Top-down Bias Research Articles

Affordances of distractors and compatibility effects: a study with the computational model TRoPICALS

AbstractSeeing an object activates in the brain both visual and action codes. Crucial evidence supporting this view is offered by compatibility effect experiments (Ellis et al. (2007). J Exp Psychol: Hum Percept Perform): perception of an object can facilitate or interfere with the execution of an action (e.g. grasping) even when the viewer has no intention of interacting with the object. TRoPICALS (Caligiore et al. (2010). Psychol Rev) is a computational model developed to study compatibility effects. It provides a general hypothesis about the brain mechanisms underlying compatibility effects, suggesting that the top-down bias from prefrontal cortex (PFC), and its agreement or disagreement with the affordances of objects, plays a key role in such phenomena. Compatibility effects have been investigated in the presence of a distractor object in (Ellis et al. (2007). J Exp Psychol: Hum Percept Perform). The reaction times (RTs) results confirmed compatibility effects found in previous experiments without the distractor. Interestingly, results also showed an unexpected effect of the distractor: responding to a target with a grip compatible with the size of the distractor produced slower RTs in comparison to the incompatible case. Here we present an enhanced version of TRoPICALS that reproduces and explains these new results. This explanation is based on the idea according to which PFC might play a double role in its top-down guidance of action selection producing: (a) a positive bias in favor of the action requested by the experimental task; (b) a negative bias directed to inhibiting the action evoked by the distractor. The model also provides two testable predictions on the possible consequences on compatibilities effects of the target and distractor objects in Parkinsonian disease patients with damages of inhibitory circuits.

Early event‐related cortical activity originating in the frontal eye fields and inferior parietal lobe predicts the occurrence of correct and error saccades

Although the cortical circuitry underlying saccade execution has well been specified by neurophysiological and functional imaging studies, the temporal dynamics of cortical activity predicting the occurrence of voluntary or reflexive saccades in humans are largely unknown. Here, we examined electrophysiological activity preceding the onset of correct (i.e., voluntary) or error (i.e., reflexive) saccades in an oculomotor capture task. Participants executed saccades to lateralized visual targets while attempting to inhibit reflexive glances to abruptly appearing distracters. Since the visual display was identical for both types of saccades, different electrophysiological patterns preceding correct and error saccades could not be explained by low-level perceptual differences. Compared to correct saccades electrophysiological activity preceding error saccades showed significant differences of the scalp electric field and of voltage amplitudes at posterior electrodes. In addition, though error saccades had significantly shorter latency than correct saccades a prolonged topographic configuration of electric potentials prior to error saccades was found ∼120-140 ms following target onset. In agreement with the known asymmetry in hemispheric dominance for spatial attention, distinct electrophysiological patterns were only found for leftward saccades. While error saccades were associated with stronger activity in the right Frontal Eye Field, correct saccades were preceded by stronger activity in the inferior parietal lobule. These findings suggest that selection of the saccade target in a conflicting situation is determined by early top-down biases originating in frontal and parietal cortical regions critical for spatial attention and saccade programming.

Early top-down attentional modulation in visual processing

An established view of attention is that it acts by biasing the competition between rival representations at various stages in visual processing (e.g., Desimone & Duncan, 1995). The bias brings about a processing advantage for attended items over unattended items, so that attended items have greater neural activity in associated representations, are better perceived, remembered, and eventually are more likely to be acted on. There is extensive evidence that the bias can originate in one of two sources. It can be data driven and result from input properties, in which case more salient items will be biased over less salient ones (e.g., Theeuwes, 1991). Conversely, the bias can be internally driven and result from the current behavioural goals of the observer (e.g., Folk, Remington & Jonhnston, 1992). Both types of bias have the effect of resolving the competition between processing for perceived items. A common view is that for all types of input, and at all processing stages, the competition between rival representations is resolved by means of a combination of bottom-up and top-down biases. In this way, attention optimally leads to selection of items that may be relevant because of their objective appearance (bottom-up selection), while at the same time allowing selection of items that are currently task-relevant (top-down selection).

Training Top-Down Attention Improves Performance on a Triple-Conjunction Search Task

Training has been shown to improve perceptual performance on limited sets of stimuli. However, whether training can generally improve top-down biasing of visual search in a target-nonspecific manner remains unknown. We trained subjects over ten days on a visual search task, challenging them with a novel target (top-down goal) on every trial, while bottom-up uncertainty (distribution of distractors) remained constant. We analyzed the changes in saccade statistics and visual behavior over the course of training by recording eye movements as subjects performed the task. Subjects became experts at this task, with twofold increased performance, decreased fixation duration, and stronger tendency to guide gaze toward items with color and spatial frequency (but not necessarily orientation) that resembled the target, suggesting improved general top-down biasing of search.

Lapsing when sleep deprived: Neural activation characteristics of resistant and vulnerable individuals

Lapses of attention, in the form of delayed responses to salient stimuli, increase in frequency for some but not all persons after sleep deprivation (SD). To identify patterns of task-related brain activation that might explain differences in vulnerability to SD, we performed fMRI on participants during a visual, selective attention task. We analyzed the correct responses in a trial-by-trial fashion to model the effects of response time. Stimulus contrast was varied to modulate perceptual difficulty. Attentional lapses and low-contrast stimuli were independently associated with increased signal in fronto-parietal regions associated with biasing attention. Sleep-deprived vulnerable individuals showed reduced top down fronto-parietal signal across all levels of image contrast and this reduction was particularly significant during lapses. There was concurrent reduction in extrastriate cortex and thalamus activation. Non-vulnerable persons showed a trend towards higher top-down biasing of attention and preserved visual cortex activation during SD lapses. A major contributor to performance degradation in SD appears to be a reduction in top-down biasing of attention that is independent of task difficulty.

Sleep Deprivation Impairs Object-Selective Attention: A View from the Ventral Visual Cortex

BackgroundMost prior studies on selective attention in the setting of total sleep deprivation (SD) have focused on behavior or activation within fronto-parietal cognitive control areas. Here, we evaluated the effects of SD on the top-down biasing of activation of ventral visual cortex and on functional connectivity between cognitive control and other brain regions.Methodology/Principal FindingsTwenty-three healthy young adult volunteers underwent fMRI after a normal night of sleep (RW) and after sleep deprivation in a counterbalanced manner while performing a selective attention task. During this task, pictures of houses or faces were randomly interleaved among scrambled images. Across different blocks, volunteers responded to house but not face pictures, face but not house pictures, or passively viewed pictures without responding. The appearance of task-relevant pictures was unpredictable in this paradigm. SD resulted in less accurate detection of target pictures without affecting the mean false alarm rate or response time. In addition to a reduction of fronto-parietal activation, attending to houses strongly modulated parahippocampal place area (PPA) activation during RW, but this attention-driven biasing of PPA activation was abolished following SD. Additionally, SD resulted in a significant decrement in functional connectivity between the PPA and two cognitive control areas, the left intraparietal sulcus and the left inferior frontal lobe.Conclusions/SignificanceSD impairs selective attention as evidenced by reduced selectivity in PPA activation. Further, reduction in fronto-parietal and ventral visual task-related activation suggests that it also affects sustained attention. Reductions in functional connectivity may be an important additional imaging parameter to consider in characterizing the effects of sleep deprivation on cognition.

An Object-Based Visual Attention Model for Robotic Applications

By extending integrated competition hypothesis, this paper presents an object-based visual attention model, which selects one object of interest using low-dimensional features, resulting that visual perception starts from a fast attentional selection procedure. The proposed attention model involves seven modules: learning of object representations stored in a long-term memory (LTM), preattentive processing, top-down biasing, bottom-up competition, mediation between top-down and bottom-up ways, generation of saliency maps, and perceptual completion processing. It works in two phases: learning phase and attending phase. In the learning phase, the corresponding object representation is trained statistically when one object is attended. A dual-coding object representation consisting of local and global codings is proposed. Intensity, color, and orientation features are used to build the local coding, and a contour feature is employed to constitute the global coding. In the attending phase, the model preattentively segments the visual field into discrete proto-objects using Gestalt rules at first. If a task-specific object is given, the model recalls the corresponding representation from LTM and deduces the task-relevant feature(s) to evaluate top-down biases. The mediation between automatic bottom-up competition and conscious top-down biasing is then performed to yield a location-based saliency map. By combination of location-based saliency within each proto-object, the proto-object-based saliency is evaluated. The most salient proto-object is selected for attention, and it is finally put into the perceptual completion processing module to yield a complete object region. This model has been applied into distinct tasks of robots: detection of task-specific stationary and moving objects. Experimental results under different conditions are shown to validate this model.

The spatial and temporal dynamics of anticipatory preparation and response inhibition in task-switching

We investigated ERP and fMRI correlates of anticipatory preparation and response inhibition in a cued task-switching paradigm with informatively cued, non-informatively cued and no-go trials. Cue-locked ERPs showed evidence for a multicomponent preparation process. An early cue-locked differential positivity was larger for informative vs. non-informative cues and its amplitude correlated with differential activity for informatively vs. non-informatively cued trials in the dorsolateral prefrontal cortex (DLPFC), consistent with a goal activation process. A later differential positivity was larger for informatively cued switch vs. repeat trials and its amplitude correlated with informatively cued switch vs. repeat activity in the posterior parietal cortex (PPC), compatible with a category-response (C-R) rule activation process. No-go trials elicited a frontal P3, whose amplitude was negatively correlated with activity in the ventrolateral prefrontal cortex (VLPFC) and basal ganglia motor network, suggesting that a network responsible for response execution was inhibited in the course of a no-go trial. These findings indicate that anticipatory preparation in task-switching is comprised of at least two processes: goal activation and C-R rule activation. They also support a functional dissociation between DLPFC and VLPFC, with the former involved in top-down biasing and the latter involved in response inhibition.

TRoPICALS: A computational embodied neuroscience model of compatibility effects.

Perceiving objects activates the representation of their affordances. For example, experiments on compatibility effects showed that categorizing objects by producing certain handgrips (power or precision) is faster if the requested responses are compatible with the affordance elicited by the size of objects (e.g., small or large). The article presents a neural-network architecture that provides a general framework to account for compatibility effects. The model was designed with a methodological approach (computational embodied neuroscience) that aims to provide increasingly general accounts of brain and behavior (4 sources of constraints are used: neuroscientific data, behavioral data, embodied systems, reproduction of learning processes). The model is based on 4 principles of brain organization that we claim underlie most compatibility effects. First, visual perception and action are organized in the brain along a dorsal neural pathway encoding affordances and a ventral pathway encoding goals. Second, the prefrontal cortex within the ventral pathway gives a top-down bias to action selection by integrating information on stimuli, context, and goals. Third, reaction times depend on dynamic neural competitions for action selection that integrate bottom-up and top-down information. The congruence or incongruence between affordances and goals explains the different reaction times found in the experiments. Fourth, as words trigger internal simulations of their referents, they can cause compatibility effects as objects do. We validated the model by reproducing and explaining 3 types of compatibility effects and showed its heuristic power by producing 2 testable predictions. We also assessed the explicative power of the model by comparing it with related models and showed how it can be extended to account for other compatibility effects.

Effects of anticipatory perceptual simulation on practiced human-robot tasks

With the aim of attaining increased fluency and efficiency in human-robot teams, we have developed a cognitive architecture for robotic teammates based on the neuro-psychological principles of anticipation and perceptual simulation through top-down biasing. An instantiation of this architecture was implemented on a non-anthropomorphic robotic lamp, performing a repetitive human-robot collaborative task. In a human-subject study in which the robot works on a joint task with untrained subjects, we find our approach to be significantly more efficient and fluent than in a comparable system without anticipatory perceptual simulation. We also show the robot and the human to improve their relative contribution at a similar rate, possibly playing a part in the human's "like-me" perception of the robot. In self-report, we find significant differences between the two conditions in the sense of team fluency, the team's improvement over time, the robot's contribution to the efficiency and fluency, the robot's intelligence, and in the robot's adaptation to the task. We also find differences in verbal attitudes towards the robot: most notably, subjects working with the anticipatory robot attribute more human qualities to the robot, such as gender and intelligence, as well as credit for success, but we also find increased self-blame and self-deprecation in these subjects' responses. We believe that this work lays the foundation towards modeling and evaluating artificial practice for robots working in collaboration with humans.

Crossmodal links between vision and touch in spatial attention: a computational modelling study.

Many studies have revealed that attention operates across different sensory modalities, to facilitate the selection of relevant information in the multimodal situations of every-day life. Cross-modal links have been observed either when attention is directed voluntarily (endogenous) or involuntarily (exogenous). The neural basis of cross-modal attention presents a significant challenge to cognitive neuroscience. Here, we used a neural network model to elucidate the neural correlates of visual-tactile interactions in exogenous and endogenous attention. The model includes two unimodal (visual and tactile) areas connected with a bimodal area in each hemisphere and a competition between the two hemispheres. The model is able to explain cross-modal facilitation both in exogenous and endogenous attention, ascribing it to an advantaged activation of the bimodal area on the attended side (via a top-down or bottom-up biasing), with concomitant inhibition towards the opposite side. The model suggests that a competitive/cooperative interaction with biased competition may mediate both forms of cross-modal attention.

Distinct control networks for cognition and emotion in the prefrontal cortex

The activation of dorsolateral prefrontal cortex (dlPFC) has been suggested to reflect the engagement of a control mechanism for top-down biasing of context processing in resource-demanding memory tasks. Here we tested the hypothesis that the dlPFC subserves a similar function also in attention and emotion tasks. 18 healthy young adults were tested in a functional magnetic resonance imaging (fMRI) study where the demands for context processing were manipulated in three different cognitive domains: auditory attention, episodic retrieval, and emotion regulation. We found that the right dlPFC was jointly sensitive to increased cognitive demands in the attention and memory tasks. By contrast, increased demands in the emotion task (reappraisal) were associated with increased activity in ventromedial PFC along with decreased amygdala activity. Our findings of divergent prefrontal control networks for cognitive and emotional control extend previous separations of cognition and emotion in the anterior cingulate cortex.

Activity in the superior colliculus reflects dynamic interactions between voluntary and involuntary influences on orienting behaviour

Performance in a behavioural task can be influenced by both bottom-up and top-down processes such as stimulus modality and prior probability. Here, we exploited differences in behavioural strategy to explore the role of the intermediate and deep layers of the superior colliculus (dSC) in covert orienting. Two monkeys were trained on a predictive cued-saccade task in which the cue predicted the target's upcoming location with 80% validity. When the delay between cue and target onset was 250 ms, both monkeys showed faster responses to the uncued (Invalid) location. This was associated with a reduced target-aligned response in the dSC on Valid trials for both monkeys and is consistent with a bottom-up (i.e. involuntary) bias. When the delay was increased to 650 ms, one monkey continued to show faster responses to the Invalid location whereas the other monkey showed faster responses to the Valid location, consistent with a top-down (i.e. voluntary) bias. This latter behaviour was correlated with an increase in activity in dSC neurons preceding target onset that was absent in the other monkey. Thus, using the information provided by the cue shifted the emphasis towards top-down processing, while ignoring this information allowed bottom-up processing to continue to dominate. Regardless of the selected strategy, however, neurons in the dSC consistently reflected the current bias between the two processes, emphasizing its role in both the bottom-up and top-down control of orienting behaviour.

Neural Integration of Top-Down Spatial and Feature-Based Information in Visual Search

Visual search is aided by previous knowledge regarding distinguishing features and probable locations of a sought-after target. However, how the human brain represents and integrates concurrent feature-based and spatial expectancies to guide visual search is currently not well understood. Specifically, it is not clear whether spatial and feature-based search information is initially represented in anatomically segregated regions, nor at which level of processing expectancies regarding target features and locations may be integrated. To address these questions, we independently and parametrically varied the degree of spatial and feature-based (color) cue information concerning the identity of an upcoming visual search target while recording blood oxygenation level-dependent (BOLD) responses in human subjects. Search performance improved with the amount of spatial and feature-based cue information, and cue-related BOLD responses showed that, during preparation for visual search, spatial and feature cue information were represented additively in shared frontal, parietal, and cingulate regions. These data show that representations of spatial and feature-based search information are integrated in source regions of top-down biasing and oculomotor planning before search onset. The purpose of this anticipatory integration could lie with the generation of a "top-down salience map," a search template of primed target locations and features. Our results show that this role may be served by the intraparietal sulcus, which additively integrated a spatially specific activation gain in relation to spatial cue information with a spatially global activation gain in relation to feature cue information.

Filtering items of mass distraction: Top-down biases against distractors are necessary for the feature-based carry-over to occur

In preview search a new target is difficult to detect if it carries a feature shared with the old distractors [Braithwaite, J. J., Humphreys, G. W., & Hodsoll, J. (2003). Color grouping in space and time: Evidence from negative color-based carry-over effects in preview search. Journal of Experimental Psychology: Human Perception and Performance, 29(4), 758–778.] Two experiments are presented which examined whether this negative color carry-over effect is dependent on an attentional-set to ignore old, irrelevant distractors. Consistent with this, the data show that the negative carry-over effect is greatly reduced if the attentional-set to ignore the old preview items is removed and replaced by a set to prioritize the old items instead. The findings demonstrate that preview search, and the carry-over effect, are at least partly determined by a top-down intentional bias against old, irrelevant information.

Top-Down Control-Signal Dynamics in Anterior Cingulate and Prefrontal Cortex Neurons following Task Switching

The prefrontal cortex (PFC) and anterior cingulate cortex (ACC) have both been implicated in cognitive control, but their relative roles remain unclear. Here we recorded the activity of single neurons in both areas while monkeys performed a task that required them to switch between trials in which they had to look toward a flashed stimulus (prosaccades) and trials in which they had to look away from the stimulus (antisaccades). We found that ACC neurons had a higher level of task selectivity than PFC neurons during the preparatory period on trials immediately following a task switch. In ACC neurons, task selectivity was strongest after the task switch and declined throughout the task block, whereas task selectivity remained constant in the PFC. These results demonstrate that the ACC is recruited when cognitive demands increase and suggest a role for both areas in task maintenance and the implementation of top-down control.

A Model for Perceptual Averaging and Stochastic Bistable Behavior and the Role of Voluntary Control

We combine population coding, winner-take-all competition, and differentiated inhibitory feedback to model the process by which information from different, continuously variable signals is integrated for perceptual awareness. We focus on "slant rivalry," where binocular disparity is in conflict with monocular perspective in specifying surface slant. Using a robust single parameter set, our model successfully replicates three key experimental results: (1) transition from signal averaging to bistability with increasing signal conflict, (2) change in perceptual reversal rates as a function of signal conflict, and (3) a shift in the distribution of percept durations through voluntary control exertion. Voluntary control is implemented through the use of a single top-down bias input. The transition from signal averaging to bistability arises as a natural consequence of combining population coding and wide receptive fields, common to higher cortical areas. The model architecture does not contain any assumption that would limit it to this particular example of stimulus rivalry. An emergent physiological interpretation is that differentiated inhibitory feedback may play an important role for increasing percept stability without reducing sensitivity to large stimulus changes, which for bistable conditions leads to increased alternation rate as a function of signal conflict.

Understanding words in context: The role of Broca's area in word comprehension

What role does meaning selection play in word comprehension, and what neural systems support this selection process? Most words have multiple meanings and are therefore ambiguous. This is true of both homonymous words (words that have multiple unrelated meanings) and polysemous words (words that have multiple related meanings). The extant evidence indicates that meaning selection is an integral part of homonym comprehension. However, it is not known whether meaning selection extends to polysemous words, or what neural systems support meaning selection during comprehension. Prior neuroimaging and neuropsychological evidence suggest that the left inferior frontal gyrus (LIFG) may play a role in resolving competition during language processing. We therefore sought to test the hypotheses that meaning selection is part of polysemous word comprehension, and that the LIFG resolves meaning competition during word comprehension. We tested healthy participants on a version of the triplet lexical decision task, with polysemous and homonymous stimuli. Results suggest that the meanings of polysemous words, like the meanings of homonyms, are selected based on context. However, homonymous and polysemous words differed in how meaning frequency affected meaning selection. We then administered the triplet lexical decision task to patients with LIFG damage to examine whether this region plays a role in context-dependent meaning selection. Results support the hypothesis that the LIFG serves as a top-down biasing mechanism that facilitates rapid meaning selection during word comprehension. We conclude that context-dependent meaning selection is an integral part of word comprehension for both homonyms and polysemous words, and that the LIFG facilitates this selection process.

Why do beliefs about intelligence influence learning success? A social cognitive neuroscience model

Students' beliefs and goals can powerfully influence their learning success. Those who believe intelligence is a fixed entity (entity theorists) tend to emphasize 'performance goals,' leaving them vulnerable to negative feedback and likely to disengage from challenging learning opportunities. In contrast, students who believe intelligence is malleable (incremental theorists) tend to emphasize 'learning goals' and rebound better from occasional failures. Guided by cognitive neuroscience models of top-down, goal-directed behavior, we use event-related potentials (ERPs) to understand how these beliefs influence attention to information associated with successful error correction. Focusing on waveforms associated with conflict detection and error correction in a test of general knowledge, we found evidence indicating that entity theorists oriented differently toward negative performance feedback, as indicated by an enhanced anterior frontal P3 that was also positively correlated with concerns about proving ability relative to others. Yet, following negative feedback, entity theorists demonstrated less sustained memory-related activity (left temporal negativity) to corrective information, suggesting reduced effortful conceptual encoding of this material-a strategic approach that may have contributed to their reduced error correction on a subsequent surprise retest. These results suggest that beliefs can influence learning success through top-down biasing of attention and conceptual processing toward goal-congruent information.

The neurodynamics of visual search

We review different functions in visual perception associated with attention and memory that have been integrated by a model based on the biased competition hypothesis. The model integrates, in a unifying form, the explanation of several existing types of experimental data obtained at different levels of investigation. At the microscopic level, single cell recordings are simulated. At the mesoscopic level of cortical areas, results of functional magnetic resonance imaging (fMRI) studies are reproduced. Finally, at the macroscopic level, the outcome of psychophysical experiments like visual search tasks are also described by the model. In particular, the model directly addresses how bottom-up and top-down processes interact in visual cognition, and shows how some apparently serial processes reflect the operation of interacting parallel distributed systems. Attentional top-down bias guides the dynamics to focus attention at a given spatial location or on given features.