Visual learning of statistical relations among nonadjacent features: Evidence for structural encoding

Recent results suggest that observers can learn, unsupervised, the co-occurrence of independent shape features in viewed patterns (e.g., Fiser & Aslin, 2001). A critical question with regard to these findings is whether learning is driven by a structural, rule-based encoding of spatial relations between distinct features or by a pictorial, template-like encoding, in which spatial configurations of features are embedded in a “holistic” fashion. In two experiments, we test whether observers can learn combinations of features when the paired features are separated by an intervening spatial “gap”, in which other, unrelated features can appear. This manipulation both increases task difficulty and makes it less likely that the feature combinations are encoded simply as larger unitary features. Observers exhibited learning consistent with earlier studies, suggesting that unsupervised learning of compositional structure is based on the explicit encoding of spatial relations between separable visual features. More generally, these results provide support for compositional structure in visual representation.     

Characterizing typographic expertise: Do we process typefaces like faces?

Studies of face recognition and discrimination provide a rich source of data and debate on the nature of their processing, in particular through using inverted faces. This study draws parallels between the features of typefaces and faces, as letters share a basic configuration, regardless of typeface, that could be seen as similar to faces. Typeface discrimination is compared using paragraphs of upright letters and inverted letters at three viewing durations. Based on previously reported effects of expertise, the prediction that designers would be less accurate when letters are inverted, whereas nondesigners would have similar performance in both orientations, was confirmed. A proposal is made as to which spatial relations between typeface components constitute holistic and configural processing, posited as the basis for better discrimination of the typefaces of upright letters. Such processing may characterize designers' perceptual abilities, acquired through training.     

Similarity grouping as feature-based selection

Our visual system groups objects with similar features, such as colour, orientation, or shape. We argue that similarity grouping is nothing more than global selection of a certain feature value, such as red, horizontal, or circular. This account makes the striking prediction that only one feature group can be created at a time. Here we provide the most direct evidence yet for this proposal, using a number estimation task that forces simultaneous processing of all objects. Multiple similarity cues failed to produce grouping in this task, in contrast to proximity, connectivity, and common region, which all showed strong grouping effects.     

More of myself: Manipulating interoceptive awareness by heightened attention to bodily and narrative aspects of the self

Psychology distinguishes between a bodily and a narrative self. Within neuroscience, models of the bodily self are based on exteroceptive sensorimotor processes or on the integration of interoceptive sensations. Recent research has revealed interactions between interoceptive and exteroceptive processing of self-related information, for example that mirror self-observation can improve interoceptive awareness. Using heartbeat perception, we measured the effect on interoceptive awareness of two experimental manipulations, designed to heighten attention to bodily and narrative aspects of the self. Participants gazed at a photograph of their own face or at self-relevant words. In both experimental conditions interoceptive awareness was significantly increased, compared to baseline. Results show that attention to narrative aspects of self, previously regarded as relying on higher-order processes, has an effect similar to self-face stimuli in improving interoceptive awareness. Our findings extend the previously observed interaction between the bodily self and interoception to more abstract amodal representations of the self.     

Attending to what matters: Flexibility in adults’ and infants’ action perception

Action perception is selective in that observers attend to and encode certain dimensions of action over others. But how flexible is action perception in its selection of perceptual information? One possibility is that observers consistently attend to particular dimensions of action over others across different contexts. Another possibility, tested here, is that observers flexibly vary their attention to different dimensions of action based on the context in which action occurs. We investigated 9.5-month-old infants’ and adults’ ability to attend to drop height under varying contexts—aiming to drop an object into a narrow container versus a wide container. We predicted differential attention to increases in aiming height for the narrow container versus the wide container because an increase in aiming height has a differential effect on success (i.e., getting the object into the container) depending on the width of the container. Both adults and infants showed an asymmetry in their attention to aiming height as a function of context; in the wide container condition increases and decreases in aiming height were equally detectable, whereas in the narrow container condition observers more readily discriminated increases over decreases in aiming height. These results indicate that action perception is both selective and flexible according to context, aiding in action prediction and infants’ social–cognitive development more broadly.     

Perceptual training to increase drivers’ ability to spot motorcycles at T-junctions

Motorcyclists too often collide with other road users who pull out of side roads in front of them. These other road users typically report making all the necessary visual checks, despite failing to see the approaching motorcycle. These Look But Fail To See errors appear to be attenuated in road users who themselves have motorcycling experience, suggesting that motorcycle exposure may lower thresholds for spotting these vulnerable road users through natural perceptual learning. This raises the possibility that perceptual training could improve car drivers’ abilities to spot motorcycles. Two experiments are reported. The first experiment demonstrated that a T-junction task, requiring participants to detect an approaching vehicle in briefly displayed images, was sensitive to participants’ motorcycle experience, with dual drivers (who both ride motorcycles and drive cars) performing better than average car drivers. Following this, a second experiment split the car drivers into 2 groups. One group undertook a Pelmanism task requiring participants to match pairs of motorcycles, while the control group had to match pairs of fruit. When the two groups were re-tested on the T-junction task, the group who had undergone perceptual training for motorcycles via the Pelmanism task, were better able to identify approaching motorcycles, but not approaching cars. The results suggest that gamification of perceptual training for motorcycle detection provides a novel opportunity to improve driver safety.     

Illusions of integration are subjectively impenetrable: Phenomenological experience of Lag 1 percepts during dual-target RSVP

We investigated the relationship between different kinds of target reports in a rapid serial visual presentation task, and their associated perceptual experience. Participants reported the identity of two targets embedded in a stream of stimuli and their associated subjective visibility. In our task, target stimuli could be combined together to form more complex ones, thus allowing participants to report temporally integrated percepts. We found that integrated percepts were associated with high subjective visibility scores, whereas reports in which the order of targets was reversed led to a poorer perceptual experience. We also found a reciprocal relationship between the chance of the second target not being reported correctly and the perceptual experience associated with the first one. Principally, our results indicate that integrated percepts are experienced as a unique, clear perceptual event, whereas order reversals are experienced as confused, similar to cases in which an entirely wrong response was given.     

Attention and cognitive penetrability: The epistemic consequences of attention as a form of metacognitive regulation

A recent approach to the cognitive penetrability of perception, i.e. the possibility that perception is shaped top-down by high-level cognitive states such as beliefs and desires, proposes to understand the phenomenon on the basis of its consequences, among which there is a challenge for the epistemic role of perceptual experience in justifying beliefs (Stokes, 2015). In this paper, I argue that some attentional phenomena qualify as cases of cognitive penetrability under this consequentialist approach. I present a popular theory of attention, the biased-competition theory, on which basis I establish that attention is a form of metacognitive regulation. I argue that attention (as metacognitive regulation) involves the right kind of cognitive-perceptual relation and leads to the same epistemic consequences as other more traditional versions of cognitive penetrability.     

Volume completion in 4.5-month-old infants

In this study, we examined 4.5-month-old infants' visual completion of self-occluding three-dimensional objects. A previous study on this topic reported that 6-month-old, but not 4-month-old infants extrapolate a convex, symmetric prism from a limited view of its surfaces (Soska & Johnson, 2008). As of yet, studies on the development of amodal completion of three-dimensional, self-occluding objects are scarce. Given 4-month-old infants' abilities to derive three-dimensional shape from a variety of visual cues, three-dimensional amodal completion may well depend on the perceptual strength of three-dimensionality in the stimulus displays. The first experiments (1A and 1B) tested this hypothesis by means of a habituation paradigm and showed that 4.5-month-old infants are indeed able to amodally complete the back of a self-occluding object when sufficient three-dimensional cues are available. Further support for volume completion in 4.5-month-old infants was found in a second experiment, again using a habituation paradigm, that measured perceived connectedness between two visually separated, self-occluding, three-dimensional objects.     

Are People Successful at Learning Sequences of Actions on a Perceptual Matching Task?

We report the results of an experiment in which human subjects were trained to perform a perceptual matching task. Subjects were asked to manipulate comparison objects until they matched target objects using the fewest manipulations possible. An unusual feature of the experimental task is that efficient performance requires an understanding of the hidden or latent causal structure governing the relationships between actions and perceptual outcomes. We use two benchmarks to evaluate the quality of subjects' learning. One benchmark is based on optimal performance as calculated by a dynamic programming procedure. The other is based on an adaptive computational agent that uses a reinforcement-learning method known as Q-learning to learn to perform the task. Our analyses suggest that subjects were successful learners. In particular, they learned to perform the perceptual matching task in a near-optimal manner (i.e., using a small number of manipulations) at the end of training. Subjects were able to achieve near-optimal performance because they learned, at least partially, the causal structure underlying the task. In addition, subjects' performances were broadly consistent with those of model-based reinforcement-learning agents that built and used internal models of how their actions influenced the external environment. We hypothesize that people will achieve near-optimal performances on tasks requiring sequences of action-especially sensorimotor tasks with underlying latent causal structures-when they can detect the effects of their actions on the environment, and when they can represent and reason about these effects using an internal mental model.