Anne Fernald - US grants

The research proposed here investigates the role that the prosody, or intonation, of mothers' speech to infants plays in the development of prelinguistic communication. When adults speak to infants, they use higher pitch and relatively smooth, simple and highly modulated intonation contours, when compared with normal adult speech. This research focuses on the developmental functions of this characteristic maternal speech style, known as "motherese". The first study examines the selective responsiveness of infants to motherese intonation, testing the hypothesis that infant cardiac orienting is greater to pitch contours typical of motherese than to pitch contours typical of normal adult speech. The second study tests the hypothesis that mothers consistently use specific intonation patterns to convey specific affective and pragmatic messages to prelinguistic infants. The proposed project has 3 general objectives: 1) to develop experimental methods for use in the study of language input, a research area that has traditionally relied on observational and descriptive methods; 2) to provide new convergent evidence on the role of maternal intonation in the development of communication and language; and 3) to use the proposed exploratory studies as a basis for developing an integrated future research program on the role of affective processes in the development of language.

The research proposed here focuses on affective processes in the development of preverbal communication and language. I will conduct an integrated series of studies with infants and adults to test the following hypotheses: a) The expression of positive affect in mothers' and fathers' speech to infants is associated with elevated pitch and characteristic exaggerated intonation contours; b) Young infants are selectively responsive, both behaviorally and psychophysiologically, to these exaggerated pitch contours in parental speech; c) The highly modulated, affective pitch contours in parental speech serve to direct the attention of the preverbal infant to objects and events in the environment; d) Parents consistently use specific intonation patterns to convey specific affective and pragmatic messages to infants; these context-specific affective vocalizations may function as the first units of vocal meaning for the preverbal infant; e) The exaggerated, highly affective pitch contours of parental speech ultimately help the infant to associate speech sounds with their referents, facilitating lexical acquisition and language comprehension. I want to continue efforts to integrate the study of preverbal communication and language development with the study of perceptual and affective processes in infancy. Traditional conceptual boundaries have separated research in auditory and speech perception, language acquisition, emotional development, and parent-infant social interaction. An interdisciplinary approach to the ontogeny of communication, as proposed here, should help to integrate these areas. I hope to provide further support for the claim that infants early affective responsiveness to parental intonation prepares the child for the later linguistic use of intonation to parse the speech stream and understand spoken language. This continuity between the infant's processing of affective ad linguistic information in parental speech is the general hypothesis that motivates and unifies the studies proposed.

The studies proposed here investigate the universality and developmental functions of the exaggerated intonation patterns characteristic of adult speech to infants. This research is motivated by three central hypotheses: 1) The use of exaggerated intonation in speech to infants is an adaptive human parenting behavior, which serves to modulate infant attention, communicate affect and facilitate language development; 2) These characteristic prosodic patterns in speech to infants will be observed in non-Western as well as European cultures; 3) Certain stereotypical intonation patterns in mothers' speech, common across culture, come to function as the first regular sound-meaning correspondences appreciated by the pre-verbal infant. The first goal of the project is to conduct a series of cross=language field studies of adult speech to infants in Japanese, Italian, Hausa, Spanish, Black American English, Tongan, and Samoan. These observational studies will investigate common patterns as well as language-specific variations in the prosodic modifications used in mothers' speech to infants in these diverse languages and cultures. The second goal is to conduct a series of cross-language experimental studies with adult and infant subjects, to test tow hypotheses within and across languages: 1) When mothers' speech to infants is filtered to remove linguistic content, adult listeners can correctly identify the communicative intent of the speaker based on intonation patterns alone, even in an unfamiliar language; 2) Pre-verbal infants are also selectively responsive to context-specific intonation contours in mothers' speech. This research strategy combines cross-cultural breadth with experiment rigor to yield convergent evidence of both language-independent and language-specific features of parental speech to infants, and on universal processes in the transition from preverbal to verbal communication.

When speaking to infants, adults across diverse cultures and languages use a special speech form characteristed by exaggerated intonation and other prosodic modifications. The use of infant-directed prosody appears to be a universal human parenting behavior which serves important affective and linguistic functions in early development. The long-term goal of this research project is prosody in maternal speech to the normal development of affective and linguistic communication in infancy. The central hypothesis motivating the proposed studies is that the melodies of mothers' speech to infants help the child gain access to meaning in adult speech, initially through emotion in the first year of life, and gradually through language in the second year. This project has three specific aims: 1) The first objective is to provide detailed descriptive data on the acoustic characteristics of infant-directed prosody. The first study will identify spectral and temporal characteristics of infant-directed vocalisations which contribute to their effectiveness as emotional signals. The second study will focus on acoustic cues which help the infant to identify linguistic units within the speech stream. 2) The second objective is to investigate experimentally how prosody in speech to infants elicits and communicates emotion, and modulates infants' attention. Extending our recent findings on young infants' responsiveness to vocal affect, one series of studies will investigate how adult vocalisations expressing positive and negative affect influence infants' emotion and behavior in uncertain situations. A second series of studies will investigate the hypothesis that the use of exaggerated intonation in speech to infants functions initially as an attentional device, and later as a clarification strategy, as the infant begins to acquire language, 3) The third objective is to investigate experimentally how prosodic cues facilitate speech processing and language comprehension in infants and young children. Several experiments will focus on the influence of prosodic features on infants' ability to recognize words, as well as syntactic units, in prosodic cues such as pitch and duration in speech processing. Both observational and experimental methods will be used to investigate how maternal vocalisations first become meaningful to infants through emotion and through language. The findings from this research will increase our understanding of how universal human parenting behaviors contribute to the early development of communication.

[unreadable] DESCRIPTION (provided by applicant): To understand language, children must learn to listen for meaning in sequences of speech sounds that unfold rapidly in time. Fluent understanding requires "listening ahead"; continually anticipating what is coming next by using linguistic and nonlinguistic information from the context in which words are spoken. Skilled adults draw on multiple sources of knowledge in order to process speech with such remarkable speed and efficiency, but little is known about the early development of this critical capacity. The first goal of the research is to investigate the development of spoken language understanding by English-learning children from 18-30 months, using powerful new methods for measuring the time course of word recognition by monitoring children's eye movements as they listen to speech. Efficiency in comprehension increases as the child learns to take advantage of regularities in the structure of continuous speech, enabling the child to anticipate upcoming words and relate word meanings to one another. The three general hypotheses motivating the proposed experiments are: 1) that infant-directed (ID) speech provides the inexperienced listener with perceptual regularities useful in early speech processing; 2) that children use their emerging knowledge of linguistic regularities in the language they are learning to facilitate word recognition; 3) that the rapid processing of predictable elements in the speech stream increases children's capacity to attend to and learn about novel elements. The first l0 experiments ask whether early word learning is enhanced by common features of ID speech such as exaggerated intonation and one-word utterances, and how children make use of language-specific cues in English such as sentence stress, morphosyntactic regularities, word order, and semantic information in understanding fluent speech. The second goal of the project is to integrate laboratory studies of infant speech comprehension with cross-language field research on ID speech, extending this research to Spanish. An observational study of Hispanic mothers' speech to infants will inform the hypotheses to be tested in 7 experiments on speech processing by Spanish-learning children from 18 to 30 months. These experiments explore how infant word recognition is influenced by characteristic features of Spanish ID speech, and how Spanish-learning children make use of language-specific cues such as gender and number-marking, word order, and semantic information in understanding fluent speech. The third goal is to refine newly developed experimental procedures for monitoring on-line speech comprehension by young language learners, and to evaluate their potential clinical relevance for assessing disorders in early language development.

[unreadable] DESCRIPTION (provided by applicant): To understand language, children must learn to listen for meaning in sequences of speech sounds that unfold] rapidly in time. Fluent understanding requires "listening ahead," continually anticipating what is coming next by using linguistic and nonlinguistic information from the context in which words are spoken. Many studies have shown how skilled adults draw on multiple sources of knowledge to process speech in real time with remarkable speed and efficiency (Tanenhaus & Trueswell, 1995). Recent studies in our laboratory using the "looking-while-listening" task have documented the dramatic gains in speed and accuracy of speed processing abilities in English-language learners over the 2nd-3rd yrs of life (Fernald et al., 2001), and the robust links between efficiency of speech processing and individual differences in the growth in lexical an grammatical abilities (Perfors et al., under review). In addition, we have embarked on exciting crosslinguisti studies examining these online language skills in children learning Spanish as a first language. The studies proposed in this supplemental application extend this line of research to a longitudinal sample of children learning both English and Spanish at the same time. We will apply newly developed measures to examine language background in this important and growing population, as well as statistical techniques (HLM) that are particularly suited to analyses of longitudinal samples. The proposed project addresses three goals: (1) To document developmental increases in speed/accuracy of processing in children learning both English and Spanish and to compare these trajectories to those of participants in our ongoing study of monolingua Spanish-learners, matched on SES and vocabulary size. (2) To examine relations between spoken language processing and offline measures of vocabulary in bilingual children. Results could indicate that the relation: previously reported for monolinguals reflect representations or processing mechanisms that are shared in the context of constructing a particular body of language knowledge (e.g., English or Spanish). Alternatively, we might find that language-processing efficiency is more related to general measures of vocabulary, suggesting that processing efficiency is a stable characteristic of an individual that facilitates vocabulary growth across different language-learning contexts. (3) To address key theoretical issues about mechanisms of language acquisition that rely on strong associations between lexical and grammar learning (Bates & Goodman, 1997; Tomasello, 2003). Studies of children learning two languages are well-positioned to evaluate claims regarding the nature and specificity of the mechanisms guiding the development of language skills in young learners. [unreadable] [unreadable]

DESCRIPTION (provided by applicant): To understand language, children must listen for meaning in speech sounds that unfold rapidly in time. Researchers have begun to explore the emergence of early efficiency in language understanding in typically developing (TD) infants, but little is known about how speech processing abilities develop in children who are at risk for language delay (LD), the question that motivates this proposal. Two longitudinal studies with English- and Spanish-learning children at risk for language delay explore relations between online measures of processing efficiency and traditional measures of language and cognitive competence on experimental and standardized tests. Participants in Study I, (n=125) recruited at 18 months, include English-learning TD children as well as late talkers (LT), some who will later move into the normal range, and some who will remain at risk for language delay. These children are tested in the looking-while-listening (LWL) procedure to assess speech processing skill at 18, 24, 30, 36, & 48 months. Participants in Study II (n=126) are Spanish-learning children at risk for language delay, as well as bilingual children learning Spanish and English simultaneously, also recruited at 18 months including TD and LT children. These Latino children are assessed at 18, 24, 36, 48, & 60 months. One goal is to adapt online processing measures developed with TD children to focus on Language Processing Challenges (LPCs) in four critical areas of language competence: familiar word recognition; use of morphosyntactic information; semantic integration; and novel word learning, all areas in which LD children have difficulties. A second goal is to establish the validity of these new online processing measures for use with very young children in relation to traditional measures of lexical and grammatical growth and to standardized clinical instruments used to assess language delay. At each age children are tested in the online LWL procedure, yielding precise, sensitive, continuous measures of speed and accuracy in interpreting linguistic stimuli in real time. They are also assessed at each age on standardized tests, enabling us to examine concurrent relations between processing efficiency in each LPC and more traditional measures of language development. Our overarching goal is to determine which measures of language function are most powerful in predicting later outcomes in diverse populations. We use online measures to identify which particular processing problems may be early precursors or markers of later language impairment. Our longitudinal designs prospectively track growth in language skills, allowing us to determine to what extent processing efficiency in infancy can distinguish LT children who catch up from those who are at risk for language and learning difficulties. Although our major focus is on the clinical relevance of these measures for children at risk for language delay, Studies I & II will also enable us to explore the full continuum of receptive language skill in diverse populations of children learning Spanish and English, documenting individual differences in children who are highly competent as well as those who are having difficulty in spoken language processing.

DESCRIPTION (provided by applicant): Each year in the U.S., one in 8 children is born prematurely and the rate of preterm birth continues to rise. Children born preterm are at substantial risk for language-based learning disabilities that may not be identified until the child enters kindergarten, a delay that precludes opportunities for early intervention. Early behavioral indicators of adverse language and cognitive outcomes are urgently needed, particularly methods that identify underlying deficits amenable to effective interventions. To meet these pressing needs, we have forged an experienced interdisciplinary team with complementary expertise in pediatrics and developmental psycho-linguistics. Our mutual long-term goals are to develop reliable and robust predictors of later outcomes for children born preterm (PT) that allow us to design effective early interventions and to elucidate the neural basis of disorders associated with prematurity and with atypical language development. Recent longitudinal studies of typically-developing infants from the Fernald laboratory have found that online measures of linguistic processing efficiency in very young children predict individual differences in language outcomes several years later. Preliminary results from a joint Feldman-Fernald effort found that 18 to 30 month old PT children had no difficulties participating in the online procedures but that PT children were slower to initiate a response and less successful at maintaining their attention compared to age-matched controls. Therefore, we propose to use these sensitive online measures in a longitudinal study of the development of language processing efficiency in PT children from 18 to 48 mos (n = 70) relative to two SES- and gender-matched comparison groups: (1) full term age-matched children (FT-AM, n = 70) with the full range of language abilities, and (2) full term language- matched children (FT-LM, n = 70) who are matched individually to the PT children on age and language level. Children will be tested longitudinally at 18, 24, 30, and 36 mos on a series of increasingly challenging, but developmentally appropriate, assessments of online language comprehension and on standardized offline tests of cognition and language. We hypothesize that PT children will perform more poorly than both groups of FT peers, and that individual differences in online processing efficiency will differentiate those children with good outcomes from those who demonstrate persistent language or associated difficulties in late preschool. We also hypothesize that clinical and/or neurological factors, such as severity of white matter injury, will correlate with language and cognitive outcomes. However, those relations will be reduced after controlling for linguistic processing efficiency, suggesting that speed and accuracy of language processing mediate the link between the neuropathology of prematurity and poor language functioning in PT children. This innovative translational project adopts a developmental psycholinguistic approach to explain individual variation in outcomes in a clinical population at high risk for language disorders, and is poised to inform future theory- and empirically-driven intervention research as well as studies of the neural basis of language disorders.

DESCRIPTION (provided by applicant): Universal newborn hearing screening provides unprecedented opportunities to identify children with early hearing loss, estimated to affect 1 to 6 per 1,000 newborns in the U.S. today. Children with hearing loss are at chronic risk of underperforming in language and literacy, placing them at considerable risk for academic and learning difficulties. Research has shown that deaf infants exposed to naturally occurring signed languages from birth from deaf signing parents academically outperform deaf individuals who learn sign later in life from non-native hearing parents. Yet, there remains a great need for scientific evidence on early language development in children learning ASL, and no studies have explored the early real-time processing of signed languages by infants and young children using online measures. This project develops the first assessment of efficiency in language interpretation by infants and toddlers learning ASL, through a novel adaptation of the looking-while-listening (LWL) procedure developed by Fernald et al., (1998; 2008). Over the last decade, the Fernald group has explored how young hearing children develop crucial processing skills as they learn to listen for meaning in spoken language, discovering significant individual differences in processing efficiency that strongly predict later language outcomes. In extending this research to children learning ASL, we build on these extensive previous studies with monolingual and bilingual children learning spoken languages, as well as on pioneering studies from the Corina lab on ASL processing by deaf adults. Participants include two groups of 18-36 mo-old children learning ASL: (a) deaf children of deaf parents (deaf-of-deaf - DODs, n=25) and (b) hearing children of deaf parents (children of deaf adults - CODAs, n=25). CODAs will also have various amounts of exposure to spoken English. We also include a third comparison group of (c) hearing children learning only English (n=25). To assess real-time language processing efficiency, DOD and CODA participants will be tested in the LWL-ASL task adapted for use with sign stimuli; CODA children and hearing children learning only English will be tested on a comparable procedure using spoken English-language stimuli. Both tasks assess children's speed and accuracy in interpreting familiar signs/words, and also their skill at using knowledge of familiar signs/words to learn novel signs/words. Sign/word vocabulary will be assessed offline with parent report and face-to-face instruments. A sub-set of the children will be re-tested 6 months later. The proposed research will be the first study of the consequences of individual differences in early real-time sign comprehension in native learners of ASL, providing the opportunity to explore the role of variation in language exposure in sign processing efficiency in both deaf and hearing sign-exposed infants. This exploratory study will also lead to the development of potentially valuable new measures for use in future assessments of children who vary in types of sign language exposure, providing benchmarks for clinicians and caregivers who need to chart linguistic competencies in deaf children receiving early interventions.

An increasing number of Latino children in the US speak primarily Spanish in the home, and have minimal proficiency in English when they enter school. Because many of these emerging bilingual children are from lower-SES backgrounds, they are at increased risk for poor academic outcomes. The strongest single correlate of later success in school is a child?s oral language skills achieved before kindergarten. A question with critical public policy and educational implications is how family learning environments in a first language (L1) contribute to shaping language processing skills that could impact learning in the second language (L2). Our extensive work examining individual differences in the efficiency with which young children understand spoken language in real time has shown that early language processing efficiency is strongly linked to outcomes within a given language, but also that benefits extend to vital academically relevant skills such as working memory. Moreover, features of children?s early caregiver-child interactions shape the development of this critical information processing skill. Here, we seek to extend this important research program to emerging Spanish- English bilinguals, a rapidly growing and highly vulnerable population in the US. The two longitudinal studies proposed here offer the unique opportunity to explore both precursors and long-term consequences of having a strong foundation in early language processing skills in Spanish, asking whether early L1 processing skills support later language, cognitive, literacy and executive function (EF) skills in school. In Study I, children from our previous studies who participated in the LWL task at 2 years will be tested on a comprehensive battery of standardized tests at 5, 6 and 7 years (n=90). To what extent does individual variation in 2-year Spanish processing efficiency predict both later outcomes in Spanish and growth in emerging spoken or written skills in English? Does variation in early Spanish processing also link to later skills that support learning independent of language, such as EF, working memory, and reasoning? In Study II, Latino children (n=100) from monolingual Spanish-speaking families participate in the LWL task at 2 years and in assessments of language, cognitive, EF and pre-literacy skills at 2, 3 and 4.5 yrs. We also examine features of learning environments that support the development of language processing skills using laboratory play sessions and day-long naturalistic recordings. Embracing the heterogeneity of learning experiences in this population, we trace trajectories of exposure to English and delineate the quantity and quality of in-home and outside-the-home experiences which may moderate relations between early processing efficiency and later outcomes. We propose a model of cumulative risk such that late emerging bilinguals with less supportive learning environments are at the greatest disadvantage due to a weak foundation in language processing skill and later exposure to English. These studies address a critical gap in the literature and will have broad-reaching implications for adjusting early education policy to most effectively support learning in this growing and at-risk population.