Infants' sensitivity to pictorial depth cues: a review and meta-analysis of looking studies

Visual segmentation of complex naturalistic structures in an infant eye-tracking search task

An infant’s everyday visual environment is composed of a complex array of entities, some of which are well integrated into their surroundings. Although infants are already sensitive to some categories in their first year of life, it is not clear which visual information supports their detection of meaningful elements within naturalistic scenes. Here we investigated the impact of image characteristics on 8-month-olds’ search performance using a gaze contingent eye-tracking search task. Infants had to detect a target patch on a background image. The stimuli consisted of images taken from three categories: vegetation, non-living natural elements (e.g., stones), and manmade artifacts, for which we also assessed target background differences in lower- and higher-level visual properties. Our results showed that larger target-background differences in the statistical properties scaling invariance and entropy, and also stimulus backgrounds including low pictorial depth, predicted better detection performance. Furthermore, category membership only affected search performance if supported by luminance contrast. Data from an adult comparison group also indicated that infants’ search performance relied more on lower-order visual properties than adults. Taken together, these results suggest that infants use a combination of property- and category-related information to parse complex visual stimuli.

3D shape-from-shading relies on a light source prior that does not change with age

The light-from-above prior enables observers to infer an object's three-dimensional shape-from-shading information. Young, Western adults implicitly assume the light source is placed not only above, but also to the left of, the observer. Previous evidence reached conflicting conclusions regarding the development of the assumed light source direction. In the present study, we measured the light source prior cross-sectionally in children aged 5-11 years, using an explicit shape judgement task. The light-from-above prior, and the left bias, were present as soon as children became sensitive to shading information, regardless of their age. Global processing preference was not related to the ability to perform the task. Similarly, scanning habits, as measured by reading proficiency and starting position in a cancellation task, were not related to the magnitude of the left bias. Children's ability to report shape-from-shading judgements increased with age, but age did not affect the direction of light priors. Thus, we concluded that the development of the light-from-above prior and leftward bias do not require an extended maturation period, but rather the direction of the light-source priors may be developmentally stable once measurable.

The Importance of the Interaction Between Ocular Motor Function and Vision During Human Infancy

Numerous studies have demonstrated the impact of imposed abnormal visual experience on the postnatal development of the visual system. These studies have provided fundamental insights into the mechanisms underlying neuroplasticity and its role in clinical care. However, the ocular motor responses of postnatal human infants largely define their visual experience in dynamic three-dimensional environments. Thus, the immature visual system needs to control its own visual experience. This review explores the interaction between the developing motor and sensory/perceptual visual systems, together with its importance in both typical development and the development of forms of strabismus and amblyopia.

Observation of the point-light animation of a grasping hand activates sensorimotor cortex in nine-month-old infants

Measuring changes in sensorimotor alpha band activity in nine-month-old infants we sought to understand the involvement of the sensorimotor cortex during observation of the Point-Light (PL) animation of a grasping hand. Attenuation of alpha activity was found both when the PL display moved towards the to-be-grasped object and when the object was deleted from the video. Before the beginning of the movement of the PL stimuli, only in the presence of the object evoked attenuation of sensorimotor alpha activity was documented, possibly interpreted either as movement prediction or as graspable object perception. Our main findings demonstrate that, during observation of stimuli moving with biological kinematics, the infants' sensorimotor system is activated when the pictorial information is absent or highly reduced, and independently of the presence of the goal-directed object. The possible compensatory function of the sensorimotor system during observation of highly degraded moving stimuli is discussed.

Asymmetry of pictorial space: A cultural phenomenon

Art experts have argued that the mirror reversal of pictorial artworks produces an alteration of their spatial content. However, this putative asymmetry of the pictorial space remains to be empirically proved and causally explained. Here, we address these issues with the "corridor illusion," a size illusion triggered by the pictorial space of a receding corridor. We show that mirror-reversed corridors-receding respectively leftward and rightward-induce markedly different illusion strengths and thus convey distinct pictorial spaces. Remarkably, the illusion is stronger with the rightward corridor among native left-to-right readers (French participants, n = 40 males) but conversely stronger with the leftward corridor among native right-to-left readers (Syrian participants, n = 40 males). Together, these results demonstrate an asymmetry of the pictorial space and point to our reading/writing habits as a major cause of this phenomenon.

A model for discovering 'containment' relations

Rapid developments in the fields of learning and object recognition have been obtained by successfully developing and using methods for learning from a large number of labeled image examples. However, such current methods cannot explain infants' learning of new concepts based on their visual experience, in particular, the ability to learn complex concepts without external guidance, as well as the natural order in which related concepts are acquired. A remarkable example of early visual learning is the category of 'containers' and the notion of 'containment'. Surprisingly, this is one of the earliest spatial relations to be learned, starting already around 3 month of age, and preceding other common relations (e.g., 'support', 'in-between'). In this work we present a model, which explains infants' capacity of learning 'containment' and related concepts by 'just looking', together with their empirical development trajectory. Learning occurs in the model fast and without external guidance, relying only on perceptual processes that are present in the first months of life. Instead of labeled training examples, the system provides its own internal supervision to guide the learning process. We show how the detection of so-called 'paradoxical occlusion' provides natural internal supervision, which guides the system to gradually acquire a range of useful containment-related concepts. Similar mechanisms of using implicit internal supervision can have broad application in other cognitive domains as well as artificial intelligent systems, because they alleviate the need for supplying extensive external supervision, and because they can guide the learning process to extract concepts that are meaningful to the observer, even if they are not by themselves obvious, or salient in the input.

Are Different Actions Mediated by Distinct Systems of Knowledge in Infancy?

This chapter considers why studies of infant looking and reaching often suggest different patterns of cognitive and perceptual development. In some cases, convergent results have emerged from studies of infant looking and reaching, but differences are common. The most typical results suggest less adult-like perception and cognition in studies of reaching than in studies of looking. Several reaching studies, however, do not fit this pattern, suggesting that reaching actions may be mediated by distinct systems of knowledge and information processing. Comparisons of research on other behaviors, such as crawling and walking, also suggest that infant knowledge systems vary across actions. Research on how adult size perception differs between verbal and reaching response behaviors is considered and used as a template to interpret the developmental results. Like adults, when infants prepare to engage in particular actions, they seem to shift their sensitivity to particular sources of information and to process that information in action-relevant ways. These tendencies suggest that distinct knowledge systems mediate different actions in infancy.

The life-span trajectory of visual perception of 3D objects

Deriving a 3D structural representation of an object from its 2D input is one of the great challenges for the visual system and yet, this type of representation is critical for the successful recognition of and interaction with objects. Perhaps reflecting the importance of this computation, infants have some sensitivity to 3D structural information, and this sensitivity is, at least, partially preserved in the elderly population. To map precisely the life-span trajectory of this key visual computation, in a series of experiments, we compared the performance of observers from ages 4 to 86 years on displays of objects that either obey or violate possible 3D structure. The major findings indicate that the ability to derive fine-grained 3D object representations emerges after a prolonged developmental trajectory and is contingent on the explicit processing of depth information even in late childhood. In contrast, the sensitivity to object 3D structure remains stable even through late adulthood despite the overall reduction in perceptual competence. Together, these results uncover the developmental process of an important perceptual skill, revealing that the initial, coarse sensitivity to 3D information is refined, automatized and retained over the lifespan.

Emergence of the ability to perceive dynamic events from still pictures in human infants

The ability to understand a visual scene depicted in a still image is among the abilities shared by all human beings. The aim of the present study was to examine when human infants acquire the ability to perceive the dynamic events depicted in still images (implied motion perception). To this end, we tested whether 4- and 5-month-old infants shifted their gaze toward the direction cued by a dynamic running action depicted in a still figure of a person. Results indicated that the 5- but not the 4-month-olds showed a significant gaze shift toward the direction implied by the posture of the runner (Experiments 1, 2, and 3b). Moreover, the older infants showed no significant gaze shift toward the direction cued by control stimuli, which depicted a figure in a non-dynamic standing posture (Experiment 1), an inverted running figure (Experiment 2), and some of the body parts of a running figure (Experiment 3a). These results suggest that only the older infants responded in the direction of the implied running action of the still figure; thus, implied motion perception emerges around 5 months of age in human infants.

A Picture You Can Handle: Infants Treat Touch-Screen Images More Like Photographs than Objects

Infants actively explore their world in order to determine the different ways in which they can interact with various objects. Although research on infant perception has focused on how infants understand the differences between 2- and 3-dimensional objects, today's infants increasingly encounter 2D images with interactive qualities on smart-phone screens, tablets, and laptops. The purpose of this experiment was to examine the types of manual behaviors infants direct toward tablet images and to compare these actions to those evoked by 2D photographs or 3D when tactile feedback is controlled. Infants between the ages of 7-10 months sat on their parent's lap in front of a table with a built-in well covered by a clear, plastic sheet while the three types of displays (photographs, objects, and screen images on a tablet) were presented for 30 s each. Infants saw three examples of each type of display presented in the built-in well so that tactile feedback information from the different displays was controlled. Coders noted the proportion of trials in which infants grasped, scratched, rubbed, or patted the display. Results indicate that infants direct significantly more grasps, scratches, and rubs toward 3D objects than 2D photographs. Infants also direct more grasps to objects compared to screen images. Our data suggests that infants are treating screen images more similarly to 2D photographs than 3D objects.

Distinct Visual Processing of Real Objects and Pictures of Those Objects in 7- to 9-month-old Infants

The present study examined 7- and 9-month-old infants' visual habituation to real objects and pictures of the same objects and their preferences between real and pictorial versions of the same objects following habituation. Different hypotheses would predict that infants may habituate faster to pictures than real objects (based on proposed theoretical links between behavioral habituation in infants and neuroimaging adaptation in adults) or to real objects vs. pictures (based on past infant electrophysiology data). Sixty-one 7-month-old infants and fifty-nine 9-month-old infants were habituated to either a real object or a picture of the same object and afterward preference tested with the habituation object paired with either the novel real object or its picture counterpart. Infants of both age groups showed basic information-processing advantages for real objects. Specifically, during the initial presentations, 9-month-old infants looked longer at stimuli in both formats than the 7-month olds but more importantly both age groups looked longer at real objects than pictures, though with repeated presentations, they habituated faster for real objects such that at the end of habituation, they looked equally at both types of stimuli. Surprisingly, even after habituation, infants preferred to look at the real objects, regardless of whether they had habituated to photos or real objects. Our findings suggest that from as early as 7-months of age, infants show strong preferences for real objects, perhaps because real objects are visually richer and/or enable the potential for genuine interactions.

Sensory development: late integration of multiple cues

Young children use multiple cues to appreciate the three-dimensional structure of the world. A new study reveals that these cues are properly integrated only years later, thus showing that sensory development is protracted well into teenage years.

Eleven-month-old infants infer differences in the hardness of object surfaces from observation of penetration events

Previous studies have shown different developmental trajectories for object recognition of solid and non-solid objects. However, there is no evidence as to whether infants have expectations regarding certain attributes of objects, such as surface hardness, in the absence of tactile information. In the present study, we examined infants’ perception of the hardness of object surfaces from visually presented penetration events using the familiarization–novelty preference procedure. Experiment 1 showed that by 11 months old infants distinguished a relatively soft surface from a crusty surface based on changes in the velocity of a moving object as the moving object penetrated the surface of the target object. Experiment 2 ruled out the possibility that infants were merely sensitive to differences in the velocity changes in the stimuli.

Infant Perception of Incongruent Shapes in Cast Shadows

A cast shadow occurs when an object blocks the light from an illumination and projects a dark region onto a surface. Previous studies have reported that adults are slower to identify an object when the object has an incongruent cast shadow than when it has a congruent cast shadow (Castiello, 2001). Here, we used the familiarization-novelty preference procedure to examine whether 5- to 8-month-old infants could detect the relationship between object shapes and cast shadows. In Experiment 1, we examined the infants' ability to detect incongruency between objects and cast shadows. Results showed that 7- to 8-month olds could detect incongruence between the object shapes and the cast shadows, whereas 5- to 6-month olds did not. Yet, our control experiment showed that infants could not detect this incongruence from stimuli in which a white outline had been added to the original cast shadow to decrease the possibility of it being perceived as a cast shadow (Experiment 2). The results of these experiments demonstrate that 7- to 8-month olds responded to the congruence of cast shadows and to consistent contrast polarity between the cast shadow and its background.

Infants' ability to respond to depth from the retinal size of human faces: comparing monocular and binocular preferential-looking

To examine sensitivity to pictorial depth cues in young infants (4 and 5 months-of-age), we compared monocular and binocular preferential looking to a display on which two faces were equidistantly presented and one was larger than the other, depicting depth from the size of human faces. Because human faces vary little in size, the correlation between retinal size and distance can provide depth information. As a result, adults perceive a larger face as closer than a smaller one. Although binocular information for depth provided information that the faces in our display were equidistant, under monocular viewing, no such information was provided. Rather, the size of the faces indicated that one was closer than the other. Infants are known to look longer at apparently closer objects. Therefore, we hypothesized that infants would look longer at a larger face in the monocular than in the binocular condition if they perceived depth from the size of human faces. Because the displays were identical in the two conditions, any difference in looking-behavior between monocular and binocular viewing indicated sensitivity to depth information. Results showed that 5-month-old infants preferred the larger, apparently closer, face in the monocular condition compared to the binocular condition when static displays were presented. In addition, when presented with a dynamic display, 4-month-old infants showed a stronger 'closer' preference in the monocular condition compared to the binocular condition. This was not the case when the faces were inverted. These results suggest that even 4-month-old infants respond to depth information from a depth cue that may require learning, the size of faces.

Delayed visual orienting responses in children with developmental and/or intellectual disabilities

BACKGROUND

Assessment of higher visual processing functions mostly requires active cooperation of participants, which is problematic in children with intellectual disabilities (ID). To circumvent this, we applied remote eye tracking to quantify (ab)normal visual orienting responses in children with ID in terms of reaction times to visual stimuli.

METHODS

We presented visual stimuli (cartoon, coherent form, and coherent motion) to 127 children (2-14 years) with developmental and/or ID (risk group) and simultaneously measured their orienting ocular motor responses. Reaction times to fixation (RTF) in the risk group were compared with RTF values of an age-matched control group.

RESULTS

Overall, in 72% of the children in the risk group, RTF values to cartoon were delayed, in 47% to form, and in 38% to motion. The presence of delayed reaction times was highest in the group of children >4 years with ID.

CONCLUSION

Our data show that a majority of children with developmental and/or ID have delayed visual orienting responses. This suggests that this group has increased risk for higher visual processing dysfunctions. Future studies are planned to correlate abnormal orienting responses to type of brain damage and to dissociate the responses from ocular motor disorders.

Infants' perception of curved illusory contour with motion

Recently, Masuda et al. (submitted for publication) showed that adults perceive moving rigid or nonrigid motion from illusory contour with neon color spreading in which the inducer has pendular motion with or without phase difference. In Experiment 1, we used the preferential looking method to investigate whether 3-8-month-old infants can discriminate illusory and non-illusory contour figures, and found that the 7-8-month-old, but not the 3-6-month-old, infants showed significant preference for illusory contour with phase difference. In Experiment 2, we tested the validity of the visual stimuli in the present study, and whether infants could detect illusory contour from the current neon color spreading figures. The results showed that all infants might detect illusory contour figure with neon color spreading figures. The results of Experiments 1 and 2 suggest that 7-8-month-old infants potentially perceive illusory contour from the visual stimulus with phase-different movement of inducers, which elicits the perception of nonrigid dynamic subjective contour in adults.

Measuring young infants' sensitivity to height-in-the-picture-plane by contrasting monocular and binocular preferential-looking

To examine young infants' sensitivity to a pictorial depth cue, we compared monocular and binocular preferential looking to objects of which depth was specified by height-in-the-picture-plane. For adults, this cue generates the perception that a lower object is closer than a higher object. This study showed that 4- and 5-month-old infants fixated the lower, apparently closer, figure more often under the monocular than binocular presentation providing evidence of their sensitivity to the pictorial depth cue. Because the displays were identical in the two conditions except for binocular information for depth, the difference in looking-behavior indicated sensitivity to depth information, excluding a possibility that they responded to 2D characteristics. This study also confirmed the usefulness of the method, preferential looking with a monocular and binocular comparison, to examine sensitivity to a pictorial depth cue in young infants, who are too immature to reach reliably for the closer of two objects.

Getting the closer object? An information-based dissociation between vision for perception and vision for movement in early infancy

In human adults two functionally and neuro-anatomically separate systems exist for the use of visual information in perception and the use of visual information to control movements (Milner & Goodale, 1995, 2008). We investigated whether this separation is already functioning in the early stages of the development of reaching. To this end, 6- and 7-month-old infants were presented with two identical objects at identical distances in front of an illusory Ponzo-like background that made them appear to be located at different distances. In two further conditions without the illusory background, the two objects were presented at physically different distances. Preferential reaching outcomes indicated that the allocentric distance information contained in the illusory background affected the perception of object distance. Yet, infants' reaching kinematics were only affected by the objects' physical distance and not by the perceptual distance manipulation. These findings were taken as evidence for the two-visual systems, as proposed by Milner and Goodale (2008), being functional in early infancy. We discuss the wider implications of this early dissociation.

The ground is dominant in infants' perception of relative distance

When making relative distance judgments, adults attend to information provided by the ground surface and generally ignore information provided by ceiling surfaces. In the present study, we asked whether this ground dominance effect is present in infancy. Groups of 5- and 7-month-old infants viewed a display depicting textured ground and ceiling surfaces. Two toys, which were attached to vertical rods, were affixed to the display. The toys/rods were positioned so that one toy was specified as being nearer by the ground surface but farther away by the ceiling surface, while the other toy was specified as being farther away by the ground surface but nearer by the ceiling surface. Under monocular viewing conditions, the infants in both age groups reached preferentially for the toy that was specified as being nearer by the ground surface. This effect was significantly stronger than that observed under binocular viewing conditions. The findings indicate that the infants responded to the distance information provided by the ground surface to a greater extent than to information provided by the ceiling.

Young infants' perception of the trajectories of two- and three-dimensional objects

We investigated oculomotor anticipations in 4-month-old infants as they viewed center-occluded object trajectories. In two experiments, we examined performance in two-dimensional (2D) and three-dimensional (3D) dynamic occlusion displays and in an additional 3D condition with a smiley face as the moving target stimulus. Rates of anticipatory eye movements were not facilitated by 3D displays or by the (presumably) more salient smiley face relative to the 2D condition. However, latencies of anticipations were reduced, implying that 3D visual information may have supported formation of more robust mental representations of the moving object. Results are interpreted in a context of perceptual constraints on developing cognitive capacities during early infancy.