The Role of Vision in Learning Disabilities - The Flawed Argument That Does

My earlier blog (see here) took aim at Irlen Syndrome as a major cause for reading disabilities and the proposed mandatory screening for all school children. It now seems that this particular legislative initiative has been abandoned in favour of a slightly more sophisticated version that targets “visual impediments to learning.” As this argument is made with the support of a very selective and uncritical review of the literature on vision and learning disabilities, I felt that a technical account of what is actually known about visual disorders and reading is needed. I can’t argue with platitudes such as vision and audition are both involved in reading (if we, for the sake of argument, conveniently ignore braille), but the devil here is both in the details and in the larger picture. Let me illustrate this with dyslexia, arguably the most common learning disability and also the most widely studied.

Developmental dyslexia – a specific difficulty in developing accurate and/or automatic word reading skills – was historically associated with visual disorders. Over the long history of dyslexia studies, it has, for example, been called “word blindness” or “strephosymbolia” and associated with a variety of presumed visual disorders most of which were initially poorly understood. By the 1970s most of these visual explanations ran both into conceptual as well as experimental trouble in that it became increasingly clear that no amount of visual training helped students to read better and that the linguistic basis of reading disabilities started to be established (Vellutino’s 1979 book summarizes these nicely).

In the 1980s and 1990s, multiple research programs attempted to explore the aetiology of dyslexia in some visual pathways. The most prominent of these hypotheses proposed that dyslexia results from a specific impairment in the subsystem, called magnocellular system, that processes faster changes in visual stimuli. Initially, the argument went that the magnocellular system inhibits the parvocellular system – the system that processes slow changes – during eye movements (saccades) and thus eliminates blur in perception. If you have seen a recording of how your eyes move during reading, it is easy to understand why the assumption of blurred vision made sense. However, we now know that the magnocellular system rather than the parvocellular system is suppressed during saccades, completely undermining the conceptual foundation of the magnocellular hypothesis. Moreover, a multitude of empirical studies failed to find specific magnocellular deficits in individuals with dyslexia and when these were found, only a small number of individuals exhibited them. Disappointingly, almost all reading studies reviewed in the paper used to justify visual deficits to Alberta legislators are from this line of research which is considered, at best, only marginally relevant. 

The original hypothesis has subsequently been revised in various ways and two different hypotheses seem to have emerged with more support than the others. One is that sluggish serial identification of visual stimuli possibly leads to “crowding” or interference from similar neighbouring stimuli (such as letters). Thus, the main problem is located in serial as opposed to parallel processing and has to do with rapidly identifying briefly and serially presented stimuli. Interestingly, this seems to be true for both auditory and visual stimuli and thus the deficit is possibly not a visual deficit but an attentional deficit.

The second hypothesis suggests that parallel rather than serial visual identification is impaired in dyslexia. It is still unclear whether these visual attention span deficits are limited to visual material that is verbally coded (and therefore possibly reflect verbal processes), but it begins to be clear that, to the extend that they identify individuals with word reading difficulties, these individuals may be a distinct subtype of dyslexics. It also seems that visual attention span deficits are, at least in some studies, dissociated from phonological deficits, magnocellular dysfunction, and sluggish attentional shifting. Interestingly, the latter two tend to co-occur with phonological deficits, suggesting a possibility of partly shared etiology. 

This last finding invites another important observation: the assumption that developmental disorders (and all learning disabilities are developmental disorders) can have a single cause – visual or verbal – at any level of analysis has been challenged by multifactorial etiological models. One of the leading researchers in this area, Dr. Bruce Pennington from the University of Denver, reviewed evidence from multiple studies examining comorbidity – joint existence – between reading and speech disorders and between reading and attention disorders. He argued that probabilistic multiple deficit models are needed to provide realistic accounts of developmental disorders and the nondeterministic relationship between disorders and their presumed causes. We have written elsewhere more about these models, but the important point here is that due to the complex genetic, neural, cognitive and behavioural aetiology of almost all learning disabilities, they tend to come packaged together. In other words, a child with learning disabilities is likely to have other issues as well, including visual problems, but almost never do the latter “cause” the former (as the Irlen syndrome people would argue).

To end this somewhat complex piece with a clear “take home” message: to the extent that the most common learning disability, dyslexia, is associated with any kinds of visual problems, these problems are not the kinds of problems that optometrists can usually identify or do anything about. Most importantly, these visual problems are not the cause of the learning disability. Causal models for all learning disabilities are much more complex and, so far, only partially understood. However, when we begin by identifying individuals who struggle with reading or with the important pre-reading skills, in other words, we start from looking at the learning disabled population and not the population with visual impairments, we find, without exception, that the vast majority (up to 80%) of them struggle with meta-linguistic skills (such as phonological awareness), and most of them have not had access to the necessary learning supports. Due to comorbidity, some children will also have visual problems. In most of the Alberta schools I have visited, the necessary expertise and resources are not there to identify reading disabilities and to help these children properly to make important gains in their language development. My question to Alberta legislators is this: if we have additional resources to invest into the educational system, why not invest them where the best available research promises the highest return on that investment? In other words, invest in early identification of language-based learning disabilities.