Educational Neuroscience: Implications for deaf children
By Mairéad MacSweeney, Wellcome Trust Senior Research Fellow at the UCL Institute of Cognitive Neuroscience.
I work with people who are born severely or profoundly deaf in order to inform our understanding of how the brain processes language. People born deaf can provide a unique insight into how the brain processes language. This is because the vast majority of the population are born hearing, and hear spoken language even before they are born, while still in the womb. They are always surrounded by it. But for a child who is born profoundly deaf, the situation is very different. By definition they have incomplete access to spoken language. In addition some deaf children may access a purely visual language – in this country – British Sign Language. We can look at language development and brain development in these children to gain unique insights into how language is processed in the absence of sound. I look at sign language processing and also spoken language processing in the form of lip-reading and reading.
Reading forms a major part of my current research. Deaf children find it incredibly difficult to learn to read, because when we read we read a spoken language written on the page and deaf children have impoverished access to spoken language. Although deafness is not a learning disability – most deaf children have normal non-verbal IQs – a typical deaf child will have a reading age of 10-11 years old when they leave school aged 16. This has major consequences for their educational and vocational attainment.
A major skill in developing reading accuracy is learning how to map the written letters onto the correct sounds. This is what phonics training is all about – mapping sounds to letters or letter combinations. Broader language skills are also very important to reading comprehension, such as semantics (the meaning of words) and syntactic structure (how words are put together to form sentences).
Some deaf children do become very good readers. If we can understand how they achieve this, despite the fact that they can’t hear the language that they are reading, then this can give us insights into how we might be able to better teach all deaf children. More broadly, an understanding of how important hearing spoken language is for learning to read has the potential to inform our understanding of reading in hearing children too, not just deaf children.
For example, one of the core deficits in hearing children with dyslexia is that they have poor phonological awareness skills, such as knowing that “chair” rhymes with “bear,” or that “split” without the “l” is “spit.” One of the main theories of dyslexia in hearing children is that they have some kind of low-level auditory processing deficit. With deaf children we can look at a group who have very little or no auditory input, and look at the impact of that on their reading.
How far is this area of research from being applied in education?
Some areas of this research are ready to be tested for educational applications, indeed, I am about to start a randomized controlled trial doing just that, funded by the Wellcome Trust. Data from some of my previous neuroimaging work with deaf adults suggest that how words are represented or processed in the brain at the ‘sub word’ level (also called – sub-lexical or phonological level) is very similar in the brains in deaf and hearing people, even though they’re based on different modalities, or types of inputs. For example, hearing people know that “chair” and “bear” rhyme because of their sounds, and their neural representations of the spoken words are based primarily on the auditory inputs. Many deaf people also know that these words rhyme, but our data, as well as research from others, suggests that this knowledge is based on lip-reading, and also how the words feel when they’re spoken. The idea that deaf people may generate phonological representations in the brain, which are very similar to those generate by hearing people, suggests that these representations may operate at an abstract, or amodal, level.
On the basis of this model, my colleagues and I are developing a lip- reading training program for young deaf children, which will use phonics training-type games and visual speech (lipreading). For example, children will see a film clip of someone saying the word, for example “ball,” and then they will have to choose a picture of a ball from a set of options. We are currently developing this and will be running a randomised controlled trial next autumn to test the idea. The trial will be conducted in a number of different schools, and will involve the children doing ten minutes of training per day, four days a week, for 12 weeks. A control group will spend the same amount of time playing games that train numeracy skills. After the training period, we will look at any changes in the lip-reading skills, phonological awareness (e.g., knowing that ‘chair’ and ‘bear’ rhyme) and reading skills of the two groups. We of course predict, and hope, that the group who have had played the lipreading games will show greater gains in their literacy skills than those who played the maths games.
The model that we’re testing has support from neuroscience, but you could argue that you could develop it on the basis of behavioural data alone. The neuroscience data do however lend strong support to the proposal that it is ‘amodal’ phonological representations that underlie the relationship we see between lipreading and reading in deaf children. In this trial, we’re not including any neuroscience measures at all. What we are doing is seeing if the intervention works, behaviourally. If it does, then we will perform another trial, involving neuroimaging measures, probably event-related potentials, to give us insight into the mechanisms that are changing during the intervention.
What are the limiting factors?
We need randomised controlled trials to assess interventions, but trials involving any neuroimaging measures are difficult and costly. I think the best way forward for the growing field of ‘educational neuroscience’ is to take interventions that work behaviourally, ones that have a good grounding in experimental and cognitive psychology, and then to test these in randomised controlled trials involving neuroimaging measures to gain insights into the mechanisms underlying the behavioural changes. I would argue that neuroscience-based interventions that don’t change behaviour are a waste of time. So, throwing money at educational interventions and scanning children’s brains before we even know if an intervention works at the behavioural level is probably the wrong thing to do.
We have to be very clear about what we expect from neuroscience in terms of implications for educational practice. A lot of people think that it will give us something unique – not previously explored in the fields of developmental psychology or cognitive science, but I don’t think it will. Neuroscience can tell us about the brain mechanisms underlying particular cognitive functions, and give us additional dependent variables to measure when we look at changes due to a particular intervention, but whether it can truly direct us to brand new educational interventions is another thing altogether. Perhaps the only exception to this is where underlying brain processes are manipulated directly, for example using Transcranial Magnetic Stimulation (TMS) or Transcranial Direct-Current Stimulation tDCS.
What are the potential benefits of applying this research to education?
With regard the trial that we will be running with deaf children, any gains we see in reading, even if very modest, would be of enormous benefit to educating deaf children. If we do see any gains, we would of course make the intervention available to more deaf children and more schools. From the behavioural data, we also have good reason to believe that the intervention may benefit hearing children, too.
What prospects are there for future development?
I think there are grounds for cautious optimism. Any future field of ‘educational neuroscience’ needs to build on developmental cognitive neuroscience. Understanding how the neural systems that support cognition develop and respond to changes in input and environment should ultimately help develop and refine educational interventions that have been shown to work behaviourally. Many current educational interventions are multi-faceted. Having particular predictions about brain changes due to one particular part of an effective intervention is a further tool to help pinpoint the factors that are having the most important effects in a large scale intervention – which can then help strengthen further developments.