Handbook of Language and Literacy Development - a Roadmap from 0 to 60 Months

children image


Vision Development (0 - 6 Months)click to print Print
Research Review / Parent

Written by: Shankar, S., Robertson, BA., and Bobier, W.R., University of Waterloo

Several key functions of the visual system characterize what is referred to as "seeing" and these include:

  1. The detection of contrast through changes in brightness (visual acuity) or wavelength (colour vision);
  2. The ability to perceive motion and integrate uniocular retinal images into a binocular percept. These sensory processes involve both cortical and sub-cortical pathways.
  3. In addition, an intricate oculomotor system is required to align, focus and stabilize the eyes in order that critical objects in the external world are imaged on the small but highly sensitive ("visually rich") foveal area of the retina.

This research review defines the development of these functions and their underlying neuroanatomy during the first 6 months of life. The rudiments of these functions are present at birth but the full maturation of each takes months and often years. However, the rate of this maturation declines with age so that the first 6 months of life show tremendous development of the visual capacity of the human infant. This development of vision is also indirectly supported by significant neurophysiological change.

Physiology of Vision Development

The underlying anatomical changes pertaining to vision are understood from animal models, mostly primate, and a few sparse findings taken from human infants. While anatomical changes can predict to some degree development of specific visual function it is equally true that development of visual functions also predict anatomical change. In its broadest term, the change in vision in the first 6 months of life appears to reflect increased cortical control of neural circuits which were primarily sub-cortical at birth . This development is also heavily affected by visual experience.

At Birth

Light sensitive cells of the retina (photoreceptors) are divided into rods and cones. The rods are designed more for light detection rather than visual discrimination. The cones are designed for visual resolution and in the adult are found more densely packed in the foveal region of the retina. This area only subtends close to 1 degree.

At birth, the retinal neurons (photoreceptors, bipolar cells, amacrine cells, horizontal cells and ganglion cells) and their connection to the brain (visual cortex) are in place. The primary visual pathway involves the retinal ganglion cells, whose axons form the optic nerve; the lateral geniculate body (LGN) and subsequent geniculo–cortical pathways. In the adult the synapses along this pathway coupled with intra-retinal circuits serve to sharpen the retinal image which is spatiotopically projected to the visual cortex and beyond. This capacity is very rudimentary at birth. The foveal area is anatomically recognizable at 24 to 26 weeks of gestation. However it is still immature at birth. The cone photoreceptors at birth are short and poorly developed, increasing both in length and numbers subsequently . The cones in the retinal periphery are also immature at birth, but the difference is less striking than that observed in the fovea . Rod and cone receptors in the mid-periphery are more mature than foveal receptors . Given this, many contend that newborn infants may use areas in their peripheral retina for visual functions, such as focusing and eye alignment rather than using the central retina like adults do. The LGN organization is immature with smaller neurons and poorer layered organization, as compared to the adult LGN. In addition, the myelination of neuronal axons in the visual pathway and the differentiation of axonal and dendritic processes are not complete. Many primary visual functions appear to be driven by spatiotopic links to the superior colliculus, another sub-cortical structure, which matures earlier than the visual cortex. However, its ability to provide detailed spatial resolution is extremely limited.

From Birth to 6 Months

The retinal area continues to expand, a process which started in gestation. Individual retinal cells change in their width and length, as well as spacing. Cell development is centrifugal being initiated at the foveal area . The foveal receptors complete their differentiation into cone-only design around about 3 months of age (2-4 months). However, other changes to the morphology of the cones and their packing in the fovea continue beyond this time until about 45 months of age. The visual cortex and retino-geniculate pathway grow towards maturity but remain immature. Cells in the LGN show changes in cell morphology and spatiotopic layering while the visual cortex does not show large changes in its cell numbers during this time. Pathways show increased myelination and thickening of fibre diameters thereby increasing the rate and quality of neural transmission of visual signals.

Conceptually, the rapid development of visual functions, to be described below, suggests that the postnatal neural development reflects more than just a general improvement in the efficiency of a system, all of whose components are to some degree functional from birth. Rather, it reflects the emergence of a series of new sensory capabilities as a result of increasingly more sophisticated neural networks which allows for better processing of the visual information.

Development of Sensory Functions of Vision

Three sensory functions of vision most evident during the 0-6 month period of visual development include acuity, colour and binocular vision.

Visual Acuity in a Black and White Environment and Related Functions

What we might refer to as "seeing" can be broken down into specific sensory functions all of which have been tested to some degree in human infants. Visual acuity is most commonly described by our ability to resolve letters on an eye chart, where typically dark letters are projected on a bright background. If, however, we define the letter E on an eye chart not in terms of its size but rather as a pattern of alternating light and dark bars where the smaller the letter the higher the frequency of this alternation we can expand our analysis of seeing to now include a contrast sensitivity function (CSF). Visual acuity is then the highest frequency of alternating light and dark patterns which we are capable of resolving. Now if the letter E is changed from black to grey we have reduced the contrast of the pattern and will reduce our ability to see it and hence our visual acuity.

"Seeing "can be measured by defining the highest spatial frequency we can resolve at a constant contrast (eye chart) or by noting the necessary level of contrast at a given spatial frequency (CSF). Visual acuity represents the highest spatial frequency that can be resolved at the highest contrast.

CSF and visual acuity can be tested in human infants by a subjective means known as Forced-Choice Preferential Looking (FPL) and by more objective means [Visual Evoked Potentials (VEP)]. The FPL method requires a "look" from the infant indicating that he/she has seen the object. In the VEP method seeing is correlated with evoked potentials taken from the surface of the visual cortex. Perhaps this explains why VEP typically shows better acuity scores compared to FPL for a given age. However, it is not certain that VEP can assume the infant actual sees the target.

The FPL technique is used to measure acuity in infants by capitalizing upon the infants' preference of patterned gratings over uniform gratings. Infants are presented with two panels (of equal luminance) one uniform grey and a second which has alternating white and dark bars defined at a specific spatial frequency (cycles/degree) and at a specific contrast. Over a series of trials, the contrast and spatial frequency are varied usually under computer control. If the infant is able to resolve the pattern presented to one screen, then an observer who can see the direction in which the infants gaze is directed, but does not know the location of the patterned grating will correctly identify the location of the grating.

The CSF measures the relative amounts of contrast needed over a range of spatial frequencies. Thus, as spatial frequencies of the target increase towards our acuity cut-off point, the amount of contrast required to see it increases. Interestingly, the function is not monotonic in that very low frequency targets require more contrast than mid range spatial frequencies due to characteristics of retinal circuitry. An iterative process is used to repeatedly present gratings of varying spatial frequencies to determine visual acuity and varying both spatial frequency and contrast to determine the CSF (Atkinson, 2000). Neonates have been shown to be resolve 1 cycle/deg at birth which increases linearly to 2-5 cycles/deg by 3 months of age . This is in sharp contrast to adult acuity which measures between 30-60 cycle/deg . For a standard eye chart to be visible for a newborn it would, in principle, need to be increased some 30 fold in size. Similarly the contrast required at all spatial frequencies is reduced during the first 3 months of life but again remains greater than that needed by an adult.

When VEPs are recorded using a stimulus which "sweeps" gratings of varying spatial frequencies rapidly before the newborn, visual acuity values are higher than with FPL . Visual acuity values were found to be 4.5, 5.5 and 6.7 cycles/degree for the first 3 months of life respectively, climbing to 12.6 cycles/degree at 6 months.

A realistic analogy would be as follows. If a mother were to have her newborn view the clock tower on Parliament Hill in Ottawa, the newborn could determine the outline of the tower but not the details of the clock itself. The outline of the tower would appear fainter to the newborn as well. During the first 6 months of life the outline of the tower would become more distinct and more details of the clock face would become apparent but not to the level of the mother.

Telling time does not hold the same importance for the infant as recognizing the face of a caregiver. However, the clock tower provides a basis to understand how infants look at their mothers. Newborns are thought to be able to discriminate face-like features from a scrambled orientation as young at 4-5 weeks (reviewed in . When presented with simple shapes oriented to look face-like, eye tracking data shows that infants under the age of 2 months fixate near the external contours of the face, referred to as the "externality" effect. After 3 months of age infants focus on internal features, such as the eyes. removed details from the external contour of infants' mothers' faces (the hairline) and 5 - 7 week old infants could no longer discriminate their mother's face from that of a stranger. Older infants were able to discriminate, presumably due to their ability to use internal features. The ability to "see" the internal features of a mother's face would be expected to be more difficult for the newborn given the higher spatial frequency demands vs. the outline of the face.

Colour Vision

Thus far we have defined seeing in a black and white environment. Colour vision is mediated by divisions of the cone pathways. Our adult level of colour vision (trichromacy) discriminates opposing wavelengths in much the same way as it detects light and dark. Two independent systems exist, one that discriminates green from red and a second perhaps weaker system which discriminates yellow from blue. Neither of these systems has been found to be present at birth. This would not be inconsistent with previously described immaturity of retinal cones at this time. Colour vision can be tested in much the same way as contrast sensitivity and visual acuity by using FPL. In this case the gratings vary not in overall luminance but in opposite spectral compositions,- red to green and yellow to blue. Such investigations have found that at birth colour discriminations are weak or absent. During the first two months red/green colour discriminations emerge with blue/yellow discriminations following so that trichromacy is evident at 3 months of age but such discriminations remain far coarser than adults.

Binocular Vision

Binocular vision is a hallmark of cortical processing. "Fusion" of the two separate retinal images occurs at the visual cortex. Sub-cortical structures such as the superior colliculus cannot provide this function. This cortical wiring leads to retinal points in one eye corresponding to that in the other. The fovea of each eye would be one important example of two points which must correspond if we are to see the world singly with both eyes. When images fall on non-corresponding points, cortical structures identify the disparate image with a sense of depth and the degree of offset (disparity) serves as a specific signal for eye realignment. This capacity is referred to as stereo acuity. Studies of the development of binocular fusion and stereo acuity have also been possible by using the techniques of FPL and VEPs . The FPL technique was confirmed as viable when it was found that infants in fact prefer the complexity of a fused image or one "in depth" (stereo acuity) over that of a single or flat image. Thus, the preferential looking behavior of infants seems to be related to stimulus complexity where gratings are preferred over blank screens and gratings seen in depth are preferred over those which are flat and two dimensional.

As would be expected from anatomical studies, binocular vision is not found at birth. Evidence from preferential looking and visual evoked potential studies suggest that binocular fusion is typically found around 3 months of age and the detection of gross stereo acuity follows very shortly (within weeks) from that . Interestingly, the sensitivity of stereo acuity increases rapidly once it is present and reaches maturity at a much faster rate than does visual acuity . Trajectories of stereo acuity taken in the early months of life suggest a capacity that would be adult-like by 6 months of age. This difference likely reflects the greater relative dependency of visual acuity on retinal development.

Motion Perception

The motion detection pathway projects from the ganglion cells to the magnocellular layers of the LGN, primary visual cortex and other extrastriate regions. The capacity for motion detection is present in young infants where they show a preference to moving over non moving stimuli . However, this does not necessarily prove that they can process motion effectively, as this finding might simply reflect sensitivity to temporal modulation. For example, infants readily show a visual preference for dynamic stimuli that do not move, such as full-field flicker.

The detection of changes in stimulus direction motion has been found at age of 10 weeks in infants using drifting random-dot patterns and VEP . Behavioural measures of directional sensitivity of both FPL and habituation (adaptation to FPL) methods show a slightly earlier onset at about 7-8 weeks. The range of speeds for which infants can discriminate oppositely directed movements increases with age from 8 weeks, but is velocity dependent where for example, velocities of 5 degrees/sec were detected upon average when infants were 74 days, but higher velocities (20 degrees/sec) were not found until 89 days.

Motion perception can stimulate two oculomotor responses. Smooth pursuit eye movements (SPEM) follow a moving object by matching the velocity of ocular movement to that of object movement. Pendular or sinusoidal movements are ideal in eliciting this response. Optokinetic Nystagmus (OKN) serves to stabilize the retinal image of a moving object, where a slow following response is then reset by a rapid saccade as the ocular excursion reaches orbital limits. Interestingly, 6-8 week old infants who showed no directional capacity were able to produce the appropriately directed OKN responses.

OKN responses can be found at birth are believed to be elicited through a subcortical pathway involving the nucleus of the optic tract (NOT) (pretectal nucleus) . This pathway however limits monocular responses where only patterns moving temporal to nasal will elicit an OKN response. The asymmetry starts reducing for low stimulus velocities at around 2 months and disappears by 5 months ; . The asymmetry is still present at 5 months for higher velocities . There is some evidence that this change reflects the maturation of cortically based pathways that connect to the NOT. However, VEP studies and studies on hemispherectomized infants ; suggest that the cortex may be involved in OKN at an earlier stage than is implied by the development of symmetrical responses, and that part of the OKN asymmetry is caused by directional asymmetries in cortical motion processing.

However, VEP studies and studies on hemispherectomized infants ; suggest that the cortex may be involved in OKN at an earlier stage than is implied by the development of symmetrical responses, and that part of the OKN asymmetry is caused by directional asymmetries in cortical motion processing.

Some studies have suggested that very young infants must use saccadic eye movements (see below) to track isolated targets, and that smooth pursuit eye movements (SPEM) are not seen before 8 weeks of age ;. Others have reported brief periods of pursuit in younger infants.Typically, young infants show a cogwheeling effect where visual pursuits fail to match the velocity of the external object and must be augmented by "catch up saccades".

A rapid improvement in smooth pursuit performance apparently occurs between birth and 3-4 months of age. Older infants show more smooth pursuit, fixate moving targets more accurately, and track faster-moving targets more readily than younger infants. As such, it is likely that the mechanisms mediating smooth pursuit per se are intact and functional early on, but that the development of the prevalence and accuracy of smooth pursuit may also be affected by immaturity elsewhere in the visual system . Especially relevant in this regard is that visual input to the superior colliculus is generally based on the magnocellular pathway of the primary visual system. This pathway, which largely mediates visual sensitivity to motion, develops quite slowly relative to the other (parvocellular) visual pathway. In addition, suggested that the accuracy of smooth pursuit is affected by interactions with other attentional functions mediated by frontal areas.

Oculomotor Development

The oculomotor system provides a number of critical functions: gaze stabilization, rapid voluntary changes in fixation; pursuit of moving objects and linked changes in focus and eye alignment in order to clearly view objects which approach or recede. Two of these systems OKN and SPEM have already been discussed, given their intricate relationship with motion perception.

Gaze Stabilization and Fixation

Gaze stabilizing reflexes [vestibular ocular reflex (VOR), and optokinetic nystagmus (OKN)] are driven through sub-cortical pathways. VOR has one of the fastest speeds of transmission in order that movements of the head do not smear retinal imagery, this system need to be present at birth. The ability to change fixation (saccades) is present at birth but undergoes some maturation modifications where the amplitude of the saccade changes from multiple copies of a constant hypometric amplitude to an adult-like pattern of two movements. The adult-like movements include an initial movement covering 90% of the range followed by a second movement in the same direction covering the final 10% of the movement. The pattern of multiple hypometric saccades remains beyond 3 months and does not decline fully until around 6 months of age.

The ability of the eyes to pursue a moving object is dictated to a large degree by foveal development. Thus, as discussed above, SPEM undergoes considerable maturation during the first months of life. As discussed, saccades often combine with SPEM to allow refixation of moving object in SPEM fails to sustain foveation.

Ocular Alignment (Vergence) and Focus (Accommodation)

A change in fixation from the distant horizon to a toy within grasp requires a change in binocular alignment (convergence) and an increase in ocular focus (accommodation). Neonates show a rudimentary ability to change alignment but only over a very limited range of a few degrees . Ocular alignment ability changes in its range and accuracy over the first 3 months of life.

Accommodation is also limited and almost fixed in newborns (but not to any particular distance) and similarly shows significant increases to adult-like ranges during the first 3 months of life . However accommodative accuracy is better when objects are close (75cm) vs. distant (150cm) until the infant reaches 6 months of age. This has been attributed to a progression of visual attention (see below) from near to far objects . Cross-links between vergence and accommodation which allow the two systems to act in synchrony are present in the first three months of life, and given their sub-cortical makeup may be in place at birth.

Studies generally indicate that the development of vergence and accommodation are heavily dependent upon maturation of their sensory processes and pathways of stereo acuity and visual acuity. Thus, as visual acuity develops, so does the infant's capacity to improve the contrast of image by changing focus (accommodation) of the eye. Similarly improved stereoacuity leads to better ocular alignment which in turn allows improved stereoacuity.

Visual Attention

Visual attention is the cognitive process of selectively concentrating visually on a subset of incoming information to process the sensory environment. It is thought of as an enhancement of visual processing in a location that is attended. Whether a neonate makes the appropriate response to a visual target is limited not only by the development of the visual functions described above, but also by the development of visual attention. As in the development of visual functions, the neural mechanisms sub serving visual attention show a maturation that reflects increased neural networking that becomes more cortical with time. The brainstem reticular activating systems have been linked to arousal and attention. Four 'attentional' pathways are thought to ascend from the brainstem to neocortical areas. In some cases the pathways serve to engage visual attention to stimuli in particular spatial locations while other pathways serve to disengage attention. In all the thalamic pulvinar, superior colliculus and areas of the occipital, parietal, inferior temporal and frontal cortex are involved.

In infancy there are at least three important postnatal periods of development with respect to visual attention . The first is from term to 2 months, when the development of the alert state takes place. The next period takes place from 2 or 3 to about 6 months, when there are rapid changes in spatial orienting and attention to objects. The third period is from 5 or 6 months and beyond, during which significant changes in endogenous attentional function takes place. conducted a meta-analysis of look duration from infant habituation and fixation paradigms, and concluded that the look duration follows a "triphasic" or cubic, developmental course. That is, look duration increases from birth to 8-10 weeks postnatal, drops from 3 to about 6 months, and then begins a gradual increase that extends from about 6 months through the second and third years. In very young infants, alertness is more readily initiated by exogenous events or by lower-level mechanisms of arousal than by endogenous (e.g. volitional) events. It may be that the emergence of the alert state in very young infants (e.g., prior to 3 months postnatal) is because of the ascending influence of subcortical pathways on cortical targets. Attention in infants has been studied by various methods including physiological measures (heart rate changes and ERPs). Others have assessed the attentional control of eye movements and measured orienting to new stimuli in the periphery, changes in scanning eye movements within the features of an object and changes in accommodation for targets at different distances.

Shifting of visual attention to different objects does not appear to have a particularly early emergence. Also, disengagement of attention, which requires the infant to cancel attention to one object and shift it to another is also present in a rudimentary form very early in life, but appears to show considerable improvement from 2 to 4 months . The newborn or 1 month old infant's attentional field appears to be relatively limited in lateral extent to about 20-30 degree, this being the visual angle over which a stimulus can elicit a foveating saccade. Infants in the first three months of life appear to have difficulty in disengaging attention from a foveally viewed target. They appear to sustain attention on the first target they fixate and do not easily shift their gaze to a newly appearing target.

Summary: The Six Month Old Infant

Rapid changes have occurred in the sensory and motor aspects of vision in the first six months of life. The 6-month-old infant enjoys rudiments of all critical aspects of adult vision. They see colour, form and movement. Information is being processed and integrated binocularly. They are able to fixate and follow objects in their environment. They can recognize their caregivers, fixate on a novel stimulus and supply sufficient focus to visually examine items within their grasp. These changes reflect growth in the visual function itself as well as improved visual attention. Motor development is close to adult levels where responses are limited by sensory development (such as visual acuity) and visual attention.

Shankar, S., Robertson, B-A., & Bobier, W.R. (2007). Research Review: Vision Development (0 - 6 Months). In L.M. Phillips (Ed.), Handbook of language and literacy development: A Roadmap from 0 - 60 Months. [online], pp. 1 - 9. London, ON: Canadian Language and Literacy Research Network. Available at: Handbook of language and literacy development