, leading to the tantalizing prospect that an observer's intentions could be inferred from their saccade behavior. We investigate the predictive value of task and eye movement properties by creating a computational cognitive model of saccade selection based on instructed task and internal cognitive state using a Dynamic Bayesian Network (DBN). Understanding how humans generate saccades under different conditions and cognitive sets links recent work on salience models of low-level vision with higher level cognitive goals. This model provides a Bayesian, cognitive approach to top-down transitions in attentional set in pre-frontal areas along with vector-based saccade generation from the superior colliculus. Our approach is to begin with eye movement data that has previously been shown to differ across task. We first present an analysis of the extent to which individual saccadic features are diagnostic of an observer's task. Second, we use those features to infer an underlying cognitive state that potentially differs from the instructed task. Finally, we demonstrate how changes of cognitive state over time can be incorporated into a generative model of eye movement vectors without resorting to an external decision homunculus. Internal cognitive state frees the model from the assumption that instructed task is the only factor influencing observers' saccadic behavior. While the inclusion of hidden temporal state does not improve the classification accuracy of the model, it does allow accurate prediction of saccadic sequence results observed in search paradigms. Given the generative nature of this model, it is capable of saccadic simulation in real time. We demonstrated that the properties from its generated saccadic vectors closely match those of human observers given a particular task and cognitive state. Many current models of vision focus entirely on bottom-up salience to produce estimates of spatial "areas of interest" within a visual scene. While a few recent models do add top-down knowledge and task information, we believe our contribution is important in three key ways. First, we incorporate task as learned attentional sets that are capable of self-transition given only information available to the visual system. This matches influential theories of bias signals by (Miller and Cohen Annu Rev Neurosci 24:167-202, 2001) and implements selection of state without simply shifting the decision to an external homunculus. Second, our model is generative and capable of predicting sequence artifacts in saccade generation like those found in visual search. Third, our model generates relative saccadic vector information as opposed to absolute spatial coordinates. This matches more closely the internal saccadic representations as they are generated in the superior colliculus.

Working memory accuracy for multiple targets is driven by reward expectation and stimulus contrast with different time-courses

The richness of sensory input dictates that the brain must prioritize and select information for further processing and storage in working memory. Stimulus salience and reward expectations influence this prioritization but their relative contributions and underlying mechanisms are poorly understood. Here we investigate how the quality of working memory for multiple stimuli is determined by priority during encoding and later memory phases. Selective attention could, for instance, act as the primary gating mechanism when stimuli are still visible. Alternatively, observers might still be able to shift priorities across memories during maintenance or retrieval. To distinguish between these possibilities, we investigated how and when reward cues determine working memory accuracy and found that they were only effective during memory encoding. Previously learned, but currently non-predictive, color-reward associations had a similar influence, which gradually weakened without reinforcement. Finally, we show that bottom-up salience, manipulated through varying stimulus contrast, influences memory accuracy during encoding with a fundamentally different time-course than top-down reward cues. While reward-based effects required long stimulus presentation, the influence of contrast was strongest with brief presentations. Our results demonstrate how memory resources are distributed over memory targets and implicates selective attention as a main gating mechanism between sensory and memory systems.