Genetics of Language

1
Genetics of Language Language Disorders

• Karin Stromswold
• Dept. of Psychology Center for Cognitive
  Science
• Rutgers University - New Brunswick

• Portions of this work were supported by
• Johnson Johnson Foundation
• John Merck Foundation
• Charles Johanna Busch Biomedical Research
  Grant
• Bamford-Lahey Childrens Foundation
• National Science Foundation (BCS-9875168,
  BCS-0042561, BCS-0124095)
• Correspondence may be sent to
  karin_at_ruccs.rutgers.edu

2
Key Questions

• Do genetic factors affect peoples ability to
  acquire and use language?
• Do these factors affect 'normal' peoples
  linguistic abilities or just those with language
  disorders?
• Do language-specific genes exist?
• Are genetic factors involved in all aspects of
  language?
• Are the same genetic factors involved in all
  aspects of language?
• How do genes/environment interact?

3
Content

• Relationship between innateness heritability
• Review/meta-analysis of genetic studies of
  language
• Family aggregation studies
• Pedigree studies
• Adoption studies
• Twin studies
• Linkage studies
• Limitations/Worries
• Conclusions

4
Innateness Hypothesis Heritability

• Typical evidence Universal, learnable, modular
• Genetic evidence If innate cognitive
  predisposition or neural structures enable us to
  use/acquire language, they must be encoded in our
  DNA
• Why we might fail to find evidence for language
  heritability
• Heritability amount of individual variation due
  to genetic factors
• The Innateness Hypothesis is wrong
• Linguistically-speaking, (normal) people are
  genetically identical Chomsky (1980)
  Language is like number of fingers Lieberman
  (1984) Language is like height

5
Innateness Hypothesis Heritability

• Individual differences may exist
• Acquisition rate for vocabulary (e.g., Goldfield
  Reznick, 1990), morphology (e.g., deVilliers
  deVilliers, 1973), syntax (e.g., Stromswold 1990,
  1995, Snyder Stromswold 1997)..
• Adult linguistic proficiency verbal fluency
  (e.g., Day, 1979), compound nouns (e.g., Gleitman
  Gleitman, 1970), sentence processing (e.g.,
  Corely Corley, 1995 Bever et al., 1989),
  second language acquisition (e.g., Fillmore,
  1979), grammaticality judgments (e.g., Ross,
  1979 Nagatu, 1992 Cowart, 1994).
• Caveat Genetic factors could account for
  differences among abnormal populations but not
  normal populations
• The case with number of fingers genetic
  syndromes associated with too many/few fingers
• (Contrast with heritability of finger length)

6
Methodology

• The power of meta-analyses
• Increase statistical power
• Methodological weaknesses of individual studies
  less worrisome
• Searched PsycINFO, ERIC, Medline databases for
• language, linguistic, articul, speech, read or
  spell AND
• hereditary, genetic, famil, twin, adoption,
  chromosom, linkage, pedigree, sex-ratio,
  segregation, aggregation, DNA, or RNA
• Excluded language disorders that were acquired,
  progressive, syndromic, or secondary to hearing
  loss, mental retardation, psychiatric/neurological
  disorder etc.

7
Family Aggregation of Spoken Disorders

• Do language disorders aggregate (cluster) in
  families?
• Yes Meta-analyses of 18 studies revealed
• SLI probands are more likely to have a positive
  family history 46 (range 24-78) vs. 18
  (range 3-46)
• SLI probands have more impaired relatives than do
  controls 28 (range 20-42) vs. 9 (range
  3-19)
• Caveat Difficult to separate the role of genes
  vs. environment (Deviant Linguistic Environment
  Hypothesis)

8
DLEH Predictions are not Borne Out

• Language impairments sometimes skip generations
• Most severely impaired children dont come from
  families with highest incidence of impairment
• Parents who speak normally (but have history of
  language delay) are more likely to have
  language-impaired children
• Even in families with very high impairment rates,
  some family members are normal
• In most families, some siblings are impaired and
  others are not
• No relationship between birth order and
  probability of impairment
• Concordance is no greater between primary care
  provider and child other first degree relatives
• Language-impaired children dont always have the
  same impairment as their relatives

9
Pedigree Studies Modes of Transmission

• Autosomal Dominant (AD) Most probands will have
  1 impaired parent, and half of siblings will be
  impaired
• Autosomal Recessive (AR) Most probands will
  have 2 unaffected parents, and one quarter of
  siblings will be affected
• X-linked Recessive (XLR) Impaired males have 1
  bad gene, impaired females have two bad genes.
  Thus, the MF is nn2 (where n is the frequency
  of the disordered allelle)
• Most genetic language disorders arent SML.
  Review of lit shows
• One-third of probands have 1 affected parent
  Genetic heterogeneity or AD with high rate of
  spontaneous mutation, or incomplete penetrance or
  expressivity
• One-quarter of probands have 2 affected parents
  High assortative mating, very high incidence of
  language disorders, SML models are wrong
• One-third of siblings are impaired Either AR or
  genetically heterogeneous
• Sex ratios generally between 21 to 31. Even
  most extreme only 61 Not XLR

10
Colorado Adoption Project (CAP)

• Rationale If genes are important for a trait,
  adopted childrens abilities will resemble their
  biological relativesabilities. If environment
  is important, adopted children will resemble
  their adopted relatives.
• Design Large (Ngt300) longitudinal study that
  compares adopted and nonadopted childrens skills
  with those of their biological and adopted
  parents and siblings.
• Language Disorders (Felsenfeld Plomin, 1997)
  156 children at age 7
• Positive biological family history was the best
  predictor of language disorders
• 25 of children with biological family history
  were impaired (9 with adopted FH)
• Sibling comparisons at age 7 (Cardon et al.,
  1992)
• Vocab verbal fluency h .90 (IQ-related .46,
  Language-specific .83)
• Fluency only h .33 (IQ-related .54,
  Language-specific .20)
• Vocab only h .47 (IQ-related .69,
  Language-specific .00)

11
CAP Parent-Child Comparison (Plomin et al., 1997)
Verbal Abilities
Spatial Abilities
Processing Speed
Recognition Memory
12
CAP Conclusions

• Heritable factors affect verbal abilities more
  than other types of abilities
• The influence of genetic factors becomes more
  apparent with age
• Specific-to-language factors are only seen at age
  7 (but this may be because overall IQ was used).
• Caveats about adoption studies
• All studies from a single group of children (what
  if not representative)
• Verbal assortative mating was greater for
  adoptive parents than biological parents This
  probably lowered heritability estimates
• Selective adoptive placement was not a problem
  (low correlation between adoptive and biological
  mothers verbal skills)

13
Twin Study Rationale

• Rationale Identical (monozygotic, MZ) and
  non-identical (dizygotic, DZ) twin pairs share
  the same environment, but MZ cotwins share 100
  of their DNA, whereas DZ twins share 50 of their
  DNA
• Therefore If MZ cotwins are more similar
  linguistically than DZ twins, this suggests that
  genetics plays a role in language.
• Can quantify the relative role of genetics and
  environment by measuring how much more similar MZ
  twins are than DZ twins.

14
Concordance Rates for Twin Pairs

• Are concordance rates for MZ gt DZ twins?
• Number of Impaired Individuals in Concordant
  Pairs
• Total Number of Impaired Individuals
• Two types of meta-analyses
• Mean rates Treat each studies MZ DZ
  concordance rates as data points, and use sign-
  and t-tests to determine if there is a
  significant difference.
• Overall rates Pool data from all studies and
  calculate overall concordance rates. Use
  Z-scores to test if MZ-DZ rates are different

15
Twin Correlational Analyses

• Are MZ twins test scores more highly correlated
  than DZ twins?
• Phenotypic variance variation for a trait in a
  population
• Heritable factors Falconers h2 2rMZ - rDZ
• Common environment factors c2 rMZ - h2
• Non-shared environmental factors e2 1 - rMZ
• Unweighted meta-analysis rMZ and rDZ are data
  points
• Weighted meta-analysis Weighted mean Fishers
  zs for MZ and DZ twins were calculated and
  compared using Z-scores

16
Other Genetic Analyses

• DeFries-Fulker Extreme Analysis When impaired
  people are ascertained by deviant scores,
  cotwins scores on the same test will regress
  toward the mean score of an unselected
  population. If genes play a role, DZ cotwins
  scores will regress more than MZ cotwins.
• Generalized DF Analysis extension for
  unselected populations
• If heritability estimate for language-impaired
  twins (h2g) is greater than for general
  population (h2), this indicates that certain
  genes contribute to the linguistic variance
  observed among language disordered people, but
  not for the variance in the general population
• Bivariate heritability Twins performance on
  test A is compared with that of his cotwin on
  test B. If rMZ is greater than rDZ, the
  phenotypic similarity on two tests is the result
  of genetic factors (but maybe not the same
  genetic factors)
• Genetic correlation (rG) Do the same genetic
  factors affect A and B?

17
MZ Concordance Rates are Higher

• Spoken language disorders 5 studies (266 MZ,
  161 DZ pairs)
• Mean 84 for MZ, 52 for DZ, p lt .0001
• Overall 84 for MZ, 48 for DZ, p lt .0001
• Written language disorders 5 studies (212 MZ,
  199 DZ pairs)
• Mean 76 for MZ, 41 for DZ, p lt .01
• Overall 75 for MZ, 43 for DZ, p lt .0001
• Combined spoken/written disorders (478 MZ, 360 DZ
  pairs)
• Mean 80 for MZ, 46 for DZ, p lt .0001
• Overall 80 for MZ, 46 for DZ, p lt .0001
• But why arent MZ concordance rates 100? Three
  possibilities
• MZ twins arent identical genetically and/or
  environmentally
• Expressivity of language disorders is incomplete
• Failure to diagnose language disorders in some MZ
  cotwins
• DZ pair-wise concordance rate (26) is similar
  to non-twin siblings (30)

18
Spoken Language Disorders
19
Written Language Disorders
20
SLI Twins Test Performance

• Bishop et al (1995) 63 MZ, 27 DZ twin pairs
• Articulation Falconers h2 1.82
• Phonological STM DF h2g 1.25
• Receptive vocabulary DF h2g 1.35
• Morphosyntax Wechsler h2g 1.10, CELF h2g
  .56, TROG h2g 1.09
• (But when nonverbal IQ partialled out, no
  significant genetic effects)
• Bishop et al (1999)
• 27 MZ, 21 DZ Pure tone sequence repetition DF
  h2g .11
• 25 MZ, 22 DZ Nonword repetition DF h2g 1.17
• Tomblin Buckwalters (1998) data minus triplets
  (58 twins)
• Falconers h2 .66, p .05
• Bivariate heritability for nonverbal IQ
  language .21
• Genetic correlation, RG .01 (ie., different
  genetic factors influence verbal nonverbal
  disability)

21
TEDS Twins Test Performance

• TEDS study Large population-based, parent
  report twin study
• Dale et al (1998) Analyzed data for twins with
  the smallest vocabularies (bottom 5tile, 135
  twin pairs). DF h2g .73 (vs. h2 .25 for all
  TEDS twins)
• Eley et al. (1999) DF h2g greater for TED twins
  with small vocabularies than twins with normal
  vocabularies.
• Eley et al. (2001) genetic continuity is
  greater for small vocab probands than other
  proband groups
• Purcell et al. (2001) Are the genetic factors
  specific to vocabulary?
• When probands were selected based on small
  vocabularies, RG for low verbal nonverbal
  scores 1.0 (i.e., the genetic factors that
  cause 2 years olds to have small vocabularies are
  the same as those that cause them to have
  nonverbal delays.)
• When probands were selected based on poor
  nonverbal scores, the vocabulary-nonverbal RG
  .36
• Why the asymmetry Differences in homogeneity of
  the samples? Problems with the measure?
  Directionality of effect?

22
Colorado Twin Study of Reading Disability

• Olson et al (1989) Genetic factors played a
  large role for phonological reading (DF h2g
  .93) but not orthographic reading (DF h2g
  .-.16).
• Light et al. (1998). DF h2g for phonological
  reading .52 overall reading .70
• Castles et al (1999) Genetic factors account for
  twice as much of the variance in phonological
  dyslexics as orthographic dyslexic (67 vs. 31)
• Gayan Olson (1999) contra Castles et al.
  (1999) argue that heritable factors play a
  significant role in all types of dyslexia.
• Olson et al. (1999) Wadsworth et al. (2000)
  genetic factors play a greater role in reading
  disability among children with high IQs than low
  IQs
• Light et al. (1998) RG for overall reading/math
  .36 (60 due to genetic factors common with IQ
  and 20 due to genetic factors common to
  phonological reading)

23
Summary Twin Language Disorders

• Some language disorders are genetically based
  (25-100)
• Genetic factors probably affect the linguistic
  abilities of disordered populations more than the
  general public (75 vs. 25)
• Genetic language disorders seem to impact
  different aspects of language, but less is known
  about phonology, morphology syntax
• Unknown if the same genetic factors cause
  different types of language disorders (and even
  if they do, what would this mean?)
• Unclear if the genetic factors identified are
  specific to language
• The few existing studies have conflicting
  results, possibly reflecting aspects of language
  assessed, methods of assessing etc.

24
Normal Twin Vocabulary Studies

• Overall 8 studies with 1577 MZ, 1389 DZ twins
• Unweighted mean rMZ .81, rDZ .57,
  Falconers h2 .48 (p .002)
• Weighted mean rMZ .93, rDZ .76, Falconers
  h2 .33 (p lt .0001)
• Early 3 studies with 1247 MZ, 1152 DZ twins
  18-24 months old
• Unweighted mean rMZ .91, rDZ .78,
  Falconers h2 .26 (p .08)
• Weighted mean rMZ .95, rDZ .80, Falconers
  h2 .29 (p lt .0001)
• Late 5 studies with 330 MZ, 237 DZ twins 3-13
  years old
• Unweighted mean rMZ .75, rDZ .44,
  Falconers h2 .62 (p .001)
• Weighted mean rMZ .71, rDZ .45, Falconers
  h2 .53 (p .02)
• Role of genes increases with age (2 long. studies
  meta-analysis)
• Unclear whether genes are specific to language (1
  study yes, 1 study no, 1 longitudinal study
  yielded different results at different ages)
• Different genes affect normal impaired twins
  vocabulary
• Bottom 5tile TEDS complete genetic overlap for
  vocab nonverbal skills.
• For all TEDS twins, vocabulary-specific genetic
  factors exist

25
Normal Twin Vocabulary Studies
26
Normal Twin Phonology/Articulation

• Phoneme Discrimation 21 pairs of 2-3 year olds
  (Fischer 1973)
• rMZ .64, rDZ .53, Falconers h2 .22 (p gt
  .10)
• Phonemic Awareness 126 pairs of 6-7 year olds
  (Hohnen Stevenson 1999)
• Weighted mean rMZ .90, rDZ .56, Falconers
  h2 .68 (p lt .001)
• Age 6 29 IQ-related genetic factors, 23
  vocab/morphosyntax, 9 phonology
• Age 7 18 IQ-related genetic factors, 67
  vocab/morphosyntax-related
• Phonological STM 100 pairs of 7-13 year old
  twins (Bishop et al. 1999)
• Heritable factors do not affect the ability to
  repeat sequences of pure tones
• Heritable factors do affect the ability to repeat
  nonsense words (h2 .71, p .01)
• Articulation 180 pairs of 3-8 yrs (Matheny
  Bruggemann 73, Mather Black 84)
• Weighted mean rMZ .93, rDZ .79, Falconers
  h2 .26 (p .03)

27
Normal Twin Phonology Articulation
28
Normal Twin Morphosyntax

• 12 twin studies of children between 20 months
  12 years.
• Diversity of methods used and aspects of
  morphosyntax assessed precludes combining data
  from these studies, but
• rMZ significantly greater than rDZ for all
  measures in 5 studies, 2/3s of measures in 1
  study, and 1/2 of measures in 2 studies. In 4
  studies, MZ-DZ differences were not significant
  in majority of measures
• rMZ gt rDZ in 33 of 36 measures, p lt .0001
• Mean rMZ gt rDZ for each of 12 studies, p lt .0001
• Significant differences were more common in
  larger studies and in studies that used cleaner
  measures of morphosyntax
• Do language-specific genes exist?
• Munsinger Douglass (1976) MZ-DZ difference
  significant even when nonverbal IQ partialled out
• Hohnen Stevenson (1999) Syntax-specific genes
  account for 20-30
• Dale et al. (1999) Genetic factors are specific
  to language but not syntax (however,
  parent-report syntax measure is worrisome)
• No evidence that influence of genetics increases
  with age

29
Normal Twin Morphosyntax
30
Normal Twin Written Language

• Reading 5 studies with 745 twin pairs
• Weighted mean rMZ .86, rDZ .66, Falconers h2
  .45, p .002
• Hohnen Stevenson (1999) Some genetic factors
  are specific to language but not reading
  (20-30), with a modest amount specific to
  reading (20-30). Genetic factors common to IQ
  have only a modest effect (10), and genetic
  factors specific to phonemic awareness have no
  effect on normal childrens reading (c.f.,
  dyslexia findings)
• Spelling 2 studies with 246 twin pairs (Osborne
  et al, 1968 Stevenson et al, 1987)
• Weighted mean rMZ .78, rDZ .48, Falconers h2
  .60, p .002
• Stevenson et al. (1987) Heritabilty estimates
  are greater for IQ-adjusted scores than
  non-adjusted scores (.73 vs. .53)

31
Normal Twin Reading
32
Summary of Twin Results

• Genetic factors play a greater role for
  language-impaired people (1/2 -2/3) than
  normals(1/4-1/2)
• Genetic factors affect all aspects of language
• Probable existence of some language-specific
  genes
• Possible existence of some genes specific to
  different aspects of language

33
Potential Worries With Twin Studies

• Gene/environment interactions the
  generalizability of heritability estimates
  obtained from twins
• Twins have higher rates of impairments/delays
  than singletons
• Twins have impoverished prenatal postnatal
  environments
• Environmental assumptions
• Prenatal Do MZ and DZ twin pairs have the same
  prenatal environment?
• Postnatal Do MZ and DZ twin pairs have the same
  postnatal environment?
• gt Swedish Separated-at-Birth Twin Study
  (Pedersen et al 1994) yielded similar
  heritability estimates for reared apart twins as
  is found for reared together twins
• Genetic assumptions
• Are MZ twins genetically identical?
• Are DZ twins genetically equivalent to siblings?

34
Linkage Studies Background

• Naming conventions
• Humans have 22 pairs of autosomal 2 sex (Y, X)
  chromosomes
• Autosomal chromosomes are numbered from 1-22 by
  size (1 is largest)
• Each chromosome has a constriction short arm
  (p) long arm (q)
• Thus, 15q21 refers to staining band 21 on long
  arm of chromosome 15
• Multiplex analyses Compare DNA of affected and
  unaffected family members in highly affected SML
  transmission families. Do marker locus and trait
  locus assort independently or is there decreased
  recombination (indicating 2 loci are neighbors)?
  Logarithm of odds score gt 3 indicates linkage.
  LOD of -2 indicates no linkage. Problem
  multiplex analyses reveal genes that can cause
  SLI, but rarely do
• Sibling pair analyses Compare DNA of affected
  and unaffected siblings. If a trait locus is
  closely linked to a marker locus, similarity
  between siblings for the marker alleles should
  correspond with phenotypic similarity, regardless
  of the mode of transmission (i.e., works with
  non-SML disorders)

35
(No Transcript)
36
The KE Family
KE family Multiplex family with AD disorder
that includes grammatical deficits (Gopnik
1990), oral dyspraxia (Fisher et al. 1998, Hurst
et al. 1990), and low nonverbal IQ and nonverbal
learning disorders (Vargha-Khadem et al. 1995)
37
7q31 Loci for Spoken Impairments

• Fisher et al. (1998) Disorder in KE family is
  linked to 7q31
• Tomblin et al. (1998) Linkage of SLI with 7q31
  in a population-based study of second graders
• Lai et al. (2000) The disorder is linked to
  7q31.2 in affected KE family members and an
  unrelated person with a similar disorder
• Lai et al (2001) All and only affected family
  KE members have an abnormal form of the FOXP2
  gene. The gene codes for transcription factor,
  and is highly expressed in fetal tissue and its
  homologue is found in mouse cerebral cortex.
• Enard et al. (2002) The FOXP2 homologue in
  non-human primates (and mouse) differs from that
  of humans.
• Genetic link between 7q31 and Tourette Syndrome
  and autism

38
Other Loci for Spoken Impairments

• Froster et al. (1993) Family with 1p22 and 2q31
  translocation associated with written spoken
  impairments
• Elcioglu et al. (1997) Isolated case of severe
  language delay but normal nonverbal abilities
  Inverted duplication of 15q13-gt15q2.
• Bartlett et al. (2000) 19 multiplex families
  with linkage near 4 dyslexia loci (1p36, 2p15,
  6p21, 15q21). No linkage to 7q31
• Cholfin et al. (2000) Multiplex family with AD
  transmission, but no linkage to 7q31
• SLI consortium (2002) 98 siblings. 16q24
  (nonword rep), 19q13
• Bartlett et al. (2002) 5 Canadian multiplex
  families 13q21

39
Going from Loci to Genes

• SLI Lai et al. (2001) FOXP2 transcription
  factor gene.
• Dyslexia possible candidate genes
• 1p34-p36
• 2p15-p16 phosphotase calcineuron (psychiatric
  disorders)
• 6p21-p23 HLA (autoimmune), GABA-beta receptor 1
  (CNS inhibitor), lyso-phospholipid coenzyme A
  acyl transferase (fatty acid and membrane
  phospholipid metabolism gene), human kinesin gene
  (C elegans mutant have behavioral disorders) .
• 6q13-16.2
• 15q21-q23 beta2-microglobin gene (autoimmune)
  neuronal tropomodulin 2 3 (a major binding
  protein to brain tropomyosin)
• 11p15.5 dopamine D4 receptor gene

40
What we dont know . Phonology

• Do genetic factors affect phonology (vs.
  articulation)?
• Stromswold Ganger (in prep) analysis of
  monthly spontaneous speech samples (22-47 mo)
  from 8 sets of normal twins
• Size of phonetic inventory is not more similar
  for MZ cotwins
• Order of acquisition of phonemes is more similar
  for MZ cotwins
• Accuracy rate is more similar for MZ cotwins
• Syllable initial 7.7 vs. 16.4
• Syllable final 9.9 vs. 15.9
• Patterns of errors may be more similar for MZ
  cotwins
• Substitution rates similar, but MZ cotwins more
  likely to make the same substitutions
• MZ cotwins more likely to make the same classes
  of substitution errors (e.g., fronting, voicing
  errors, stopping)
• Deletion rates more similar for MZ than DZ
  cotwins

41
What we dont know . Syntax

• To what extent do genetic factors play a role in
  syntax?
• Published syntax studies generally are small
  and/or use worrisome measures
• To do large-scale studies, we need a syntax test
  that parents can administer
• The Parent Assessment of Language (PAL)
• Weve designed and are norming a series of
  parent-administered test for children ages 3 and
  above. Each years PAL tests childrens
  comprehension of syntactic constructions that
  children are mastering at that age (actives,
  passives, reflexive, pronouns, relative
  clauses,modals, subjunctives, subject and object
  control structures, etc.).
• Longitudinal twin study using the PAL (current N
  120)

42
PAL Syntax Items (Picture-pointing)

• Age 3 Age 4
• 4 Full Actives The bear licked the dog 4 Full
  Passives The bear was licked by the dog
• 2 Easy Reflexives The bear licked himself 2
  Easy Pronouns The bear licked him
• Age 5 Age 6
• 1 Full Active The bear was licking the dog 1
  Truncated Active The bear was licking
• 3 Full Passives The bear was licked by the
  dog 3 Truncated Passives The bear was licked
• 2 Easy Reflexives The bear was licking
  himself 2 Easy Pronouns The bear was licking
  him
• Ages 7 8
• 1 Full Active The bear was licking the dog 1
  Truncated Active The bear was licking
• 3 Full Passives The bear was licked by the
  dog 3 Truncated Passives The bear was licked
• 1 Easy Reflexives The bear was licking
  himself 1 Med Reflexive The dog's friend was
  licking himself
• 1 Easy Pronoun The bear was licking him 1 Med
  Pronoun The dog's friend was licking him
• Ages 9 10
• 1 Full Active The bear was licking the dog 1
  Truncated Active The bear was licking
• 3 Full Passives The bear was licked by the
  dog 3 Truncated Passives The bear was licked
• 1 Med Refl. The dog's friend was licking
  himself 1 Hard Refl The friend of the dog was
  licking himself
• 1 Med Pronoun The dog's friend was licking
  him 1 Hard Pronoun The friend of the dog was
  licking him
• Age 11 and above
• All preceded by One of these two dogs is hot and
  followed by the query Which dog is hot?

43
(No Transcript)
44
(No Transcript)
45
(No Transcript)
46
(No Transcript)
47
(No Transcript)
48
(No Transcript)
49
PAL Syntax Yes/No/Maybe Task (Ages 9)

• Sally said Shouldnt you make the knot tight?
  Did Sally think the knot should be tight?
• Billy wont go to the park unless John goes.
  Will Billy stay home?
• Katie promised Lucy, who was thirsty, to buy
  juice. Did Lucy say she would buy juice?
• Mary who is going to the party with Steve does
  not like to dance. Does Mary enjoy dancing?
• Michaels cat chased the mouse and ran away.
  Did Michaels cat run away?
• Jim thinks Tom is bad at sports. Is Tom bad at
  sports?
• Maybe the band would have played last night if
  the drummer hadnt quit. Did the band play last
  night?
• The doctor who was looking for the nurse walked
  home from the hospital. Did the doctor walk
  home from the hospital?

50
What we dont know . Specificity

• Do language-specific genes exist?
• Need more, large studies that assess development
  in many different areas (not just cognitive
  abilities, but also fine motor, gross motor, oral
  motor, social etc.)
• Do specific genes for different aspects of
  language exist? (e.g., syntax-specific,
  phonology-specific, lexicon-specific)
• Need to assess multiple aspects of language in a
  large group of children
• Data that we are collecting in our twin study
• PAL assesses articulation, lexical access,
  reading/pre-reading, and syntax
• ASQ parent assessment of gross motor, fine
  motor, cognitive, language social-emotional
  skills
• Developmental milestones (gross motor, fine
  motor, cognitive, language, social)
• Special educational/therapy services
• Neuropsychological diagnoses

51
Sample PAL (Age 4)
Articulation of onsets
List any sounds the child regularly says wrong,
and give a typical mispronounced word
Lexicon Rapid naming (number of foods named in
30 seconds) Pre-reading Capital letter
identification. (Orthographic and phonologic
word reading starting at age 6 PAL.) Syntax
Picture pointing comprehension task 4 actives
(e.g., the dog licked the bear) 4 passives
(e.g., the fox was tickled by the lion) 2
reflexives (e.g., the cat scratched himself) 2
pronouns (e.g., the monkey splashed him)
52
What we dont know Gene x Environment

• Koeppen-Schomerus et al. (2000) Heritable
  factors play a negligible role in linguistic and
  cognitive abilities of very premature TEDS twins.
  
• What is the relative importance of prenatal and
  postnatal environment?
• We are comparing heritability estimates for twins
  with easy/hard prenatal courses
• Gestational age
• Birthweight
• Birthweight percentile
• Brain injuries
• Short (discharged before or by due date ) vs.
  long hospital stays
• Composite neonatal morbidity measure
• We are comparing heritability estimates for twins
  with different postnatal environments (SES,
  therapeutic interventions, traditional vs.
  developmental NICUs)
• Are there specific perinatal factors that place
  twins at risk (e.g., steroids, MgSO4,
  intrauterine infection, placental infarction,
  ventilation, TTTS, etc.)?
• Quantifying the role of prenatal environment we
  are comparing outcomes for MZ twins with very
  similar birth weights and very different birth
  weights (MZS -MZD ) and DZ twins with
  similar/different birth weights

53
What we dont know Going from genes to disorders

• The genotype to phenotype mapping problem
• One GenotypeMany Phenotypes / One PhenotypeMany
  Genotypes
• The developmental problem phenotypes change
• Direct vs. indirect genetic effects The case of
  clotting disorders
• Indirect If a mother has a genetic clotting
  disorder, her children are at risk even if they
  not carry the mutation.
• Direct A child with a genetic clotting disorder
  is at risk for perinatal strokes (and the
  language areas of the brain are particularly
  vulnerable)
• Maternal/child interactions possible when both
  have the disorder
• Environmental interactions high estrogen, low
  folic acid, delayed child-bearing
• Specificity problem
• Familial Dysautonomia (9q31 IKBKAP). AR
  disorder with normal IQs and profound oral motor
  dyspraxia (but they also have ANS problems)
• FOXP2 Do people with 7q31-linked autism and
  Tourette Syndrome have the FOXP2 mutation?
• Just so stories

54
What if Language is Like Height?

• Quantative Trait Loci (QTLs) Multifactorial-polyg
  enic
• Hypothesis In normal people (and in most
  language-impaired people), variance in linguistic
  ability results from many genes (each of which
  has a small effect) acting together and in
  combination with the environment. Thus,
  linguistic abilities are normally distributed,
  and the observed heritability is due to QTLs
• How to find language QTLs
• People practice linguistic assortative mating.
• Assortative mating increases genetic variance in
  successive generations
• Assortative mating additive genetic variances
  makes QTLs easier to find
• It is easier to detect QTLs by looking at the
  high end of the distribution (at the low end,
  random mutations environmental insults obscure
  QTL effects)
• Linguists (particularly second generation
  linguists) should donate their DNA

55
(No Transcript)