Rochelle S. Newman, Ph.D. - US grants

The modern household can be a chaotic place, full of noise from televisions, computers, and playing children. This proposal comprises two sets of studies that, taken together, will greatly enhance our understanding of how infants learn language in such an acoustically hostile environment. These studies investigate both how infants disentangle the speech of different talkers speaking simultaneously, and how much experience they need in order to then separate that fluent speech into individual words. Most research on infant language acquisition has taken place in quiet laboratory rooms with no outside distractions, but in the real world, speech directed to infants often occurs in the presence of competing acoustic signals. Most current theories of language acquisition suggest that children "bootstrap" their way into language comprehension by exposure to large amounts of language information. Infants' ability to separate speech from background noise limits the amount of this information they can use. Thus, studying these basic attention abilities provides an important underpinning for theories of language acquisition. In addition, understanding the degree to which infants are affected by background noise is relevant in evaluating child-care decisions, as settings involving greater numbers of children are likely to result in greater degrees of noise. If infants have difficulty selectively attending to a caregiver in noisy situations, infants placed in quieter environments may have an advantage in learning language. The first set of studies examines how infants learn language in non-ideal environments. Recent research has shown that infants, like adults, have the ability to separate streams of auditory information (Newman and Jusczyk, 1996). Little is known about the limits of this ability, however, and the studies described here examine this more fully. A second set of studies investigates the amount of exposure to language an infant requires to begin the language-learning process. Many theories assume infants need linguistic experience to pick out individual words from fluent speech, but how much experience is required remains unclear. The present studies explore this issue, as well as whether infants can learn language from non-interactive situations, such as watching television. These studies further explore the role of early attentional and segmental mechanisms on language development, with the goal of constructing a theoretical model of infant language acquisition, one which is grounded in a better understanding of the problems infants face outside of the laboratory.

Infants sometimes hear speech in quiet settings, but oftentimes they must separate speech from many other sounds in the background (the television, the dishwasher, traffic, other people talking, and so on). It is remarkable that infants can separate speech from background noises so effectively, especially given research indicating that infants are unable to use the kinds of perceptual cues that adults use when perceiving speech. With support of the National Science Foundation, Dr. Newman is conducting a series of experiments on infant speech perception. The experiments are designed to investigate two types of perceptual cues that infants are hypothesized to use in separating speech from background noises: 1) voice qualities associated with the gender of a speaker, and 2) phonetic qualities associated with the language being spoken. Previous studies have shown that infants can distinguish between different voices and different languages only a few days after birth, so it is possible that the perceptual mechanisms used to make these distinctions are also used to more generally distinguish speech from background noises. However, different theories predict different results as a function of age, so the experiments will test infants at different ages to bring evidence to bear on the competing theories, thereby helping to refine and advance them. The research also has broader implications for understanding of how auditory and language systems develop in the formative years of infancy. Thus the basic knowledge created by this research promises to be informative for parents and other caregivers in terms of the impact that household sounds and other auditory distractions may have on infant development.

Human language is both universal within the species and highly variable across populations. This Integrative Graduate Education and Research Traineeship (IGERT) project will train young scientists and engineers to understand language diversity by combining the tools of behavioral, computational and biological research, drawing upon an extensive collaborative network that spans nine departments in five colleges at the University of Maryland. The project aims to promote sustainable change in the science of language by building infrastructure for interdisciplinary research on diverse languages through local and international collaborations and outreach efforts, by strengthening links between basic science and clinical and engineering applications, and by building awareness of the science of language through high school and undergraduate partnerships. The training plan provides coursework, research training, and an environment geared towards preparing students for interdisciplinary research and equipping them to build collaborative networks in their future careers. Preparation for interdisciplinary research will be provided by broad coursework, integrative pro-seminars and a post-candidacy lab rotation that will pair trainees with students from other disciplines. A central component of the project is the Winter Storm, an intensive two-week workshop that will provide foundational skills training, research planning, and professional development. The project will enhance the use of computational and neuroscientific techniques in studies of atypical language and second language learning, and will partner with an NSF-supported Science of Learning Center based at Gallaudet University that focuses on visual language. IGERT is an NSF-wide program intended to meet the challenges of educating U.S. Ph.D. scientists and engineers with the interdisciplinary background, deep knowledge in a chosen discipline, and the technical, professional, and personal skills needed for the career demands of the future. The program is intended to catalyze a cultural change in graduate education by establishing innovative new models for graduate education and training in a fertile environment for collaborative research that transcends traditional disciplinary boundaries.

A great deal of research has investigated how infants learn their native language. Such research has suggested that a number of early perceptual skills and experiences pave the way for language acquisition, and might be considered as critical underpinnings for developing language skills. In particular, research has suggested that infants? ability to separate streams of speech into individual words, and their ability to track statistical relationships in the input (statistical learning) might be prerequisite skills for language-learning, and that the maternal input an infant receives may also play a substantial role. The current research aims to test this hypothesis experimentally. The investigators will test the same group of infants on both segmentation and statistical learning between 7 and 11 months of age, while appraising maternal input to them. The investigators will also evaluate these same children a year later, on a range of language and cognitive measures. In so doing, we will be able to evaluate whether differences in infants' early abilities and experiences might predict distinct patterns of more and less successful language development.

Given the substantial emphasis on statistical learning and segmentation in the literature, the insights gained from these studies will have profound implications for research in the area of language acquisition. A better understanding of the relationship between segmentation and statistical learning will inform the debate on the relative importance of each or both to language development, both theoretically and practically. This research will greatly enhance our understanding of the relationship among different infant abilities and later language development, and lead to more nuanced theories of how language acquisition builds on early perceptual skills and experiences. Finally, results could eventually lead to earlier identification of children at known-risk for language delay and in turn provide guidance to researchers and interventionists about potential next steps in designing and testing intervention programs. This project also will provide educational opportunities for a large number of undergraduate and graduate students. The impact of this research will be enhanced by broad dissemination of the results. In addition to publication in research journals and at conferences, the PI's prior research on this topic has appeared on a variety of television and radio news reports across the country and the world, and has appeared in several parenting magazines intended for broad audiences.

Language development is shaped by the language that infants experience from families and caregivers. Much of our knowledge about early language and speech development is based on infants experiencing only one language (monolinguals), usually English. However, most infants around the world, and many in the United States, regularly experience more than one language at birth (bilinguals). Research must give more attention to the role of bilingual experience in order to better explain early language development. Bilingual children and adults differ from monolinguals in several ways. For example, bilinguals tend to have smaller vocabularies than monolinguals, but they are better at a wide range of general cognitive control skills. These skills include the ability to attend to one thing while ignoring something else (e.g., pay attention to a person's voice while music is playing), and to shift attention between different objects or events. If these differences are present even in young learners, they might have important implications for how these children learn language. New research suggests that some of the effects of bilingualism may be present even in infancy; bilingually-exposed infants show some enhanced cognitive control skills. To better understand how being bilingual changes an infant's cognitive abilities, the current project examines the processing abilities of infants within their first year of life, comparing infants exposed to one versus two languages. We also compare infants exposed to three different specific language combinations. Specifically, we ask the following questions: does bilingual exposure affect how well infants can focus on words while ignoring background sounds? Does it enable infants to better learn new patterns even in the presence of distracting information? Among bilingual infants, is there an advantage if the two languages are very similar vs. very distinct from one another? To address these questions, the project also develops a new, innovative testing system that enables running the same studies in multiple locations, for improved access to diverse language populations.

Parents, early childhood educators, early intervention specialists, policy makers, and researchers all benefit by improved understanding of the potential influences of bilingual experience during infancy. The new knowledge created by the current project will directly benefit basic research on speech and language development. Undergraduate and graduate student learning and training will be enhanced through mentored research experience at four different institutions. The new testing system will be available for download from a public website, increasing access to and opportunities for participation in research and thus advancing knowledge and scientific discovery. The discoveries from the current project will provide evidence for improving decisions by parents and caregivers interacting with bilingual infants, and will suggest recommendations for policies related to language, caregiving, and education.

Most infants around the world, and many in the USA, are raised learning two languages. Recent research suggests that growing up bilingual leads to cognitive benefits. Bilinguals show advantages during tasks that rely on short-term memory (involved in the temporary holding of information) and attention (necessary for selecting and inhibiting different forms of information). However, bilinguals' language and vocabulary development differs from that of monolinguals, and as a result they are sometimes misidentified as having language disorders. Most studies have examined ways in which bilinguals are better or worse than monolinguals. However, it is possible that bilinguals simply approach tasks differently, which leads to advantages in some tasks and disadvantages in others.

The current research examines whether differences in language exposure (monolingual vs. bilingual) and age influence individuals' ability to learn and comprehend words in in difficult listening conditions (e.g., when speech is heard in the presence of background noise). Listeners of different ages often find themselves in noisy settings where they are spoken to in the context of competing sound. Hence, examining language processing under these circumstances has great relevance to everyday life. The project includes 4 experiments with infants and adults that rely on behavioral measures (e.g., eye-tracking, verbal responses). It examines monolinguals' and bilinguals' ability to focus their attention on a particular sound signal (i.e., the target speech), while ignoring a competing signal (i.e., the background noise) during two tasks that are extremely relevant for becoming a proficient language user (learning and comprehension).

Given the rapidly growing rate of bilingualism worldwide, understanding whether or not human faculties such as cognition and language develop differently depending on the number of languages in the environment is of great importance. Discoveries from this project will provide evidence for generating more accurate tools for diagnosing language delays, developing teaching approaches, and developing recommendations for policies related to language, caregiving, and education. This project will also support the scientific training of a promising scholar.

The first step in speech comprehension is to break up the speech signal and identify the individual words. To do so, one must locate where one word ends and the next begins. Unlike in written text, where there is a visual gap between pairs of words, no such reliable cue for word boundaries exists in spoken speech. Research has shown that native English listeners can utilize a set of cues to segment continuous speech, including the structure and meaning of sentences, knowledge of what constitutes a real word in English, sensitivity to whether a sound is allowed in a certain position in the word, and the emphasis given to a part of the word. However, it is not clear whether these segmentation cues are used by nonnative listeners in a similar fashion.

The use of cues may be influenced by nonnative listeners' English proficiency. Beginning or intermediate second language (L2) learners may have less established knowledge about what constitute a real word in English. Also, L2 learners' use of cues may be influenced by the characteristics of sounds in their native languages. Four groups of participants, including a monolingual English group and three nonnative groups (Mandarin, Korean, and Spanish), will participate in four experiments. Each experiment will examine how L2 learners differ from native listeners in the use of cues and how the three L2 groups differ from each other as a result of their native language experiences.

Understanding speech is a critical component in language acquisition and identifying cues that can facilitate this process will benefit both learners and teachers. There are nearly 40 million L2 learners in the U.S.; improved English proficiency will help them become more competitive and productive in the work place. This study also contributes to the training of a promising young researcher.

Language learning, in humans and machines, has far-reaching relevance to global technology, commerce, education, health, and national security. This National Science Foundation Research Traineeship (NRT) award prepares doctoral students at the University of Maryland, College Park with tools to advance language technology and language learning. The program provides trainees with an interdisciplinary understanding of learning models from cross-training in linguistics, computer science, and psychological and neural sciences, and with the tools to work with multi-scale language data. The training program contributes to the public understanding of science through a policy internship program that engages trainees with federal agencies and Washington-area professional organizations. Moreover, by contributing to the development of a free public digital linguistic tool, Langscape, it will provide a valuable resource for researchers, the public, the government, and nongovernmental agencies to discover geographical and linguistic information about languages of the world.

Flexible and efficient language learning, in humans and machines, is the research focus of this NRT program. The research hypothesis is that improvements in learning in machines and in humans will come from the ability to use training data more efficiently at multiple scales. Through interdisciplinary team approaches, trainees will explore efficient use of language data, with a focus on the informativity of data to human and machine learning. Through a suite of training activities that includes intensive summer research workshops, engagement with undergraduates and K-12 schools, and policy internships, trainees will become flexible communicators in writing and speaking and also learn to apply their research to diverse contexts.

? DESCRIPTION (provided by applicant): The input children receive is a critical factor in their language development, but much of the language they hear directed to themselves and others occurs in noisy or multi-talker environments. Yet children must rely on these degraded signals to attempt to learn their native language. The language learning systems of the brain evolved in what were presumably far quieter ambient environments than present-day settings, where noise from traffic, television, and electronic devices are ubiquitous. Recent work suggests that children are affected by background noise much more than are adults, potentially limiting the extent to which they can benefit from the language input they receive. Yet the underlying reason for these age-related differences remains unclear, as is the effect that such differences actually have on learning. Noise has a substantial public health impact, ranging from sleep disturbance and stress to decreased performance in school. Noise and other unwanted auditory signals can potentially affect speech processing in at least two ways: they can degrade the quality of the target signal (e.g., by making portions of it inaudible) and they can serve as a potential source o distraction or interference. We predict that the latter plays a larger role in young children's particular difficulty with noise, and that this is tied to children's developing attention abilitie in general. To examine this, the current proposal has three specific aims: i) to distinguish effects caused by the degradation of the signal from effects of distraction, and identify which factor plays the greater role in children's perceptual difficulties, and how this relates to other nonlinguistic attentional skills (specifically selective attention in the presence of distractors);ii) to investigate the range of stimuli for which young children show effects of distraction; and iii) to investigate the effects of noise on the learning of new words, not just recognition of already-known words. Distinguishing effects of signal degradation from effects of distraction has both theoretical and applied significance. Theoretically, understanding the nature of children's difficulties provides insight into the mechanisms involved in speech perception, as well as into the nature of speech processing development. Moreover, most theories of early language learning presume either that the child is hearing language input in ideal listening conditions or that the input has already been separated from background noise. The proposed studies will thus spur the development of more nuanced views of language acquisition, and provide critical information for evaluating and improving such models. On an applied level, understanding the limitations of children's ability to understand and learn from spoken language in the context of other speech will have vital implications for child-rearing practices, and for understanding potential causes for language delay. The results from these studies will also impact how we think about intervention and the effect of child care quality.

Humans hear the speech of others almost every day. Understanding that speech is often quite difficult, as can be seen when interacting with automated speech recognition technologies. Doing so requires the use of complex yet surprisingly effective cognitive abilities. But are the mental tools that humans use to understand speech used for speech only, or are there ones that are applied to multiple purposes? This project seeks to link language learning and perception to other tasks to determine the extent to which speech perception shares an underlying basis with other cognitive processes. This project will enrich the understanding of cognition. Furthermore, it could open up new avenues for designing technologies to better improve speech processing as well as lead to new methodologies to train people learning a second language.

To study the domain-specificity of speech perception, this project will center on two particular aspects of speech: category learning and segmentation. Accurate comprehension of spoken language demands the segmentation of continuous speech into discrete words, just as the perception of actions demands the segmentation of perceived activity into discrete events. And listeners must learn to deal with the variability in speech sounds in order to treat some sounds as belonging to the same category, just as they must group, say, disparate dog sounds as belonging to a single "barking" category. One experiment will investigate the extent to which rate information can affect the segmentation of events, while another will assess the extent to which biases that seem to be present in phonetic category learning can also be found in non-speech category learning. A third experiment will use magnetoencephalography (MEG) to probe the acquisition of certain types of speech sound categories. All told, the research will illuminate whether and which processes in language and in other domains parallel each other, which relates to the notion of modularity, the idea that the brain houses separate components that have evolved to perform individual functions in the world.