New thoughts on ageing
Over the past few years the visual changes that come with ageing have become a priority research topic. There are many reasons for this. Firstly, in industrialised countries the extension of life expectancy, happening in parallel with a fall in the birth rate, is leading to a considerable rise in the average age of the population. The number of people aged over 65 is set to almost triple by 2050, to reach some 14% of world population and 20% of the population in industrialised countries . Moreover, technological developments will permit the treatment of disabling afflictions with increasing safety, allowing for the use of more interventionist methods on elderly people.
These developments offer the means, as well as leading to the demand, for a better understanding of the difficulties encountered by elderly people. Not to mention the economic benefits of treating their visual problems, since regained autonomy results in so much reduction in the financial burden on society in terms of dependency.
The consideration currently given to elderly people is part of renewed social thinking with regard to the most vulnerable individuals, who have very little voice. Children and young and middle aged adults with disabilities will also benefit from it.
A study of ageing of the visual system meets with a variety of methodological difficulties. Where to position the limits of normality? To what extent should senile crystalline lens and macular reorganisation be considered physiological? Particularly since most functional problems in ageing increase in the presence of age-related macular degeneration, cataract or glaucoma, and these are disorders that affect nearly one in five people aged between 65 and 74 years [19, 07]. Some authors have suggested criteria in order to attempt to define the limits of normality .
Alterations of the visual message
The clinical expression of age-related changes is highly diverse. Its severity is moreover extremely variable from one individual to another, ranging from a complete absence of clinical symptoms through to a disabling condition. Surprisingly, ageing is even sometimes accompanied by an improvement in some of the processes that contribute to image processing in the brain.
Retinal illumination is reduced in elderly people, notably due to the narrowing of the pupil, by around 2 to 2.5mm between the ages of 10 and 80 , and the increased filtering of light rays when they cross the cornea and the crystalline lenses, particularly for short wavelengths (Fig.1) . The reorganisation of retinal and post-retinal neurons also contributes to weakening of perception of the luminous signal. Neuron density is reduced in the retina , lateral geniculate bodies and the V1 occipital area . Also, the metabolism of those nervous cells that are spared is disturbed, for example that of retinol in the photoreceptor pigment epithelium complex , and of the cholinergic modulation of the cortical neurons .
Fig. 1: Reduction in retinal illumination with age, due to narrowing of the pupil and crystalline lens filtering. Reduction is considerable and becomes established gradually, throughout life. 
Intra-ocular reduction of the luminous flow is all the greater since elderly people are only exposed to outdoor light (> 1000 Lux) for half the amount of time that young people are , and still less for people living in institutions .
The reduction of luminosity is harmful for several reasons: firstly, of course, for appreciation of the details of an image (colours, spatial and temporal contrasts, …) but also for the regulation of neurovegetative and immune mechanisms, or for cognitive and thymic functions, all of which are regulated by light. In order to work in satisfactory conditions, over the age of 50 many people need twice as much light as their young colleagues . Also, 30% of people over the age of 65 suffer from depression  and almost half from insomnia ; these disorders are commonly associated with a reduction of light.
Sensitivity to spatial contrasts is, for its part, relatively preserved in elderly people in daylight conditions; it deteriorates, however, the when the image is presented at high temporal frequency, due to the reorganisation of receptors and post-receptors . In semi-darkness, a deficiency is observed in particular in low spatial frequencies, probably related to magnocellular damage . Sensitivity to movement, as well as an ability to define direction, are also affected, mainly due to nerve alterations . Colour distinction is reduced, for values close to the perceptive threshold, particularly in terms of blue/yellow, due to spectral changes to the luminous flow caused by crystalline lens problems and receptor and post-receptor reorganisations; nerve regulation mechanisms do however compensate for the spectral imbalance and retain satisfactory colour recognition . Adaptation to darkness is delayed and less extensive, due to miosis and metabolic disturbances to the rods .
Structural and functional transformations occurring with age do in fact have complex effects, which can themselves interact, sometimes in synergy and at other times contrary to one another. Thus, pupil narrowing reduces the light flow all the more since rays cross the crystalline lens through the central portion only, which is the thickest, precisely where filtration is strongest. Pupil narrowing does have some advantages too, though, because it attenuates the visual problems resulting from optical aberrations and diffraction caused by crystalline lens opacities, and therefore the glare they generate. Similarly, nuclear sclerosis of the crystalline lens provides retinal tissues with better protection against ultraviolet rays.
Higher visual functions. Facilitating, suppressive and compensatory mechanisms
Integrating processes and semantic recognition capacities are generally not altered with age . Some processes are less efficient however and, conversely, some are optimised .
One of the main processes in cerebral image processing enables the identification of objects located in an environment with visual noise, known by Anglo-Saxons as “figure-ground discrimination”. This is based mainly on two mechanisms, one a facilitator and the other a suppressor. The first integrates into a single image the fragmentary elements of its presentation , whereas the second attenuates the distracting elements present within the visual context  (Fig. 2). With ageing, the facilitating mechanisms are preserved but the suppressor mechanisms would appear to be less efficient, which would lead to a reduction in the ability to ignore irrelevant information and a weakening in the power of concentration.
Paradoxically, the weakening of suppressor mechanisms is beneficial in some respects. In fact it optimises the memorisation of images that are apparently considered to be insignificant and thus enables elderly people, more than young people, to remember the detail of scenes observed previously . The weakening of these mechanisms in elderly people also appears to facilitate the integration of multisensory information (see below).
In the image identification and extraction process, the integrator mechanism is performed by long horizontal intercellular links in the V1 cortical area ; the second, the suppressor, depends on short GABAergic inhibitor fibres .
Fig 2. Image identification. On the left Gabor stimuli evoke a fragmented outline of a circle. On the right these stimuli are presented in an environment with visual noise. Identification of the circle involves an initial process of integrating the fragments of the outline and a second process of perceptive suppression of the detractors. 
Nerve networks in the brain restructure themselves in response to the deterioration of certain sensorial functions . When visual perception tasks are performed, occipital activation tends to decrease with age whereas, in parallel to this, activation of the areas associated with “top-down” control functions intensifies, more particularly in the frontal cortex, . Analysis of cerebral connectivity also shows in elderly people a reduction of the median occipital signal and an accentuation of that of the (pre-)frontal areas in particular (Fig. 3) . This reorganisation enables a compensatory recruitment of nervous structures that integrates the endogenous data analysis process .
Fig 3. Evolution of functional connectivity with age (functional cerebral MRI). The blue pixels indicate reduced activity in older people, the yellow/red pixels increased activity. IC28: internal occipital cortex (visual); IC29: frontopolar cortex, dorsomedial prefrontal and dorsolateral, and (para)anterior cingulate cortex (executive functions). 
Multisensory integration and visual dominance
Our perception of the world is multimodal by its very essence, because objects emit signals of various types, visual, audio and olfactory, for example. One might expect that this diversity of sensorial attributes could cause confusion. And yet our perceptive organisation enables us to integrate all these parameters within a harmonious spatiotemporal space, using a complex, continually adapting process .
The integration of multisensory information accelerates the recognition of objects, as much in young as in older people. This faculty is more pronounced in elderly people  and could act as a strategy to compensate for the deterioration of the functions of each sensorial mode . It is probable that in elderly people the attenuation of suppression mechanisms referred to above is actually beneficial to the integration of multimodal afferents . But it could be that the optimisation of multisensory integration acts to the detriment of precision of analysis . In elderly people this optimisation is associated with increased activity in the inferior parietal and prefrontal median lobes .
When visual information is in conflict with that from another sensory mode, or if it gives rise to contradictory interpretations, visual information usually predominates . This is known as visual dominance. This dominance would appear to be more pronounced in elderly people than in young people [18, 14]
Limitations of current studies and exemplarity of the visual function
An important parameter is still missing in our assessment of perceptive functions in elderly people: that of the contribution made by past experience in the analysis of information and its probable compensatory role for sensory changes. Current studies are for the time being mainly directed towards the recognition of elementary shapes and the performance of low level tasks. In everyday life, however, analysis processes frequently involve the experience of the person involved, even when it is a matter of recognising the presence of a simple object in a natural environment with visual noise. As we have said, studies undertaken on functional imagery in elderly people do show during visual tasks an increased part played by the cerebral structures that integrate into information processing a “top-down” contribution of endogenous data. This work prefigures the development of more elaborate behavioural research, which will better reflect the aptitudes of elderly people in the processing of visual information.
The visual system is a preferred model for the study of information processing. In the past, vision research often led to recognition of functional organisation that was to prove common to various modalities. Over and above its sensorial context, the study of the visual system offers exceptional access to the understanding of the cerebral processes that govern our interactions with the world around us. Observations made on the ageing of vision, and on the development of the compensatory mechanisms that accompany it, provide us with important information in terms of the knowledge of the processes linked to ageing.