The main aim of the talk is to demonstrate how 'pure' scientific analysis and techniques can make a real contribution to the understanding of developmental dyslexia, the most common of the developmental disorders, and one that affects millions of children and adults in the UK. Developmental dyslexia is normally identified by unexpected problems in learning to read for children of average or above average intelligence —“a disorder in children who, despite conventional classroom experience, fail to attain the language skills of reading, writing and spelling commensurate with their intellectual abilities” (from the definition by the World Federation of Neurology, 1968, p26). A typical estimate of the prevalence of dyslexia in Western school populations is 4%, with roughly four times as many boys as girls being diagnosed. It has major financial and social repercussions at school and at work both for those with dyslexia and those working with them.
A major problem with developmental dyslexia is that it is diagnosed by means of reading difficulties, and so a child needs to fail to learn to read for two years or so before a diagnosis is considered valid. It is of course a very destructive process for a child to have the crucial early years at infant school blighted by failure to acquire one of the fundamental skills. For the later school years there is also the danger of a vicious circle with poor reading leading to poor motivation, avoidance of text-based school work, emotional trauma, and adoption of maladaptive strategies such as clowning around, disruption or truancy. Even in adulthood, many dyslexic people still feel intensely angry about the way they were treated at school. It is important to note, however, that dyslexia is defined in terms of a discrepancy between actual reading and the reading performance that would be expected on the basis of the child's intelligence. Many dyslexic children turn out to be creative and successful, and it has been suggested that many of our most creative artists and scientists were dyslexic.
The main aim of dyslexia researchers has been to find the underlying cause of dyslexia. If the underlying cause can be established, then it should be possible to diagnose dyslexia pre-school, thereby giving the opportunity to provide initial reading support better tuned to the way that dyslexic children learn best. This 'stitch in time' support should then lead to much improved reading, allowing the child to keep up with the rest of the class, and so avoid all the traumas caused by failing to learn to read - allowing the dyslexic child to enjoy his/her strengths without suffering from the core reading problem. There is very great interest internationally in dyslexia, and, in particular, in a long-standing core programme the US National Institute for Health has invested over $200M since 1984 in research aimed at discovering the underlying cause(s).
One of the most promising approaches to finding the underlying cause derives from findings in the early 1980s that dyslexic children had particular difficulty in hearing the individual sounds in words. For instance, at the age of 5 years, children who would later turn out to be dyslexic had considerable difficulty in hearing that, say, 'cat', 'mat' and 'bat' rhyme. In general, they seem to have limited 'phonological awareness' (sensitivity to the sound structure in words). This 'phonological deficit' leads to difficulties in learning to read and spell because one of the early stages in learning to spell is to split a word into its component sound chunks, each of which then has to be spelled in order. The phonological deficit account was for many years the dominant hypothesis for dyslexia research, but the key question is why dyslexic children have this problem.
In the talk, I shall describe the research undertaken over the past 12 years by my colleague Dr. Angela Fawcett and me, in our attempt to discover the underlying cause. The trail has led us from cognitive psychology to cognitive neuroscience to developmental science, but we believe that we now have a coherent account of the major symptoms, the underlying cause and the way the reading problems develop. Whether our account proves to be a complete explanation, or whether there are other important contributory causes, remain topics for further research, but at present, as I shall describe in the talk, there appear to be remarkably few loose ends.
In our original research, published 10 years ago, we argued that, unlike language, reading is not a 'special' skill for humans. We are not evolutionarily adapted to read, since after all, few people could read at all until the last two or three hundred years. Consequently, an analysis of the learning processes should cast some light on why dyslexic children fail to learn to read. One of the critical aspects of learning a skill fluently is to make it automatic, so that one can do it without thinking about it. A clear adult example of the importance of automatisation is in learning to drive. The beginner can either steer, or change gear, but not both at the same time, because of the need to consciously attend to each procedure. An expert driver changes gear and steers 'automatically', thus leaving more 'capacity' for watching the traffic, planning a manoeuvre, or holding a conversation. Of course, automatisation is a key requirement for reading, and there is extensive evidence that dyslexic children, even when reading well, are less fluent, requiring more time and effort to read than would a non-dyslexic child of the same reading age. Automatisation of the processes in reading is no different from the general processes of automatising any other complex skill, and so we started by putting forward the bold hypothesis that dyslexic children would have difficulty in automatising any skill (cognitive or motor). Rather to our surprise, this hypothesis was clearly supported by a set of experiments in which we asked dyslexic children to do two things at once. If a skill is automatic, then one ought to be able to do something else at the same time (assuming it does not directly interfere with the first skill) with little or no loss of performance. Our most startling finding was for balance - a highly automatic skill with no language component. We found that although a group of dyslexic adolescents were normally able to balance as well as 'controls' (non-dyslexic children matched for age and IQ), their balance deteriorated very significantly when they had to do something else at the same time, whereas the controls' balance was not affected at all. We tried a range of secondary tasks, including counting or pressing a button on hearing a tone (and also we tried blindfolding them to prevent the children consciously attending to visual cues when trying to balance) and got the same pattern. We concluded therefore that our hypothesis was indeed supported, and that dyslexic children were not automatic even at the fundamental skill of balance. For some reason, dyslexic children had difficulty automatising skills, and had therefore to concentrate harder to achieve normal levels of performance. We have used the analogy of driving in a foreign country - one can do it, but it requires continual effort and is stressful and tiring over long periods. On our account, life for a dyslexic child is like always living in a foreign country.
It should be stressed that automatisation is not a conscious process - by dint of practice under reasonably consistent conditions most humans just 'pick up' skills without having to think at all. Our account gave an intuitively satisfying account not only of the reading problems but also of the phonological difficulties (because phonological awareness is a skill that is picked up initially just by listening to one's own language). Furthermore, it explained why it is that everything needs to be made explicit in teaching a dyslexic child, whereas for non-dyslexic children one can often get away with just demonstrating the skill. Perhaps most satisfying, many dyslexic people and dyslexia practitioners came to us to say that our account seemed exactly right to them - they did have to concentrate on even the simplest skills. On the other hand, what was not clear was why dyslexic children have problems in skill automatisation, and until this puzzle has been solved, it was difficult to see how we would be able to test for dyslexia before school.
Interestingly, dyslexia has an established genetic basis - a male child with dyslexic parent or sibling has a 50% chance of being dyslexic. There should therefore be some underlying abnormality of the brain reflecting this genetic inheritance. Researchers have investigated the language area of the cerebral cortex, together with the relative size of corresponding regions of the right and left cerebral hemispheres (most right handed people have the temporal lobe of the left hemisphere specialised for language processing). However, promising early leads seemed to peter out on further investigation. There have also been recent investigations of the magnocellular pathways - sensory pathways from the eye and ear that carry information rapidly to the brain, but it is not clear why sensory input difficulties might cause problems, say, in spelling. It had long been known that the cerebellum (the 'hind brain' - a primitive but very complex brain structure at the back of the brain) is involved in acquisition and execution of motor skills such as walking and reaching. Interestingly, however, when brain imaging techniques such as PET scanning were introduced in the early 1990s, it became clear that the cerebellum was highly active in a range of skills - when imagining a tennis stroke, when speaking, or even when trying to keep a list of words in memory. These findings tallied with an emerging view that the cerebellum was a key brain structure for the acquisition and use of a range of cognitive skills, including 'language dexterity'. Putting together the 'cognitive neuroscience' results on the role of the cerebellum in skill automatisation, balance and language dexterity with our own findings with dyslexic children, it became clear that the cerebellar abnormality was a prime candidate for the cause of the difficulties suffered by dyslexic children.
Over the past five years we have completed a series of stringent tests of our cerebellar deficit hypothesis. First we undertook clinical tests of cerebellar dysfunction - both dysmetria (difficulty in precisely measured movements) and dystonia (low muscle tone) - on our panel of dyslexic and control children. We established that the dyslexic children showed marked deficits on almost all of these clinical tests, and we then replicated these findings on further populations of dyslexic and control children, establishing that around 80% of our sample of dyslexic children showed clear 'cerebellar' symptoms. These findings were completely unexpected from the literature and were not predicted from any other theory of dyslexia, and consequently they provided strong support for the hypothesis. Nonetheless, it could still be argued that it was not the cerebellum itself, but perhaps some input to, or output from, the cerebellum that was causing the problems. This is by no means an unlikely hypothesis because the cerebellum has two way connections with almost all parts of the brain, including the language areas. Consequently, in collaboration with colleagues in other laboratories, we carried out two direct tests of the cerebellar deficit hypothesis. Particularly striking results were obtained from a PET (brain imaging) study involving learning a sequence of finger presses, known to result in considerable activation in the cerebellum with non-dyslexic adults. We established that our dyslexic adults showed only 10% of the normal cerebellar activation both when executing a previously overlearned (automatic) sequence and when learning a new sequence. This suggests strongly that, unlike non-dyslexic adults, dyslexic adults do not activate the cerebellum in these learning and automatic tasks - presumably because it does not help them in the normal way. Finally, there is a collection of dyslexic and non-dyslexic brains in the Beth Israel / International Dyslexia Association Brain Bank (Boston, USA) and our PhD student, Andrew Finch, was given permission to undertake neuroanatomical investigation of the cerebellar regions of these specimens. He established significant abnormality, characterised by greater cell size, in both the cerebellum and in the inferior olive (a nucleus in the brain stem that sends input to the cerebellum).
Consequently, at least for the dyslexic children in our panel, we have found both behavioural and neurological evidence of cerebellar abnormality, thereby providing strong support for our cerebellar deficit hypothesis. Of course there is a great deal of research still to do, and indeed the hypothesis suggests a range of fascinating further studies, but those are for the future. This converging multidisciplinary evidence of cerebellar abnormality led us to develop an 'ontogenetic causal chain' analysis in which we propose that cerebellar abnormality from birth leads to slight speech output dysfluency, then receptive speech problems (i.e., difficulties in hearing the speech sounds) and thence deficiencies in phonological awareness. Taken together with the problems in skill automatisation and coordination associated with the cerebellar impairment, this analysis provides not only a good account of the pattern of difficulties suffered by dyslexic children, but also how they arise developmentally. This causal chain analysis still awaits confirmation via studies of pre-school 'dyslexic' children, but if supported by further research, it provides a very significant analysis. It demonstrates how abnormality in a brain structure (the cerebellum) can lead, via difficulties in cognitive processes such as automatisation and phonology, to deficits in arguably the pinnacle of cognitive skill, namely reading.
Finally, let us return to one of our initial motivations, that of identifying dyslexia pre-school and providing support proactively so that a dyslexic child can learn to read relatively normally. Several years ago we were able to put together our findings on balance and motor skill difficulties with 9 further tests known to be difficult for dyslexic children (including, phonological, memory, speed, copying and pre-literacy tasks) into a 30 minute screening test that could be administered by a teacher in a child's reception year at infant school. This 'Dyslexia Early Screening Test' has now been taken up by over 3000 schools in the UK, and can form the first stage in a screening-support system. We feel strongly that the development of this important applied educational test could not have been achieved if it had not been for our careful scientific analysis and investigation of the underlying causes of dyslexia.
topics - Dyslexia