Visual Perception and Cognition in infancy

Carl Granrud, Carl E. Granrud

• Publisher: Psychology Press
• Year: 2013
• Print ISBN: 9780805807059
• eBook ISBN: 9781134757336
• Topic: Psychology
• Subtopic: Developmental Psychology
• Index: Psychology

The chapters in this book are based on papers presented at the 23rd Carnegie Mellon Symposia on Cognition. At this exciting event, speaker after speaker presented new discoveries about infants' visual perception in areas ranging from sensory processes to visual cognition. The field continues to make significant progress in understanding the infant's perceptual world. Several advances have come from the development of new methods for exploring infant perception and cognition that have brought new empirical findings. Advances have also been made in understanding the mechanisms underlying perceptual development. Outstanding examples of this ongoing progress can be seen in the chapters of this volume.

1

Spatial and Chromatic Visual Efficiency in Human Neonates

Martin S. Banks

Elizabeth Shannon

University of California at Berkeley

The early stages of vision are primarily serial. Visual stimuli pass sequentially through the eyes’ optics, which are responsible for forming the retinal image; the photoreceptors, which sample and transduce the image into neural signals; and two to four retinal neurons, which transform and transmit those signals to the optic nerve and eventually to the central visual pathways. Considerable information is lost in these early stages of the visual process as evidenced by the close correspondence between the filtering properties of the optics and receptors, and some measures of visual sensitivity (e.g., Banks, Geisler, & Bennett, 1987; Coletta, Williams, & Tiana, 1990; Pelli, 1990). In this chapter, we examine how immaturities among these early stages of vision limit the spatial vision and the color vision of human neonates.

It is well established that human neonates see poorly. In the first month of life contrast sensitivity (a measure of the least luminance contrast required to detect a visual target) and visual acuity (a measure of the finest detail that can be detected) are at least an order of magnitude worse than in adulthood (reviewed by Banks & Dannemiller, 1987). Chromatic discrimination (the ability to distinguish targets on the basis of their wavelength composition) is much reduced, too (reviewed by Teller & Bornstein, 1987). One would think that the anatomical and physiological causes of such striking functional deficits would have been identified, but the specific causes are still debated. Some investigators have pro-posed that optical and retinal immaturities are the primary constraint (Jacobs & Blakemore, 1988; Wilson, 1988), whereas others have emphasized immaturities in the central nervous system (Bronson, 1974; Brown, Dobson, & Maier, 1987; Salapatek, 1975; Shimojo & Held, 1987). Our purpose is to establish as well as possible the limitations imposed by optical and receptor immaturities and to discuss how those limitations should be incorporated into our descriptions and theories of vision in early life. We conclude that much, but not all, of the spatial and chromatic deficits exhibited by neonates can be explained by optical and receptor immaturities. Immaturities among post-receptoral mechanisms are responsible for the unexplained portion of the deficits.

The approach we use relies on ideal observer theory. By definition, the performance of an ideal observer is optimal given the built in physical and physiological constraints (Green & Swets, 1966). Ideal observers have been useful tools in vision research because their performance provides a rigorous measure of the information available at chosen processing stages (Barlow, 1958; Geisler, 1989; Pelli, 1990; Watson, Barlow, & Robson, 1983). For instance, the performance (e.g., contrast sensitivity or visual acuity) of an ideal observer with the optical and photoreceptor properties of an adult eye reveals the information available at the receptors to discriminate various spatial and chromatic stimuli. Similarly, comparing the performance of two ideal observers with different optics and receptors reveals how changes in those properties affect the information available for visual discrimination. In this sense, ideal observer analyses allow an atheoretic assessment of constraints imposed by various stages of visual processing.

Unlike more conventional neural theories that assume specific neural mechanisms at different ages (e.g., Bronson, 1974; Salapatek, 1975; Wilson, 1988), we attempt to reduce assumptions to a minimum. An ideal observer is derived here with the optical and photoreceptor properties of the human neonate (see also Banks & Bennett, 1988). The properties of the young fovea are now understood well enough to minimize the number of necessary assumptions. The performance of this ideal observer is then computed for various spatial and chromatic tasks. Its performance is the best possible for a visual system with the front-end of the newborn system. Moreover, comparisons of the performance of newborn and adult ideal observers reveal the information lost by optical and photoreceptor immaturities across the life span. We will show that much, but not all, of the deficits human neonates exhibit in contrast sensitivity, grating acuity, and chromatic discrimination can be understood from information losses in the optics and photoreceptors.

OPTICAL AND PHOTORECEPTOR DEVELOPMENT

This section briefly reviews the literature on the development of the eye and its optics and on the development of the photoreceptors. More detail is provided in Banks and Bennett (1988).

The eye grows significantly from birth to adolescence, with most of the growth occurring in the first year. For instance, the distance from the cornea, at the front of the eye, to the retina, at the back (the axial length), is 16 to 17 mm at birth, 20 to 21 mm at 1 year, and 23 to 25 mm in adolescence and adulthood (Larsen, 1971; Hirano, Yamamoto, Takayama, Sugata, & Matsuo, 1979). Shorter eyes have smaller retinal images. So, for example, a 1 ° target subtends 204 µ (microns) on the newborn's retina and 298 µ on the adult's (Banks & Bennett, 1988). Thus, if newborns had the retina and visual brain of adults, one would expect their visual acuity to be about two thirds that of adults.

Another ocular dimension relevant to visual sensitivity is the diameter of the pupil. The newborn's pupil is smaller than the adult's, but this difference probably has little effect, if any, on visual sensitivity because the eye is shorter, too. More specifically, the eye's numerical aperture (the pupil diameter divided by the focal length of the eye) is nearly constant from birth to adulthood (Banks & Bennett, 1988); so, for a given target, the amount of light falling on the retina per degree squared should be nearly constant across age.

Still another ocular factor relevant to visual sensitivity is the relative transparency of the ocular media. Two aspects of ocular media transmittance are known to change with age: the optical density of the crystalline lens pigment and that of the macular pigment. In both cases, transmittance is slightly higher in the young eye, particularly at short wavelengths (Bone, Landrum, Hime, Fernandez, & Martinez, 1987; Werner, 1982). Thus, for a given amount of incident light, the newborn's eye transmits slightly more to the photoreceptors than does the mature eye.

The ability of the eye to form a sharp retinal image is yet another relevant ocular factor. This ability is typically quantified by the optical transfer function. There have been no measurements of the human neonate's optical transfer function, but the quality of the retinal image almost certainly surpasses the resolution performance of the young visual system (see Banks & Bennett, 1988, for details). Thus, it is assumed here that the optical transfer function of the young eye is adult-like. This assumption is not critical for our purposes because moderate changes in optical quality would not affect the main conclusions of this chapter.

If optical imperfections do not contribute significantly to the visual deficits observed in young infants, receptoral and postreceptoral processes must. The retina and central visual system all exhibit immaturities at birth (for reviews, see Banks & Salapatek, 1983; Hickey & Peduzzi, 1987), but recent work has also found striking morphological immaturities in the fovea, particularly among the photoreceptors.

The development of the fovea in the first year of life is dramatic, but subtle morphological changes continue until at least 4 years of age (Yuodelis & Hendrickson, 1986). To illustrate some of the more obvious developments, Figs. 1.1 and 1.2 display Yuodelis and Hendrickson's micrographs of retinas at different ages. Figure 1.1 shows low-power micrographs of the fovea at birth, 4 years, and adulthood. The black lines and arrows mark the rod-free portion of the retina, the so-called foveola. The diameter of the rod-free zone decreases from roughly 5.4° at birth to 2.3° at maturity.

FIG. 1.1. Low-power micrographs of the human fovea at birth (1), 45 months (2), and 72 years (3). The black lines and arrows mark the width of the rod-free foveola. Because at birth the foveola is very wide, only half of it is shown. The most central cone synaptic pedicles are indicated by a white arrow in (2) and (3). P, photoreceptor nuclei; G, ganglion cell layer; OS, outer segment of photoreceptors; PE, pigment epithelium. From Yuodelis and Hendrickson (1986).

Figure 1.2 displays, at higher magnification, the human foveola at birth and adulthood. An individual cone is outlined for clarity in each panel. The cones’ outer segments are labelled OS. The inner segments are just below the outer segments.

In the mature cones, the inner segment captures light, and through waveguide properties, funnels it to the outer segment where the photopigment resides. As the light travels down the outer segment, many opportunities are provided to isomerize a photopigment molecule and thereby create a visual signal.

The micrographs of Fig. 1.2 illustrate the striking differences between neonatal and adult cones. Neonatal inner segments are much broader and shorter, and, unlike their mature counterparts, they are not tapered from the external limiting membrane to the interface with the outer segment. The outer segments are distinctly immature, too, being much shorter than their adult counterparts.

In order to estimate the efficiency of the neonate's lattice of foveal cones, we calculated the ability of the newborn's cones to capture light in the inner segment, funnel it to the outer segment, and produce a visual signal (Banks & Bennett, 1988). We began by estimating the effective collecting area of cones at different ages. We found that the newborn inner segment cannot funnel light to the outer segment properly: The inner segment is so short and broad that light rays approaching and reflecting off the inner segment wall at acute angles cannot reach the outer segment.

FIG. 1.2. Development of human foveal cones illustrated by light micrographs. A single cone is outlined in both figures; magnification is constant. Ages: (6) = 5 days postpartum; (9) = 72 years. PE, pigment epithelium; OPL, outer plexiform layer; M, Muller glial cell processes; CP, cone pedicles; OS, outer segments. From Yuodelis and Hendrickson (1986).

If the funneling property of the inner segment does not work, the effective aperture or collecting area of newborns’ cones must be the outer segment. Taking the smaller size of the newborn eye into account, the angular diameter of the effective collecting area is about 0.35 minutes of arc. The dimensions required to compute this value are given in Table 1.1. The effective aperture of adult foveal cones is, of course, the inner segment because its funneling properties are rather good. Thus, the aperture of adult cones is about 0.48 minutes.

Calculations were also made of the average spacing of cones in the newborn and adult fovea from Yuodelis and Hendrickson's data (see Banks & Bennett, 1988, for details). Table 1.2 shows the cone-to-cone distances in minutes of arc. Cone-to-cone separation in the center of the fovea is about 2.3 min, 1.7 min, and 0.58 min in neonates, 15-month-olds, and adults, respectively. It is very important to note that these lattice dimensions impose a limit on the highest spatial frequency that can be resolved without distortion or aliasing (Williams, 1985). This highest resolvable spatial frequency is called the Nyquist limit. From cone spacing estimates, Nyquist limits of 15, 27, and 60 c/deg were calculated for newborns, 15-month-olds, and adults, respectively. Adult grating acuity is similar to the Nyquist sampling limit of the foveal cone lattice (Green, 1970; Williams, 1985). One naturally asks, then, whether a similar relationship is observed in newborns. The answer is evident from a comparison of Table 2.2 and Fig. 1.7. Although newborn Nyquist limits are much lower than adult, they are not nearly as low as the highest grating acuity observed early in life. Thus, in human newborns the Nyquist sampling limit of the foveal cone lattice far exceeds the observed visual resolution, implying that coarse sampling by widely spaced receptors is not a major cause of low acuity in newborns.

TABLE 1.2
Nyquist Limits

	D	Nyquist Limit
Neonate central fovea	2.30 min arc	15.1 c/deg
Neonate foveal slope	1.66 min arc	20.9 c/deg
15 month central fovea	1.27 min arc	27.2 c/deg
Adult central fovea	0.58 min arc	59.7 c/deg

We used the information listed in Table 1.1 to construct model receptor lattices for newborns and adults; these are shown in Fig. 1.3. The white bars at the bottom of each panel represent 0.5 minutes of arc and serve as references. The light gray areas represent the effective collecting areas: the cone apertures. Nearly all of the light falling within these apertures reaches the photopigment and is, therefore, useful for vision. The effective collecting areas cover 65% and 2% of the retinal patches for the adult and newborn foveas, respectively. Consequently, the vast majority of incident photons are not collected within newborn cone apertures and are, therefore, not useful for vision.

The next factor to consider is how efficiently the outer segment—where the photopigment resides—absorbs photons and produces the is...

Table of contents

• Front Cover
• Half Title
• Carnegie Mellon Symposia on Cognition
• Title Page
• Copyright
• Contents
• Preface
• 1 Spatial and Chromatic Visual Efficiency in Human Neonates
• 2 Motion Nulling Techniques and Infant Color Vision
• 3 What Can Rates of Development Tell Us About Underlying Mechanisms?
• 4 Perception of Visual Direction in Human Infants
• 5 Kinematic Foundations of Infant Visual Perception
• 6 Infants’ Perception of Biomechanical Motions: Intrinsic Image and Knowledge-Based Constraints
• 7 Infants’ Sensitivity to Motion-Carried Information for Depth and Object Properties
• 8 Future-Oriented Processes in Infancy: The Case of Visual Expectations
• 9 The Object Concept Revisited: New Directions in the Investigation of Infants’ Physical Knowledge
• 10 Commentary: Extending the Ideal Observer Approach
• 11 Commentary: Cheers and Lamentations
• Author Index
• Subject Index