Semantic access occurs outside of awareness for the ground side of a figure

The hidden arrow in the FedEx logo: Do we really unconsciously "see" it?

The FedEx logo makes clever use of figure-ground ambiguity to create an "invisible" arrow in the background space between "E" and "x". Most designers believe the hidden arrow can convey an unconscious impression of speed and precision about the FedEx brand, which may influence subsequent behavior. To test this assumption, we designed similar images with hidden arrows to serve as endogenous (but camouflaged) directional cues in a Posner's orienting task, where a cueing effect would suggest subliminal processing of the hidden arrow. Overall, we observed no cue congruency effect, unless the arrow is explicitly highlighted (Experiment 4). However, there was a general effect of prior knowledge: when people were under pressure to suppress background information, those who knew about the arrow could do so faster in all congruence conditions (i.e., neutral, congruent, incongruent), although they fail to report seeing the arrow during the experiment. This was true in participants from North America who had heard of the FedEx arrow before (Experiment 1 & 3), and also in our Taiwanese sample who were just informed of such design (Experiment 2). These results can be well explained by the Biased Competition Model in figure-ground research, and together suggest: (1) people do not unconsciously perceive the FedEx arrow, at least not enough to exhibit a cueing effect in attention, but (2) knowing about the arrow can fundamentally change the way we visually process these negative-space logos in the future, making people react faster to images with negative space regardless of the hidden content.

BOLD activation on the groundside of figures: More suppression of grounds that competed more for figural status

A fundamental aspect of object detection is assigning a border to one (figure) side but not the other (ground) side. Figures are shaped; grounds appear shapeless near the figure border. Accumulating evidence supports the view that the mechanism of figure assignment is inhibitory competition with the figure perceived on the winning side. Suppression has been observed on the groundside of figure borders. One prediction is that more suppression will be observed when the groundside competes more for figural status. We tested this prediction by assessing BOLD activation on the groundside of two types of stimuli with articulated borders: AE_nov and AE_fam stimuli. In both stimulus types, multiple image-based priors (symmetry, closure, small area, enclosure by a larger region) favored the inside as the figure. In AE_fam but not AE_nov stimuli, the figural prior of familiar configuration present on the outside competes for figural status. Observers perceived the insides of both types of stimuli as novel figures and the outsides as shapeless grounds. Previously, we observed lower BOLD activation in early visual areas representing the grounds of AE_fam than AE_nov stimuli, although unexpectedly, activation was above baseline. With articulated borders, it can be difficult to exclude figure activation from ground ROIs. Here, our ground ROIs better excluded figure activation; we also added straight-edge (SE) control stimuli and increased the sample size. In early visual areas representing the grounds, we observed lower BOLD activation on the groundside of AE_fam than AE_nov stimuli and below-baseline BOLD activation on the groundside of SE and AE_fam stimuli. These results, indicating that greater suppression is applied to groundsides that competed more for figural status but lost the competition, support a Bayesian model of figure assignment in which proto-objects activated at both low and high levels where image features and familiar configurations are represented, respectively, compete for figural status.

Scene-selective brain regions respond to embedded objects of a scene

Objects are fundamental to scene understanding. Scenes are defined by embedded objects and how we interact with them. Paradoxically, scene processing in the brain is typically discussed in contrast to object processing. Using the BOLD5000 dataset (Chang et al., 2019), we examined whether objects within a scene predicted the neural representation of scenes, as measured by functional magnetic resonance imaging in humans. Stimuli included 1,179 unique scenes across 18 semantic categories. Object composition of scenes were compared across scene exemplars in different semantic scene categories, and separately, in exemplars of the same scene category. Neural representations in scene- and object-preferring brain regions were significantly related to which objects were in a scene, with the effect at times stronger in the scene-preferring regions. The object model accounted for more variance when comparing scenes within the same semantic category to scenes from different categories. Here, we demonstrate the function of scene-preferring regions includes the processing of objects. This suggests visual processing regions may be better characterized by the processes, which are engaged when interacting with the stimulus kind, such as processing groups of objects in scenes, or processing a single object in our foreground, rather than the stimulus kind itself.

Semantic Expectation Effects on Object Detection: Using Figure Assignment to Elucidate Mechanisms

Recent evidence suggesting that object detection is improved following valid rather than invalid labels implies that semantics influence object detection. It is not clear, however, whether the results index object detection or feature detection. Further, because control conditions were absent and labels and objects were repeated multiple times, the mechanisms are unknown. We assessed object detection via figure assignment, whereby objects are segmented from backgrounds. Masked bipartite displays depicting a portion of a mono-oriented object (a familiar configuration) on one side of a central border were shown once only for 90 or 100 ms. Familiar configuration is a figural prior. Accurate detection was indexed by reports of an object on the familiar configuration side of the border. Compared to control experiments without labels, valid labels improved accuracy and reduced response times (RTs) more for upright than inverted objects (Studies 1 and 2). Invalid labels denoting different superordinate-level objects (DSC; Study 1) or same superordinate-level objects (SSC; Study 2) reduced accuracy for upright displays only. Orientation dependency indicates that effects are mediated by activated object representations rather than features which are invariant over orientation. Following invalid SSC labels (Study 2), accurate detection RTs were longer than control for both orientations, implicating conflict between semantic representations that had to be resolved before object detection. These results demonstrate that object detection is not just affected by semantics, it entails semantics.

Learned value and predictiveness affect gaze but not figure assignment

Many factors affect figure-ground segregation, but the contributions of attention and reward history to this process is uncertain. We conducted two experiments to investigate whether reward learning influences figure assignment and whether this relationship was mediated by attention. Participants learned to associate certain shapes with a reward contingency: During a learning phase, they chose between two shapes on each trial, with subsets of shapes associated with high-probability win, low-probability win, high-probability loss, and low-probability loss. In a test phase, participants were given a figure-ground task, in which they indicated which of two regions that shared a contour they perceived as the figure (high-probability win and low-probability win shapes were pitted against each other, as were high-probability loss and low-probability loss shapes). The results revealed that participants had learned the reward contingencies and that, following learning, attention was reliably drawn to the optimal stimulus. Despite this, neither reward history nor the resulting attentional allocation influenced figure-ground organization.

Semantic category priming from the groundside of objects shown in nontarget locations and at unpredictable times

Previous research demonstrated that familiar objects that are suggested, but not consciously perceived, on the groundside of the contours of a figure activate their semantic category during perceptual organization, at least when the figure appears at fixation at an expected time. Here, we investigate whether evidence for such semantic activation extends to stimuli presented at unpredictable times in peripheral locations. Participants categorized words shown centrally as denoting natural or artificial objects (Experiments 1 and 2a) or positive or negative concepts (Experiment 2b). Prior to the word, two distractor silhouettes appeared above and below fixation; both depicted novel figures. On experimental trials, portions of well-known (familiar) objects were suggested on the groundside of the borders of one (Experiment 1) or both (Experiment 2a and 2b) silhouettes. In Experiment 1, reaction times were slower when targets words were preceded by experimental distractor silhouettes regardless of whether the object suggested on the groundside of their borders was in the same or a different category as the object denoted by the word. Overall slowing may have occurred because (a) semantic category access by objects suggested on the groundside of experimental distractor silhouettes was sufficient to require filtering but not category-specific priming, (b) more competition for object status slowed processing of experimental compared to control silhouettes, or (c) featural differences increased the difficulty of processing the experimental versus the control silhouettes. The use of two identical experimental silhouettes in Experiment 2a allowed a semantic category priming effect to emerge, showing that the categories of objects suggested on the groundside of silhouette borders can be activated at unpredictable times in nontarget locations and in more than one location of the visual field. Experiment 2a suggested that (a) better explains the results of Experiment 1 than (b and c). Experiment 2b further ruled out explanations (b and c) as reasons for the Experiment 1 results by showing that the same pattern is not obtained when the semantic category of the objects suggested on the groundside of the experimental silhouettes borders is not task-relevant and does not require filtering. Thus, spatial prime-target congruence and temporal certainty are not necessary for priming by objects suggested on the groundside of figures. Implications for our understanding of the complex processes involved in perceptual organization are considered.

Intrusive effects of semantic information on visual selective attention

Every object is represented by semantic information in extension to its low-level properties. It is well documented that such information biases attention when it is necessary for an ongoing task. However, whether semantic relationships influence attentional selection when they are irrelevant to the ongoing task remains an open question. The ubiquitous nature of semantic information suggests that it could bias attention even when these properties are irrelevant. In the present study, three objects appeared on screen, two of which were semantically related. After a varying time interval, a target or distractor appeared on top of each object. The objects’ semantic relationships never predicted the target location. Despite this, a semantic bias on attentional allocation was observed, with an initial, transient bias to semantically related objects. Further experiments demonstrated that this effect was contingent on the objects being attended: if an object never contained the target, it no longer exerted a semantic influence. In a final set of experiments, we demonstrated that the semantic bias is robust and appears even in the presence of more predictive cues (spatial probability). These results suggest that as long as an object is attended, its semantic properties bias attention, even if it is irrelevant to an ongoing task and if more predictive factors are available.

Neural evidence for competition-mediated suppression in the perception of a single object

Multiple objects compete for representation in visual cortex. Competition may also underlie the perception of a single object. Computational models implement object perception as competition between units on opposite sides of a border. The border is assigned to the winning side, which is perceived as an object (or "figure"), whereas the other side is perceived as a shapeless ground. Behavioral experiments suggest that the ground is inhibited to a degree that depends on the extent to which it competed for object status, and that this inhibition is relayed to low-level brain areas. Here, we used fMRI to assess activation for ground regions of task-irrelevant novel silhouettes presented in the left or right visual field (LVF or RVF) while participants performed a difficult task at fixation. Silhouettes were designed so that the insides would win the competition for object status. The outsides (grounds) suggested portions of familiar objects in half of the silhouettes and novel objects in the other half. Because matches to object memories affect the competition, these two types of silhouettes operationalized, respectively, high competition and low competition from the grounds. The results showed that activation corresponding to ground regions was reduced for high- versus low-competition silhouettes in V4, where receptive fields (RFs) are large enough to encompass the familiar objects in the grounds, and in V1/V2, where RFs are much smaller. These results support a theory of object perception involving competition-mediated ground suppression and feedback from higher to lower levels. This pattern of results was observed in the left hemisphere (RVF), but not in the right hemisphere (LVF). One explanation of the lateralized findings is that task-irrelevant silhouettes in the RVF captured attention, allowing us to observe these effects, whereas those in the LVF did not. Experiment 2 provided preliminary behavioral evidence consistent with this possibility.