Language disorders in children: What can they tell us about genes and brains?

1. Language disorders in children:What can they tell us about genes and brains? Dorothy Bishop University of Oxford

2. Specific language impairment (SLI) • Diagnosed in children when language does not follow normal developmental course • Problems with language structure (phonology and syntax) common • Not due to hearing loss, physical abnormality, acquired brain damage • Normal development in other areas

3. What causes SLI?

4. http://news.bbc.co.uk/1/hi/education/2638889.stm Thursday, 9 January, 2003, 11:15 GMT Daily grunt parents hold children back Not enough talking going on Parents who do little more than grunt at their children every day are damaging their language development, a literacy expert has said. Alan Wells, director of the Basic Skills Agency, says parents no longer talk to their children and instead just let them sit in front of the television or computer for hours.

5. Evidence against poor language environment as cause of SLI • Most children develop adequate language despite very variable levels of language input from parents • E.g. different cultures • Children in exceptional circumstances, including hearing children of deaf parents • Twin studies: dizygotic twins growing up in same environment can be very different

6. Is brain damage implicated? • Early idea of “continuum of reproductive casualty” - i.e. brain damage incurred around time of birth might lead to learning disabilities Intraventricular hemorrhage : a common feature of prematurity

7. Evidence against brain damage as cause of SLI • No obvious signs of neurological impairment in most cases of SLI • No excess of perinatal problems in SLI • Children who do have focal lesions affecting the language areas don’t develop SLI - see Basser 1962!

8. Genetics

9. SLI: Family aggregation Rates of language/learning difficulties higher in relatives of those with SLI, compared with controls

10. Twin Study Method MZ twins: genetically identical DZ twins: share 50% of polymorphic genes Question: Is concordance for disorder higher in MZ than in DZ twins?

11. Twin studies of SLI Probandwise concordance: same-sex twins MZ DZ Lewis & Thompson, 1992 .86 .48 Bishop et al, 1995 .70 .46 Tomblin & Buckwalter, 1998 .96 .69 Hayiou-Thomas et al, 2005 .36 .33 Bishop & Hayiou-Thomas, 2008 .88 .35 (same sample, criteria of SALT contact)

12. Conclude…….. Evidence for substantial genetic influence on SLI But lots of questions remain, e.g.: • Can SLI be caused by a single gene? • Is SLI a single condition, or are there etiologically distinct subtypes? • What (if anything) does this say about evolution of language?

13. Single gene disorder: KE family Hurst et al, 1990; Gopnik, 1990; Vargha-Khadem et al, 1995

14. Nature, October 4th 2001 A forkhead-domain gene is mutated in a severe speech and language disorder Lai, C. S., Fisher, S. E., Hurst, J. A., Vargha-Khadem,F., & Monaco, A. • FOXP2: gene on chromosome 7q31 • Single base mutation in affected individuals • No allelic variation in people with normal language

15. Nature, August 22nd 2002 Molecular evolution of FOXP2, a gene involved in speech and language Enard, W., Przeworski, M., Fisher, S. E., Lai, C. S. L. Wiebe, V., Kitano, T., Monaco, A. P., Paabo, S. • FOXP2 differs in man and mouse in only 3 amino acid positions • FoxP2 differs in man vs. chimp, gorilla, rhesus macaque in 2 amino acid positions • i. e. 2/3 differences between human/mouse occurred on human lineage after separation from chimpanzee

16. Is FOXP2 a 'grammar gene'? No! • FOXP2 is a transcription factor: regulates expression of many other genes, and expressed in multiple organs • Affected members of KE family have problems beyond grammar • Abnormality of subcortical regions involved in motor control (Watkins et al, 2002)

17. Nevertheless,affected members poor at syntactic tasks, even when no speech is required, e.g. The elephant pushing the boy is big

18. How to characterise FOXP2 • A transcription factor that affects expression of many other genes, and is expressed in multiple organs • Human version of gene may be important for building brain areas involved in extracting hierarchical structure from linear input

19. Q1. Can SLI be caused by a single gene? • Yes, in case of FOXP2 • But! This is the exception rather than the rule • No abnormality of FOXP2 seen in majority of cases of SLI • Pedigree analysis does not show straightforward pattern suggestive of single gene disorder

20. SLI as complex multifactorialdisorder • Aggregates but does not segregate in families • Typical etiology for common disorders e.g. heart disease, diabetes

21. The quest for heritable phenotypes • Need to move from clinical diagnosis to measures of underlying impairment

22. Some promising component skills • Auditory processing skills (e.g., Tallal) • Morphosyntax (e.g., Rice & Wexler) • Phonological short-term memory (e.g., Gathercole & Baddeley)

23. Children’s Nonword Repetition Test (CNRep) Gathercole and Baddeley, 1990 Child listens to spoken nonwords and repeats, e.g. 2 syllables: hampent 3 syllables: dopelate 4 syllables: confrantually 5 syllables: pristoractional

24. Nonword repetition NB performance near ceiling on 2 syllable nonwords data from Bishop et al, 1996

25. Genetic analysis: quantitative rather than categorical approach

26. nonshared environment only shared environment only genes only co-twin co-twin DeFries-Fulker (DF) analysis DZ proband MZ proband -1 0 z-score

27. Nonword Repetition Test: DF analysis h2g = 1.17 (SE = .319)

28. Involvement of phonological STM in syntax learning? • Baddeley et al (1998) notion of phonological loop as buffer to hold information while computing syntactic relations • Joanisse & Seidenberg (1998) problems that seem specific to grammar may be secondary consequences of phonological memory deficit To get beyond correlational evidence, can use twin method to test for common origins in different deficits

29. Twin Sample6 yr olds from community twin sample (TEDS), biased to include those with language difficulties • 65 MZ and 67 DZ pairs with ‘low language’ on basis of parental report at 4 yr • 37 MZ and 29 DZ control pairs (neither ‘low language’) Bishop et al., 2006

30. Nonword repetition NB test not given to 19 children with unintelligible speech Residual score: total correct on 3-, 4- and 5-syllable nonsense words predicted from the total of 2-syllable words: gives purer measure of memory

31. Nonword repetition: heritability of extreme scores h2g = .84 (95%CI = .58)

32. Rice-Wexler task: test of grammatical morphology “Here the boy is raking; now he is done. Tell me what he did”.

33. Here’s a farmer. Tell me what a farmer does.

34. N children inflecting verbs Mann Whitney; z = 5.75, p < .001 Bishop et al., 2006

35. Rice-Wexler task: heritability of extreme scores (< 93% inflected) h2g = .68 (95%CI = .59)

36. Nonword repetition probands; cotwin z-scores on verbs h2g.xy = .32 (95%CI = .71) co-twin verb proband phonSTM co-twin verb proband phonSTM

37. Genetic risk 2 Genetic risk 1 Conclusions from twin study Morphosyntax deficit Phonological STM deficit MZ twins more similar to each other than DZ, but only within-trait MZ twins more similar to each other than DZ phonological STM deficit phonological STM deficit

38. Are there separate subtypes of SLI? • Poor phonological STM • Poor grammatical morphology

39. Phenotypic overlap: phonological STM / verb morphology deficits phi = .29, p < .001

40. children with co-occurring deficits are most likely to be identified with SLI phonological STM deficit morphosyntax deficit Suggests language is a strongly canalised function: will develop despite variation in underlying cognitive bases

41. Q2. Is SLI a single condition, or are there heritable subtypes? Neither seems true! • Notion of disorder that arises only when there is more than one underlying deficit • Predicts relatives will show deficit without necessarily showing disorder • Predict genetic variants will be risk factors rather than act as single causes

42. Study of parents Barry, J. G., Yasin, I., & Bishop, D. V. M. (2007). Heritable risk factors associated with language impairments. Genes, Brain and Behavior, 6, 66-76.

43. Mean language/literacy scores in parents NB significant differences maintained when self-report LI excluded

44. SLI consortium: molecular genetic studies • Have found 3 genetic variants that are associated with nonword repetition in SLI samples • Two genes on chromosome 16 act additively; other gene is downstream target of FOXP2 • Poor performance on verb morphology linked to site on chromosome 19

45. Genes associated with nonword repetition deficits Very different from FOXP2 • 'Risk' variants are common in the general population • Associated with a drop in nonword repetition score, but of moderate magnitude Two genes on chromosome 16 Gene 1 Gene 2 figures from Dianne Newbury

46. Implications • Supports notion that language involves multiple cognitive processes • Syntactic deficits not reducible to more basic cognitive process (at least, not to those measured here) • Search for single cause explanations of SLI may be doomed - language is resilient enough to survive one deficit

47. Thanks to: Caroline Adams & Courtenay Norbury 6 yr old twin testing project Robert Plomin Twins Early Development Study Mabel Rice making prepublication test materials available Dianne Newbury, Tony Monaco, Simon Fisher making prepublication genetic results available

48. Dorothy Bishop Oxford Study of Children’s Communication Impairments, Department of Experimental Psychology, South Parks Road, Oxford, OX1 3UD, England. bishop@psy.ox.ac.uk References available on: http://www.psy.ox.ac.uk/oscci/