Two reviews on visual rehabilitation appear in this issue of the Journal of Neuro-Ophthalmology. One deals with perceptual aspects (1) and the other with the motor aspects (2) of vision. Both conclude that high-quality evidence for their effectiveness is lacking. Part of the reason may be that to get good answers, one must ask good questions. Given the complexity of the visual system, and the limited insight we have in its mechanisms beyond the eye, asking the most appropriate questions is difficult. To provide a high-level overview of the issues involved, it is useful to first ask 2 seemingly simple questions: “What is vision?” and “What is vision rehabilitation?”
When asked “What is vision?,” a common answer is that it refers to the ability to recognize letters on a visual acuity chart. On further reflection, however, this answer appears to be too narrow (pun intended). When looking through a cardboard tube, one can easily read the chart, but will have problems navigating the room. It is obvious that foveal letter chart acuity vision describes only one aspect of total visual functioning.
The next question is “What is rehabilitation?” Different people may answer this question differently. Some may consider any intervention that improves function as rehabilitation. We find it helpful to make a distinction between traditional medical/surgical care and rehabilitation. The condition of patients with a deficit may be compared with a glass that can either be described as half empty or half full. Medical and surgical interventions aim at restoring what has been lost. Rehabilitation works with what remains, using it as the basis for building new potential. Comprehensive care must involve both approaches. Medical and surgical interventions address ocular problems. Rehabilitation takes advantage of the flexibility and trainability of the cerebral components of the visual system to enhance sensory processing.
Some examples may clarify this distinction. Cataract surgery restores lost transparency of the optical media; glaucoma treatment (medical or surgical) restores the lost balance between the production and outflow of aqueous fluid. Rehabilitation is different. Magnification devices do not repair the retina; they provide a larger image that can still be recognized by the brain, even if retinal detail is reduced. Talking books use intact auditory input to replace defective visual input. In all these cases, the goal is to achieve the desired functional output by modifying the input. One may compare this to a cell phone connection; as the user moves from cell tower to cell tower, the physical connections change, but the functional connection remains. To assess vision rehabilitation requires an understanding of the cerebral processing that connects the input to the output. We will discuss some aspects of visual processing that are often overlooked.
As noted above, a common misconception is that visual acuity is an adequate measure of vision. Yet, reading optotypes on a chart assesses only the retinal area where the letter is projected. Even for a 20/200 letter, this is less than 1 degree. The articles by Mansouri et al (1) and Rucker and Phillips (2) emphasize that the visual field aspect and the motor aspect also are important.
Many discussions describe visual processing as a single stream from the environment, to the retinal image, to the visual cortex, and to higher centers. This single-source and single-pathway concept is too simplistic. Indeed, our perceptions and subsequent actions are based on matching 2 separate and very different sources. One is provided by the sensory input; the other is provided by concepts and templates stored in memory. When the 2 sources match, a mental model of our environment is formed. It is on this model that our perceptions and actions are based. Occasionally, an incorrect match is made; we then speak of an optical illusion.
When the correct match is made, we say that we recognize the object or person. Because the mental model results from 2 sources, it contains far more information than the retinal image. The main function of the retinal image is to trigger the memory store, which is multisensory and permanent, since it is built over time. When we close our eyes, the retinal image disappears, but the mental model of our environment persists. When we walk and look around, our retinal image changes constantly; yet, we perceive ourselves moving through a stable environment. That stability is an attribute of the memory concept, not of the retinal image. That the memory concepts are multisensory explains how a smell or a sound can trigger a visual memory, or vice versa.
The matching process is not completely predetermined. It is subject to the effects of attention and to its enhancement or suppression of certain stimuli. Attentional effects can be observed at fairly low levels, such as when the addition of a nonvisual second task during a visual field test reduces the size of the measured field. This effect shows that attention is not specific to any sensory modality; it is a limited resource which, like the memory storage, is shared across senses. Attentional effects also occur at higher levels, as shown in the well-known video, where attention to fast-moving ball players in bright shirts completely blocks the perception of a much larger, slow-moving person in a dark orangutan costume (3).
Attention control is a part of vision rehabilitation that is not easily captured in traditional tests. This explains how some people may report improved functioning, even if this is not reflected in a traditional visual field test.
The importance of attentional control also highlights the distinction between conscious and autonomous processing of visual information, another aspect that is often overlooked. Conscious visual processing is connected to vast memory banks of previous information, which can lead to object recognition, cognition, and conscious decision-making. In the eye, it is dominated by the foveal area, and in the brain by the ventral stream. The other processing stream is autonomous and never enters consciousness; it is connected to various motor systems. It can react faster, but offers less detail, and its actions are not stored in permanent memory. In the eye, it is dominated by peripheral vision and in the brain by the dorsal stream. The autonomous system helps us to avoid obstacles, guides our feet around obstacles or along a winding path, and opens our hand just wide enough, but not too wide, to grasp an object. The 2 systems have many interconnections, so that one can switch seamlessly from one system to the other.
Understanding this distinction between conscious and autonomous processing helps us to understand why patients are naturally aware of foveal defects, but not of visual field defects. Peripheral defects may cause one to bump into objects; scotomata in the central field may interfere with the ability to read. Yet, patients who have these visual field defects often are not aware of them unless brought to their attention.
Testing the conscious part of vision is easy, but testing the autonomous part is difficult because as soon as we ask for a conscious response, we enter the conscious system. Button clicks in most traditional visual field tests are such a conscious response. To demonstrate an autonomous response that may be more relevant to daily activities, it may be better to measure a natural response, such as the timing of a visual search task. Tests that map the condition of the retina must rely on steady fixation, which is not a natural condition. Fixation shifts that are undesirable for field tests may be helpful for daily activities because they reduce the extent of the absolute defect. One explanation for the boundary shift in hemianopia is a shift in fixation, away from the edge.
Another aspect that is often overlooked is that vision did not develop to perceive stationary visual scenes, as often used in laboratory experiments. Vision evolved to assist visually guided behavior in dynamic scenarios; the study of vision, therefore, is not complete without considering its motor effects (dorsal stream). In a dynamic situation, it is important to predict what may happen in the next instant. When playing a ball game, the most important information is not where the ball is now, but when and how it will reach the player, so that the player can start a deliberate movement, anticipating that their foot, hand, or racquet will meet the ball at the right time and the right place. Motor responses are not only important for games; they also affect mobility and manipulation tasks. When reading, the motor ability to move the eyes from word to word is essential and can be disrupted by scotomata.
Shifts in fixation result from shifts in attention. Attention is a brain function, not a retinal function; it is not restricted by the limits of the visual field. In the absence of hemineglect, patients with hemianopia can direct their attention to anticipated or remembered objects on their blind side. This conscious effort can be far more effective for avoiding obstacles than a few degrees that can be demonstrated on visual field testing. To assess how well individuals can replace their autonomous obstacle detection with intentional attention shifts requires a real-life mobility course, rather than visual field testing or fMRI.
In their article, Phillips and Rucker state: “The fundamental goal is to promote the mechanisms that underlie visual recovery.” We challenge that statement and would rather say that “The fundamental goal is to promote the patient's quality of life and optimal functioning in society;” restoring underlying functions is a means toward that end. Scientific studies strive for objective over subjective evidence. This is a worthwhile goal that should not be abandoned. However, if the goal is quality of life, subjective satisfaction, such as increased visual confidence, should not be dismissed, even if objective tests cannot confirm it. Timed task performance and patient-reported questionnaires may provide more appropriate validation tests.
Some researchers have been criticized for not waiting for spontaneous recovery. Yet this approach would overlook the possibility that early training might facilitate recovery of partially damaged connections. In other situations, it has been proven advantageous to start rehabilitation early, instead of waiting for medical stability.
It was stated earlier that vision aims at facilitating interaction with the environment, which is accomplished through various motor systems. The report by Rucker and Philips (2) concentrates on various forms of ocular motor training. Their survey concludes that such training can be effective for convergence insufficiency, but that its effect on more complex conditions, such as adult deficit hyperactivity disorder and dyslexia is problematic.
Based on the previous discussion, this finding is not surprising. Convergence insufficiency training directly addresses the ocular motor system and provides immediate feedback (single vision). Both factors contribute to its success. For other disorders, which combine both perceptual and behavioral components, an understanding of the pathophysiologic processes is lacking. Therefore, it is no surprise that a uniform approach and uniform evaluation also are lacking. Visual problems may contribute, but it is far from clear if they are a major factor. The interaction of a variety of components is exceedingly complex and cannot be measured adequately at this time.
In summary, we applaud the interest in vision rehabilitation. It should be recognized that rehabilitation is not merely an extension of traditional medical and surgical care, which is focused on the structure and function of organs, and on “How the organ functions.” Rehabilitation is interested in “How the person functions” in a societal context. The definitive evaluation tools must vary accordingly. The motto of the American Academy of Ophthalmology reflects these dual goals: “Preserving sight. Empowering lives.”