Is vision developed from scratch?
It would seem that infants develop vision rather than learn how to see. Babies are born with considerable ability to see and although they lack the visual acuity of adults, they are able to detect a great many things. An example of this is that a baby is able to copy facial expressions, such as sticking out its tongue, within a few hours of being born. During this short time, it cannot have “learnt” how to see such things. Nor can it have learnt how to see prior to birth since its only environment is the darkness of the womb. At this point, the baby is taking in what its visual receptors provide it with. It does not need to learn how to use these – they just work outside of our control (we cannot, for example, switch our photoreceptor cells off).
It would therefore be wrong to think of learning how to see as developing an ability from scratch. It clearly is not, as babies are born sighted with no means of learning to see prior to birth. There could conceivably be a case for learning how to use their intrinsic visual abilities in much the same way as learning to walk – a child will just do it somewhat spontaneously. However, the evidence does not fully correlate with this proposition.
Development of visual acuity
Firstly, the improvement in visual acuity (detected by preferential looking experiments involving a grating and a plain surface) is unlikely to be due to an infant learning better use of its eyes. More likely it is due to the development of the receptive fields of its retina – the more receptive fields, the greater the acuity. A further probable reason is the development of lateral inhibition in the receptive fields, which is essential to the resolution of images. The development of greater acuity could also be due to the visual cortex being immature in that it has yet to develop the great abundance of interconnections between neurons which is found in more mature visual cortices. It is the number of interconnections that allows for the finer structuring together of receptive fields, and hence the ability to deal with finer details and subtler contrasts. Garey and De Courten (1983) found that these interconnections increase dramatically in number between the ages of two and six months. This is an important finding as it could be a reason for the improvement in an infant’s vision beyond acuity.
How innate is colour discrimination?
There is evidence to show that infants are innately able to detect colours. At the very least, their cone cells (the colour-sensitive cells of the retina) are intrinsically sensitive to certain wavelengths of light and this sensitivity is therefore something that cannot be, and is not, learnt. What is learnt is the interpretation of colour. We are taught that a particular colour is red, or blue, or whatever – they are not innate distinctions. This can be seen in different interpretations of colour between cultures. Some cultures see green and blue as one and the same thing, while in Russia they treat “light blue” and “dark blue” as separate things. What remains the same is the intrinsic, perhaps developed, but definitely not learned, biological basis of distinguishing wavelengths of light.
Development of depth perception
Infants also develop binocular vision rather than learn it. A review by Braddick and Atkinson (1983) of many experiments on this aspect of vision has shown that binocular vision is first found at around 12 and 17 weeks after birth. This is due to neurological development in that prior to gaining binocular vision, an infant has cortical neurons which respond only to input from one eye only, whereas later they develop neurons that require correlated input from both eyes for optimal stimulation. So binocular vision is not learnt but biologically developed, however, it could be argued that an infant’s development of stereoscopic depth perception is a learned ability. This would involve the infant learning, perhaps by using a trial and error-like method, how to interpret binocular disparity. But there is evidence to support the view that there is an element of learning in binocular vision. This is the fact that strabismics (strabismus is a condition where the eyes cannot be brought parallel, also known as a "squint") still develop binocular vision as they accommodate with their misaligned eye. But infants do appear to have some innate depth perception as they prefer objects with depth to flat objects. They also try to grab objects, which implies they have some spatial awareness.
Eye movements
Infants also seem to have innate oculomotor (eye movement) abilities which are far more mature than their other motor abilities. This is seen in the relatively advanced saccadic movements of infants eyes when attempting to foveate (or, fixate) a target object. However, it has been found that it takes a one year-old a sequence of repeated saccades to foveate an object of interest rather than the single saccade of a three month-old and above (Aslin & Salapatek, 1975). It has also been demonstrated that only after two months of age can an infant track a smoothly moving object without having to use a series of saccadic jumps (Aslin, 1985). It is thought that these effects are due to an underdeveloped retina and not a lack of experience of eye movement.
Further developed aspects of vision
There is further evidence that suggest that visual ability is not learned. Meridional amblyopia is a condition where a severe astigmatism (a refractive error which causes different orientations of lines to be out of focus at the same distance, due to a more cylindrical shape to the eye) in childhood means that the retina has never received any well-focussed contours in a particular orientation. Acuity for such lines is found to be poor in later life even if the astigmatism is corrected so that sharp images can be attained. This shows that acuity for these lines cannot be learned, but must instead be developed at an early age during an infant’s period of plasticity, or the time when neural connections form at an optimal rate.
Colin Blakemore’s experiments with kittens also show that the development of vision id dependant on environment. Blakemore raised kittens in an environment consisting of stripes of one particular orientation and after a certain period, the kittens were unable to detect stripes of the opposite orientation. This shows that despite an innate ability to detect all orientations, without exposure the receptors specialising in particular orientations become redundant. This is due to filtering out irrelevant sensory information. The kittens did not really learn to see only stripes of a certain orientation, rather they lost the ability to see other stripes.
Innate object awareness
Children do not appear to learn other aspects of vision such as object constancy. Preferential looking experiments have shown that infants recognise that an object is the same from all angles. This appears to be another innate ability.
Conclusions
It would appear that the development of vision is very largely biologically based. It is the development of more complex neurophysiology in the visual system of infants that allows them to gain better vision. Of course, there is an element of learning insofar as an infant will have to on some level “discover” how to use their developing abilities. This so-called learning is not necessarily an active sort of learning, but more passive, going along with the development of the visual system. But still, the ability to see is developed rather than learnt.
Based on the lectures of Prof. Oliver Braddick, Head of Department of Experimental Psychology, University of Oxford.