Specific Language Impairment and the FOXP2

Specific Language Impairment and the FOXP2

An important (but difficult to investigate) question Are the brain mechanisms that process and represent language dedicated to language or employed in lots of cognitive domains? Relates to the larger question about modularity: Is the mind/brain organized into separate modules, responsible for different cognitive functions, or is the brain an general problem solver? 2

MODULARITY What s a module? Fodor: not assembled from more primitive processes of fixed neural architecture (specified genetically) domain-specific (a module computes a constrained class of specific inputs bottom-up, focusing on entities relevant only to its particular processing capacities) fast, autonomous, mandatory (a module's processing is set in motion whenever relevant data present themselves) automatic, stimulus-driven, and insensitive to central cognitive goals. informationally encapsulated. In other words, other parts of the mind can neither influence nor have access to the internal workings of a module, only to its outputs. Loss leads to characteristic pattern of breakdown. 3

Information Encapsulation - No access to information from other modules or central system - modules are cognitively impenetrable - Examples: - Even if you know that I would never poke you in the eye, if I put my finger too close to your eye, you will blink - Muller-Lyer illusion

MODULARITY Selective loss of a specific cognitive function via brain damage is support for the modularity view. But the modularity thesis doesn t necessarily predict that there should be specific brain areas dedicated to specific cognitive functions. So modularity could be true even if there was no selective language impairment via brain damage. Fodor: of fixed neural architecture (specified genetically) However, the way brains are is determined by genes and so selective loss or sparing of language genes should lead to selective loss or sparing or language. 5

MODULARITY What would we expect of a person who has lost their language? Or spared their language? 6

DEVELOPMENTAL DISORDERS WITH A GENETIC BASIS Specific Language Impairment (SLI): Language problems without any obvious cause. fifty to seventy percent of children with SLI have at least one other family member with the disorder Williams Syndrome (WS): (more or less) general cognitive retardation except in language Results from the deletion of the long arm of chromosome 7.! If the brain is a general purpose problem solver, one s general intelligence should be predictive of abilities such as language.! For people affected by WS or SLI that s simply not true. 7

SLI and genetics For the SLI population at large, we do not know what genetic mutations cause SLI. That there is a genetic basis to SLI is supported by the fact that it runs in families. The proportion of pairs where both twins are affected with SLI is significantly higher for identical (monozygotic single egg) than for non-identical (dizygotic separate eggs) twins. However, there is one famous SLI family for whom the genetic mutation is known, the KE family in Britain. The affected members of this family are clearly not just language impaired. 8

On the general SLI population 9

Specific Language Impairment (SLI) Developmental language disorder in the absence of neurological, sensori-motor, non-verbal cognitive or social emotional deficits Affects about 7% of the population, males more often than females A delay or deficit in the use of function morphemes. Omission of function morphemes long after age-matched children with typical language development show consistent production of these elements. Diagnosis based on behavioral evaluation but is likely to have a genetic basis. Control group is either age-matched, IQ-matched or languagematched (usually MLU(mean length of utterance))-matched. 10

Hypotheses about what goes wrong in SLI Gopnik: It s a grammar problem. SLI individuals are impaired in their ability to use grammatical morphemes. Wexler (& Rice): It s a grammar problem. SLI individuals are impaired in specific type of verbal inflection. Tallal: It s not a grammar problem! It s a problem in rapid auditory processing (which has consequences for language). Fundamental problem: Since the SLI individuals are a heterogeneous group, these research groups are not necessarily studying the same types of individuals. 11

Gopnik: problem with functional morphology 12

Gopnik: problem with functional morphology Inflection Derivation 13

Gopnik: problem with functional morphology Gopnik has data from lots of different languages and for lots of different types of morphology. Gopnik s hypothesis that SLI is an impairment in grammatical morphology predicts that SLI individuals should make lots of morphological mistakes. But they mostly seem to make one type of mistake, dropping morphemes. They don t swap past tense morphology for plural morphology, for example. This doesn t follow from Gopnik s account. Wexler has an approach to this, kinda. 14

Rice & Wexler: SLI are stuck in the Optional Infinitives Stage Wexler (1990, 1992, 1994): In language development, kids go through the Optional Infinitive (OI) Stage. lasts in normal children from birth to around 3;0. Infinitive: uninflected form of the verb. Some verbs select for infinitives as their complements (direct objects), but on their own, infinitives are ungrammatical. (i) OK: Daddy wants to walk to work. (ii) Bad: Daddy walk to work. In the Optional Infinitive Stage: a. Root infinitives (as in ii) are possible grammatical sentences b. These infinitives co-exist with inflected (finite) forms c. Apart from tense and agreement, the children know relevant grammatical principles 15

Rice & Wexler: SLI are stuck in the Optional Infinitives Stage In Dutch, infinitives are always sentence-final: I want shoes to wash. But inflected verbs come in second position, before the object: I washed shoes. A Dutch kid in the OI Stage: (8) pappa schoenen wassen daddy shoes wash-inf 'Daddy shoes wash (non-finite) Infinitival in final position, as it should be. Although the child substitutes the infinitive for the inflected form, he/she uses the infinitival grammar. (9) ik pak t op I pick it up 'I pick (fin) it up' Finite verb in second position, as it should be. 16

Rice & Wexler: SLI are stuck in the Optional Infinitives Stage Development of 47 normally developing Dutch children (Wexler, Schaeffer and Bol, 1999). 17

Rice & Wexler: SLI are stuck in the Optional Infinitives Stage Extended Optional Infinitive Hypothesis: Children with SLI are just like normal children except that they go through the OI period for a much longer time than normal children, perhaps never really emerging from it. 18

Rice & Wexler: SLI are stuck in the Optional Infinitives Stage Prediction for children with SLI: Children with SLI: a. Use OI's in languages where younger normal children do b. Show the same patterns of grammatical knowledge as normal children Wexler, Schaeffer and Bo: In the 6;00-8;02 year range, Dutch children with SLI still had 15% OI's (50 of 334). The OI stage persisted much longer in the children with SLI. 19

Rice & Wexler: SLI are stuck in the Optional Infinitives Stage Problem: SLI kids don t just have problems with verbal inflection. 20

SLI nonlinguistic hypotheses Tallal: SLI is a non-language specific deficit in the rate of auditory processing. Children with SLI have difficulty processing brief or rapidly presented auditory stimuli. 21

Evidence for auditory processing difficulty in SLI (Tallal) Normal children are able to discriminate two 75-msec tones separated by an interstimulus interval (ISI) as short as 8 msec, while individuals with SLI required an ISI exceeding 300 msec to perform the same discrimination at the same level of accuracy. 22

Evidence for auditory processing difficulty in SLI (Tallal) The degree of auditory processing deficit correlates with the degree of language comprehension deficit. Early deficits in rapid auditory processing abilities both precede and predict subsequent language delays: At 6-9 months, significant differences in auditory perception between infants born to families with a history of SLI and children without history of SLI Threshold for rapid auditory processing at 7.5 months is the single best predictor of language outcome. At age 3, rapid auditory processing threshold and being male, together predict 39/41% of the variance in language outcome. 23

Evidence for auditory processing difficulty in SLI (Tallal) So what would be the connection between auditory processing difficulty and having trouble with functional morphology? 24

Evidence for auditory processing difficulty in SLI (Tallal) Leonard: it s hard to hear little functional morphemes like ed or s! As a consequence, there are difficulties in learning morphological paradigms. Predicts crosslinguistic differences in SLI profiles. In languages where functional morphemes have more perceptual salience, SLI kids should have less trouble with them. Leonard shows that that s actually true in Hebrew and Italian. This is in contrast to Gopnik s findings that even phonologically salient morphemes cause trouble. 25

KE family 26

KE family Half of the KE family are affected by SLI. Female Male Shading = Language Impaired 27

KE family Vargha-Khadem et al (Science, 2001): In the affected members of the KE family, there is a mutation in the gene FOXP2. A language gene? 28

KE family behavior on linguistic tasks 29

KE family behavior on linguistic tasks The language problems of the SLI members of the KE family are much more global than in the general SLI population. 30

KE family behavior Not just language problems. 31

KE family brains Watkins et al (2002): In motor and speech-related brain regions, significantly different amounts grey matter in affected family members as compared to unaffected and control groups, which did not differ from each other. Volume of caudate nucleus (part of the basal ganglia) reduced in affected members. Involved in regulating voluntary movements 32

FOXP2 a language gene? No, mice have it too and it s very similar. Good news: we can study it in mice. What would happen to mice if they were given the human version of FOXP2? What would happen to mice their FOXP2 underwent the mutation that SLI individuals have?

http://www.cell.com/supplemental/s0092-8674(09)00378-x

Mice with a humanized FOXP2 Cellular changes in the striatum, a part of the basal ganglia. Decreased exploratory behavior.! Altered ultrasonic (>20KHz) vocalizations.

Mice with SLI-mutated FOXP2 On isolation from the mother/nest, healthy pups emit ultrasounds that elicit retrieval by the parent (cf. crying baby). Mice lacking a functional Foxp2 do not produce these calls.

SLI upshot Messy when it comes the general group as the genetic basis is not known and different research groups may be studying different types of populations. The KE family offers an interesting research opportunity, but it is unclear to what extent the affected members of this family are representative of the SLI population in general. The gene that is mutated in the KE family, Foxp2, can be studied in mice, which allows systemic investigation of causal connection between Foxp2 and brain&behavior (in mice).