Skip to Content

Department of Cognitive Science

Background

Language impairment in autism and SLI

Impaired communication is one of the defining features of autism. However, within the ‘autistic spectrum’, there is huge variability in the severity and nature of the communication deficit, ranging from subtle pragmatic conversational difficulties through to the complete absence of speech. Researchers have tended to focus more on the impairments in social communication because these were considered more universal. Recently, however, there has been a resurgence of interest in structural language impairments (i.e., in phonology, semantics, and grammar), motivated primarily by the apparent overlap between autism and specific language impairment (SLI).

SLI is diagnosed in children who have language difficulties in the absence of any obvious cause (such as autism, hearing impairment, or low IQ). It is considered to be a separate disorder from autism and research on the two disorders has largely proceeded independently. However, a subgroup of individuals with autism shows a similar profile of language deficits to those with SLI [1] and studies of the relatives of children with autism suggest a common genetic link [2]. In this project we will investigate both disorders in a series of studies motivated both by theories of autism and by theories of SLI.

Cognitive and neural integration in autism

Theories of autism see language impairment as a consequence of broader cognitive and neural disorder. One of the most prominent current theories states that many of its symptoms are caused by a deficit in integrating multiple sources information [3,4], which arises because different regions of the brain are disconnected from one another [5,6]. Neuroimaging studies have provided evidence for structural and functional ‘underconnectivity’ [7] and research is currently underway to identify genes that affect neural connectivity and may thus increase the risk of autism [8]. However, despite the intuitive appeal of this account, the relationship between neural anomalies, cognitive dysfunction, and actual behavioural symptoms is not transparent. Nor is it clear how specific these deficits are to autism.

Language comprehension represents a model system for investigating these links. At the cognitive level, the meaning of any word or phrase is dependent on the overall context in which it is heard, and the integration of these contextual cues is clearly an essential part of ‘high-level’ comprehension. Studies have shown that despite having good single word understanding, people with autism typically perform poorly on sentence comprehension tests and often fail to choose the contextually appropriate interpretation of ambiguous words [9,10]. However, some recent studies have provided contradictory evidence [11]. In particular, Brock et al. [12] recorded children’s eye-movements as they listened to spoken sentences containing contextual information. Children with and without autism did not differ in their tendency to look towards contextually irrelevant objects. Instead, this measure was strongly associated with children’s language ability, suggesting that integration of linguistic context is a function of language impairment rather than autism.

The story is similar at the neural level. Language comprehension involves the interaction of an extended network of brain regions and previous studies using functional magnetic resonance imaging (fMRI) have shown that the integration of sentence context leads to an increase in the temporal correlation of left hemisphere activation, implying increased functional connectivity [13]. In a seminal fMRI study of sentence comprehension in autism, Just et al. [6] found abnormally increased activity in Wernicke’s area, decreased activity in Broca’s area, and, most importantly, reduced functional connectivity between these regions. However, the language ability of participants in this study was not clear and similar results have been reported in fMRI studies of both dyslexia [14] and degenerative language impairment [15].

In sum, current theories imply that ‘underconnectivity’ and cognitive integration deficits are ubiquitous in autism and specific to the disorder. However, an alternative possibility is that specific patterns of connectivity are associated with specific cognitive deficits and with specific behavioural symptoms. This perspective allows for the heterogeneity within autism and for overlap with other disorders such as SLI.

In this project, we will again use eye-movements as a measure of ongoing sentence comprehension. We will also use magnetoencephalography (MEG) to measure neural responses to sentences, allowing us to identify precisely which brain regions are activated during the integration of contextual information and to measure their functional interaction. Investigation of group and individual differences between children with autism, SLI, and typical development will establish the links between cognitive and neural integration and the specificity of any impairments to autism.

Auditory and phonological processing in SLI (and autism)

Theories of SLI have focused on more specific, localized causes. One class of theory posits that SLI is caused by a deficit in phonological processing. The most consistent evidence comes from the poor performance of children with SLI on nonword repetition tasks, in which they are required to immediately repeat a made-up word (e.g., ‘blonterstaping’) [16]. However, this task is deceptively complicated and can be failed for a number of different reasons including inability to identify the constituent phonemes, poor short-term memory, and articulatory difficulties [17,18,19]. A number of recent studies have also reported poor nonword repetition in autism [1], but it is unclear whether individuals with autism and SLI perform poorly for the same reasons. Computer-based tasks will be included in the initial test battery to assess the component skills required for nonword repetition.

Other theories of SLI focus on auditory perception. Some researchers have argued that these problems are restricted either to rapidly presented stimuli [20] or to speech-like sounds [21]. However, McArthur and Bishop [22,23,24,25] have shown that many individuals with SLI show atypical brain responses to both speech and non-speech sounds regardless of presentation rate. Auditory processing anomalies have also been reported in studies of autism [26,27]. In this project, direct comparison of brain responses to auditory tones during MEG testing will allow us to investigate the role of auditory processing deficits in language impairment in individuals with and without autism.

At the neural level, Leonard and colleagues [28,29,30] have shown that children with language (and reading) disorder have atypically small and symmetrical auditory cortices. It is unclear at present whether there is a direct association between these structural anomalies and impaired auditory or phonological processing, or whether similar anomalies are present in children with autism who have language impairment. Structural brain scans will be collected as part of this project and analyses of individual differences will allow us to address both issues.

References

  1. M Kjelgaard, H Tager-Flusberg, 2001, Language and Cognitive Processes, 16, 287.
  2. S Folstein et al. 1999, Journal of Child Psychology and Psychiatry, 40, 1117.
  3. U Frith, 1989, Autism: explaining the enigma. Oxford: Blackwell, 1989.
  4. N Minshew et al. 1997, Journal of the Neuropsychological Society, 3, 303.
  5. J Brock et al. 2002, Development and Psychopathology, 14, 209
  6. M Just et al. 2004, Brain, 127, 1811.
  7. G Rippon et al, 2007, International Journal of Psychophysiology, 63, 164.
  8. D Geschwind, P Levitt, 2007, Current Opinion in Neurobiology, 17, 103.
  9. U Frith, M Snowling, 1983, British Journal of Developmental Psychology, 1, 329.
  10. T Jolliffe, S Baron-Cohen, 1999, Cognition, 71, 149.
  11. C Norbury, 2005, Journal of Experimental Child Psychology, 90, 142.
  12. J Brock et al, 2008, Cognition, 108, 896.
  13. J Obleser et al. 2007, Journal of Neuroscience, 27, 2283.
  14. E Paulesu et al. 1996, Brain, 119, 143.
  15. S Sonty et al. 2007, Journal of Neuroscience, 27, 1334.
  16. K Graf Estes et al. 2007, Journal of Speech Language and Hearing Research, 50, 177.
  17. L Archibald, S Gathercole, 2007, Psychonomic Bulletin and Review, 14, 919.
  18. L Archibald et al, 2009, Journal of the Acoustical Society of America, 125, 1712.
  19. J Brock, C Jarrold, 2004, Journal of Speech Language and Hearing Research, 47, 1334.
  20. P Tallal, 2000, in D Bishop & L Leonard (Eds.), Speech and language impairments in children, 131. Hove: Psychology
  21. ress.
  22. M Mody et al. 1997, Journal of Experimental Child Psychology, 64, 199.
  23. G McArthur, D Bishop, 2004. Cognitive Neuropsychology, 21, 79.
  24. D Bishop, G McArthur, 2004. Developmental Science, 7, F11.
  25. D Bishop, G McArthur, 2005). Cortex, 41, 327.
  26. G McArthur, D Bishop, 2005. Brain and Language, 94, 260
  27. A Lincoln et al. 1995, Journal of Autism and Developmental Disorders, 25, 521.
  28. N Gage et al. 2003, Developmental Brain Research, 144, 201.
  29. C Leonard et al. 2002, Journal of Communication Disorders, 35, 501.
  30. C Leonard et al. 2006, Brain, 129, 3329.
  31. C Leonard et al. 2008, Neuropsychology, 22, 147.