Guest post: Auditory processing disorder – a cause of language problems or an incidental finding?
What causes a child’s language problems? Is it a problem with hearing? A problem with the brain’s interpretation of speech? Is it genetic? Specialists sometimes diagnose ‘Auditory Processing Disorder’ but the term itself is a complicated affair, writes Dorothy Bishop.
Five-year-old Charlie doesn’t speak very clearly, and doesn’t always understand what people are saying. His speech and language therapist says he has a specific language impairment, but his mum thinks he may be deaf. His family doctor refers him to an audiologist, who gives a hearing test and shows that Charlie can detect sounds just as well as other children. The audiologist suggests, however, that he may have an auditory processing disorder, or APD. The family doctor, however, has never heard of APD and is not sure whether it means that Charlie needs a different kind of treatment.
Parents are understandably confused when their child receives multiple diagnoses, some of which may not be found in paediatric textbooks. Use of the APD diagnostic label shows massive geographical variation. The diagnosis is common in the USA and Australia, and rare in the UK. I’ve commented elsewhere on the commercial pressures that may affect the variable use of the term. Here, I want to explain how scientists can go about trying to pinpoint the source of a child’s difficulties.
The field of APD is full of controversy, but let’s start with a point that most people will agree on, namely there’s more to auditory perception than is measured in an audiogram.
An audiogram is the result of a standard hearing test where you are played loud and soft sounds and have to indicate when you can hear them. Poor hearing is often the result of a problem in the peripheral auditory system, i.e. the middle or inner ear. But as well as being able to detect soft sounds: you also have to tell different sounds apart and recognise sounds (see Figure 1). Just as you can be colour-blind despite 20/20 vision, it’s possible to have problems with discrimination and identification of sounds despite having a normal audiogram.
Problems of this kind originate in the brain rather than the ear. The ear’s role in hearing is to turn a sound wave into a neural signal. This is then transmitted up the auditory nerve to different brain regions that decode the signal. Studies of adult neurological patients have shown that if certain brain regions are damaged then it can be difficult to interpret what is heard, even though you know a sound occurred. We also know from studies of animals that there are cells in the auditory regions of the brain that respond to specific sound features, such as pitch, duration or spatial location. So it seems entirely plausible that there may be children who have developmental abnormalities of the central auditory system that affect their ability to perceive particular sound features, with consequent knock-on effects on language development.
So why is this controversial? One problem is that nobody can agree on how APD should be defined. From time to time professional groups get together to try and sort out agreed criteria for diagnosis, but this has not led to consensus, perhaps because there is only a slender research basis. One key issue noted by Moore (2006) is that existing tests tend to use speech stimuli to diagnose APD. This leads to ambiguity because poor auditory performance can arise from lack of language skills rather than the other way around. For instance, Japanese listeners often have difficulty distinguishing ‘r’ and ‘l’ sounds, not because there is anything wrong with their hearing, but because their language does not make this distinction. And English speakers do just as poorly when required to distinguish sound contrasts that our own language does not have, such as length contrasts that are used in Finnish.
The importance of language knowledge was made clear in a study that used an American speech-based APD test with children in the UK. They did far worse than American children when tested using speech stimuli delivered in an American accent (Dawes & Bishop, 2007). For this reason, researchers in the UK have developed tests of auditory processing for children that involve sounds such as tones, rather than speech (Moore et al., 2010).
But even this may not sort out the chicken and egg problem of whether an auditory problem causes language difficulties or vice versa. Figure 2 shows possible relationships between these deficits.
Interpretation A is often assumed if auditory deficits are found in a child with language problem, i.e. people assume that the auditory problem has caused the language difficulties. That is a reasonable enough account, given that the child needs to hear language properly to learn it. However, the opposite relation could also explain the results. Suppose children’s language difficulties arose for quite different reasons e.g. problems in interpreting word meanings, so they take longer to learn new words. This might lead them to be less attentive to auditory features, just because they are less skilled interpreters of speech. Or consider the case of listening to a foreign language, where the listener has perfectly normal auditory processing but limited knowledge of the language. Compared to a native Finnish speaker, you or I may have more difficulty hearing differences in the length of two tones just because length differences are important for distinguishing speech sounds in Finnish but not in English. Thus, if we find a correlation between auditory deficit and language deficit, it can be difficult to distinguish interpretation A and B.
To make matters even more complicated, there is also interpretation C, where the co-occurrence of auditory and language deficits does not involve any causal link between the two; instead both deficits arise as the consequence of another factor. What might such a factor be? Well, to take just one example, some children have delayed maturation of certain brain regions. This could impact on a range of skills. Or another example would be a genetic abnormality that affects several different aspects of development.
These different interpretations have important implications for intervention. If interpretation A is correct, then it would make sense to train the child’s auditory skills, as this could have beneficial effects on language. However, if B or C are correct, then auditory training would not be effective.
To disentangle these alternatives we need converging evidence from several sources. One body of research takes children with language problems as a starting point. It explores their ability to hear different types of sound contrast to establish which, if any, auditory deficits are associated with language difficulties. This might seem like a fairly straightforward question, yet after some 40 years of research there is still considerable disagreement. There are problems both with the evidence itself and with the interpretation of that evidence.
Part of the difficulty in this field arises because young children are not very happy at sitting and making judgements about to hard-to-distinguish sounds for minutes on end. Suppose we want to find out if a child can hear differences between sounds that differ in pitch. Ideally, we want to establish a threshold level at which the difference is just noticeable. So we could play children pairs of sounds and ask them to judge if they are the same or different.
The problem is that the average 7-year-old lasts about two minutes on such a task before asking ‘how much more is there?’ and looking bored. Rather like a parent driving off on holiday with a child in the back seat, the experimenter has the unfortunate task of explaining that there are only another hundred trials to go. Some of my work in this area has been focused on the theoretically unexciting but practically important task of inventing auditory ‘games’ that the child will cheerfully play for long enough to give a sensible auditory discrimination threshold (see Fig 3). Even so, if we find that children with language difficulties do less well than typically-developing children, it can be hard to certain that we have identified a real auditory problem rather than poorer attention or concentration.
Because of this difficulty, I have become interested in alternative methods that do not require the child to attend and explicitly respond to stimuli. Using electrophysiological recording from electrodes situated on the scalp it is possible to measure the minute electrical discharges that occur in the brain when a sound is presented to the ear. By averaging over many stimulus presentations, we can detect a distinct kind of waveform representing activity over time in underlying auditory cortex. Of particular interest is a method whereby we compare the brain responses to different auditory stimuli, such as tones of different pitch or the sounds ‘ba’ and ‘da’.
Unfortunately, although this sounds neat, results are not always consistent from one study to another (Bishop, 2007). In a recent study we found, however, that whereas children with language problems had normal brainwaves at a point that corresponded to stimulus discrimination, they showed abnormalities at a later stage that appeared to reflect stimulus identification (Bishop et al., 2010). This was evident in both school-aged and teenaged children, and was seen for non-speech as well as speech stimuli.
There is a widespread tendency to assume that if you find differences in the brain responses of children with language impairment compared with a control group, that this means you have found the origin of children’s language difficulties. However, the same logic applies as with the behavioural results (Figure 2): we need to consider interpretations B and C as well as A.
Currently, the field tends to be polarised between those who think auditory deficits are an important cause of children’s language problems (interpretation A), and those who think they are either the consequence of language problems (interpretation B) or an incidental finding (interpretation C). My current view based on research to date? I think auditory processing problems may play a contributory role in causing language and literacy problems, but they are not the cause.
There are several reasons for this conclusion. First, attempts to remediate language difficulties by auditory training have been largely disappointing (Loo et al, 2010). Second, there is mounting evidence that genetic differences between children play an important role in the cause of children’s language difficulties, but this is not so for the auditory deficits that can accompany language problems (Bishop, 2006). The ability to identify and remember the sounds of speech seems to be a specific skill that is crucial for language learning and that can be disrupted in some children, even when their nonlinguistic auditory processing is entirely normal.
In sum, I would be reluctant to throw out the baby with the bathwater and deny any role of auditory impairment in causing problems with language learning, but the evidence to date suggests we should be cautious about using the term ‘auditory processing disorder’. Very often, difficulties in doing auditory tasks reflect the child’s poor language or attentional skills, rather than a primary perceptual cause of their difficulties. An important conclusion from research in this area is that we need more multidisciplinary working between audiology and other professions concerned with developmental disorders to get a fuller picture of how hearing and language problems are related (see Dawes & Bishop, 2009).
Dorothy Bishop is Professor of Developmental Neuropsychology and a Wellcome Trust Principal Research Fellow at the Department of Experimental Psychology at the University of Oxford. She blogs at BishopBlog and tweets as @deevybee.