New research led by the University of St Andrews finds that a common eye condition has little or no impact on an aspect of a person’s three-dimensional perception.
The research, led by scientists from the School of Psychology and Neuroscience and published by the Royal Society Open Science, reveals, for the first time, that a common eye condition called strabismus, also often referred to as a squint, does not impair a person’s ability to perceive 3-D layout.
Strabismus, an ophthalmological condition of the misalignment of the eyes typically arising from neurological disorders, results in an inability to focus both eyes on a single point in space.
The condition is known to be linked to neuropsychological impairments including amblyopia (lazy eye) and deficits in depth perception.
The exact nature of the deficits in 3-D and depth perception is poorly understood as knowledge of the way individuals with strabismus experience depth has often been based on anecdotal reports and conjecture. Previous studies have focused on one aspect of 3-D perception, the capacity to conduct motoric tasks that require estimates of distances to nearby objects.
Relative depth perception is a fundamental aspect of 3-D vision that underlies a person’s capacity to perceive 3-D shapes and layout of objects.
The team of scientists investigated, for the first time, the perception of relative depth from the cue of linear perspective in strabismus using pictorial images. The findings reveal that strabismics are unimpaired in the perception of relative depth in comparison to observers with typically developed binocular vision.
Lead author, Dr. Giedre Zlatkute, from the School of Psychology & Neuroscience and currently at the School of Medicine at the University of St Andrews, said:
Senior author of the study, Dr. Dhanraj Vishwanath, said: “These results support a theory that the brain constructs at least three distinct types of representation that support different competencies in 3-D vision: the perception of distances of objects near the viewer (in personal space), the perception of object distances at walking distances, and the perception of 3-D object shape/layout (relative depth).
Our results, combined with a few other studies, suggest that individuals with strabismus may be impaired only in the first component (perception of near distances) but have largely intact perception in the other two.”
Dr. Zlatkute explained: “In addition to difficulties in perceiving distances in near space, individuals with strabismus can also be impaired in perceiving the ‘3-D effect’ of 3-D movies. Further studies are required to fully understand how these impairments are linked and how they impact on a person’s overall 3-D perception in comparison to those with typically developed binocular vision.”
Strabismus (a.k.a. squint) is an ophthalmological condition of the misalignment of the eyes. Due to a range of neurological or physiological issues ranging from extra-ocular muscle paralysis to developmental delays and also genetic conditions such as Down’s syndrome, individuals with strabismus are unable to coordinate their eye muscles to fixate the two eyes on a single target point in space [1].
This results in the two eyes receiving information from two non-corresponding areas of space which can lead to difficulties in vision and particularly three-dimensional (3D) space perception [2]. Understanding the extent of perceptual and functional deficits arising from this binocular deficit is important considering that close to 5% of the population are diagnosed with strabismus [3,4].
Strabismus disrupts sensory fusion, the cortical process of combining the images from the two eyes into a single binocular image [3–6]. The main perceptual consequences of lack of fused binocular images is diplopia (double vision) and a lack of binocular depth perception. If strabismus is congenital or manifests in early childhood, the human visual system finds ways of adapting and avoiding the ambiguity of double vision.
In most cases, this leads to the cortical suppression of one eye’s image resulting in amblyopia, while under more specific circumstances, such as infantile constant strabismus, anomalous retinal correspondence or ‘new’ fovea may develop [2,4]. The lack of binocular fusion also implies that binocular disparities cannot be used to compute depth and induce the characteristic impression of stereopsis.
While there is extensive literature on the consequences and treatment of amblyopia [7–13], the assessment of deficits in depth perception in strabismus has been more limited. Until recently, much of the understanding of the qualitative and quantitative perception of depth in individuals with strabismus has relied on anecdotal reports and conjecture.
Popular media accounts [14,15] combined with the limited empirical studies, paint a mixed picture regarding the specific capacities of strabismics to perceive depth. Some conjectures in popular media and textbooks, along with early studies, suggest that despite the absence of binocular depth perception, strabismics can perceive depth on the basis of monocular cues [5,14,16,17].
Some recent studies suggest that despite the absence of measurable stereoacuity, observers are even able to appreciate the ‘3D effect’ associated with stereopsis in video images that contain depth information from motion [18–20]. By contrast, the most detailed and widely known subjective report of depth perception of an individual with infantile strabismus suggests that perceived depth in strabismus from monocular cues is significantly inferior to depth perception derived from binocular disparities at the qualitative, quantitative and functional level [15].
Barry, a neuroscientist, had no stereovision for most of her life but recovered normal binocular correspondence and stereopsis in late adulthood (age 47) through a series of orthoptic exercises. Barry’s introspective reports comparing her perception prior to acquiring stereopsis and after suggest that the depth perceived even from monocular cues (e.g. when watching a two-dimensional (2D) movie) is inferior in strabismus and probably inferred in a top-down manner [15].
Varieties of depth perception
In order to better understand the nature of depth perception capacities and deficits in strabismus, it is first necessary to discriminate between the different varieties of depth perception that an individual may experience. Depth perception is often described in textbooks as the ability to judge the distance of objects and the spatial relationship among objects at different distances [21].
The initial part of the definition refers to the perception of egocentric distance (or simply distance perception), namely the distance between an observer and the viewed objects. The second part of the definition refers to depth perception, i.e. the perception of separations in depth between one object and another or between different parts of a single object. A further distinction can be made between relative and absolute (scaled) depth perception [22,23].
Relative depth is the capacity to perceive the depth relations among points in space without knowing the actual values of depth separation. Relative depth can further be broken down into the perception of depth order (ordinal depth) which is simply perceiving the ordering of objects in depth without any impression of the degree of separation between them, and the perception of depth ratios (perceiving the relative magnitudes of separations in depth within or among objects).
The perception of 3D object shape or layout constitutes the perception of relative depth ratios. Absolute (scaled) depth is the perception of the actual depth separations between objects or points within an object scaled to some egocentric motor metric and is required in order to execute visuomotor tasks such as grasping (see figure 1).
The distinction between the perception of relative and scaled (absolute) depth. In the top panel, the observer has information about the relative ratios of separations between the four objects, but not the actual values of the separations (units of distance). The observer’s perception of 3D shape and layout is, therefore, ambiguous up to a uniform scaling factor. In the two lower panels, the observer perceives the actual separations between the objects in some egocentric units (units of distance) relevant for interaction.
Egocentric distance perception in near space (less than 2 m) is believed to rely primarily on extra-retinal cues such as accommodation, binocular convergence and defocus blur, while the judgement of the egocentric distance of objects located on the ground plane in farther space (5–25 m) appears to rely on ground plane information and declination from eye level [24–31].
Since convergence is impaired in strabismus, judging distances in near space should be compromised, while the perception of far distances along the ground plane is not expected to be compromised since it is based primarily on visual declination. Recent studies that have examined distance perception in strabismics for near and far space are consistent with this [28,32,33].
Within near (personal) space, perception of egocentric distance has been assessed by having subjects perform high-precision visuomotor tasks without any haptic feedback [32–34]. Consistent with the lack of binocular convergence information, clear deficits are seen in high-precision tasks within reach space where subjects misjudge the distance to slots when placing pegs and misjudge the distance of a bead to be threaded [32,33].
For far distances, strabismics show the normal ability to blind walk to a previewed target placed on the ground but show deficits in performing these tasks when the same target object is suspended mid-air, a condition which putatively complicates the derivation of distance purely from the ground plane declination information [28].
The predictions with respect to relative and scaled (absolute) depth, however, are more complex. Both monocular depth cues (motion parallax, perspective convergence, texture gradients, relative size, shading) and binocular disparities specify only relative depth relations and need to be scaled by egocentric distance cues such as convergence in order to estimate scaled (absolute) depth.
Due to strabismics’ lack of normal binocular convergence, both the distance cue of vergence and the depth cue of binocular disparity cannot be used, which should compromise the perception of absolute depth at near distances affecting tasks such as grasping or tasks that require nulling of absolute depth such as threading a needle.
Consistent with this, deficits in bead threading (requiring absolute depth nulling) have been demonstrated in strabismics [32,33]. However, to our knowledge, no direct examination of absolute depth perception in strabismics has been conducted by measuring either pantomimed grip apertures or online measurement of grasping dynamics.
Despite deficits in the use of the depth cue of binocular disparity, strabismics should, in principle, be able to use pictorial monocular depth cues (perspective, shading, texture) and motion parallax cues to make judgements of relative depth relations such as 3D shape, surface slant and curvature in depth. However, detailed anecdotal reports suggest that relative depth perception from monocular cues (e.g. motion parallax, shading, perspective) is compromised in strabismus [15].
Specifically, these descriptions suggest that stereotypical observers are able to use monocular cues more effectively (for example when watching standard 2D movies) because they can use ‘a lifetime of past visual experiences (with binocular vision) to re-create the missing stereo information’ (Barry [15, p. 102]).
These reports support a more general view that only binocular disparity ‘directly’ provides a bottom-up quantitative perception of depth, while depth from monocular (pictorial) cues rely on indirect top-down inferences [35], putatively built up through experience with 3D perception from binocular disparity. If the latter were true, we would expect strabismic observer to be impaired in making such top-down inferences of depth using monocular cues due to their lack of learned correlation of the cues with depth from disparity.
Moreover, there is significant evidence that different forms of 3D perception (distance, relative depth and absolute depth) might be dissociated and/or processed separately in the visual pathways [22,23,26,27,36] with potentially distinct developmental trajectories. Strabismus may affect each of them in different ways under different conditions. To our knowledge, no studies have examined the perception of relative depth based on monocular cues in strabismus.
In this study, we aimed to examine, for the first time, the use of a monocular depth cue for the judgement of relative depth in strabismus. Specifically, we aimed to determine the bottom-up susceptibility to the monocular depth cue of perspective convergence as well as the capacity to use the cue to make accurate judgements of relative depth in pictorial images comparing strabismic observers to those with normal stereovision.
While pictorial images are routinely used to examine the role of monocular cues in depth perception, it is generally believed that binocular viewing of pictorial images by observers with normal stereovision will yield shallower estimates of depth than what the cue specifies because of the conflicting disparity cue specifying the flat picture surface.
This belief suggests that strabismics who cannot use binocular disparity might perceive more depth in pictorial images than those with normal stereovision. However, most empirical evidence shows no difference in judgements of depth comparing monocular and binocular viewing for a range of tasks ([37–40]; though see Koenderink et al. [41]). Thus, if strabismics have the same capacity to use relative depth cues, we hypothesize no difference in judgements between the two groups under both monocular and binocular viewing.
Susceptibility to monocular depth cues and the capacity to use them for accurate judgements
The consensus view of monocular (pictorial) depth cues in observers with normal stereovision is that they provide bottom-up quantitative visual information for depth perception, though there are alternative views [35]. A persuasive argument for the consensus view is that when viewing pictorial images that contain monocular depth cues, depth is perceived despite the strong conflicting information from binocular disparity signalling a flat surface [42,43], experimental settings where disparity can be overridden by monocular depth cues [44], and the fact that these cues combine in a lawful way with binocular disparity information [45].
The automatic susceptibility to monocular depth cues is further highlighted in observers’ inability to cognitively suppress perspective cues in making judgements of object size and shape in pictorial images. Observers show systematic misperception of 2D object size and shape, with their judgements falling between the actual 2D stimulation and the implied 3D configuration, despite being instructed to make judgements on only the 2D shape.
Thouless [46] referred to this automatic tendency as ‘phenomenal regression to the real object’ whose magnitude is measured by the Thouless ratio. Such an automatic susceptibility to monocular depth cue of perspective convergence is demonstrated in figure 2. Measuring the Thouless ratio in such stimulus conditions can, therefore, help determine if strabismics are subject to the same bottom-up susceptibility to monocular relative depth cues as stereotypical observers or if they rely on cognitive inferences which can easily be suppressed through cognitive effort.
Demonstration of the Thouless-type regression to the real object when judging interval equidistance of lines superimposed on a perspective pictorial image. The required equidistance judgement is to determine if the separations between the four lines on the 2D plane of the page (ignoring the 3D pictorial content) are equal. (a) The intervals are equidistant on the 2D plane of the surface of the photograph but are typically perceived to be non-equidistant. (b) The spacing consistent with correct equidistant spacing in 3D pictorial space. Images adapted from Fig. 5 in Erkelens [47].
Another way to identify differences between strabismics and non-strabismics is to measure the capacity to make accurate relative depth judgements derived from monocular cues. Judgements of relative depth on the basis of monocular cues typically show underestimation in comparison to ground truth even in stereotypical observers [48], and specifically for perspective convergence [37,47,49].
Determining if strabismics show the same or greater level of underestimation will help identify processing differences between the two groups.
Both these measures (susceptibility and accuracy), taken together, can provide evidence of differences in the early and/or late processing of monocular depth cues between strabismic and stereotypical observers. In the present study, we report on two experiments where we examine differences in the susceptibility to, and accuracy in the use of, one of the primary monocular cues to relative depth perception: perspective convergence (linear perspective).
reference link : https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7813253/
More information: Giedre Zlatkute et al. Unimpaired perception of relative depth from perspective cues in strabismus, Royal Society Open Science (2020). DOI: 10.1098/rsos.200955
