Linda B. Smith, Ph.D. - US grants

The ability to compare objects according to discrete dimensions such as size, color or shape seems basic to human intelligence. However, developmental research indicates that knowledge of dimensions is complex and emerges gradually throughout the preschool and early school years. This research investigates a componential model in which an understanding of identity, wholistic similarity, attribute, dimension, direction, intensive order, and continuity comprises the structural core of dimensional knowledge. The proposed project tests the developmental independence of these components and their order of acquisition. Performance in verbal and nonverbal tasks will be compared to provide a complete description of the development of dimensional concepts. Certain aspects of this growth may reflect experience with objects, other aspects may reflect experience with language. Children aged 2-8 will be the primary participants. The experiments compare children's conceptual knowledge of relations and their acquisitions of the words that refer to relations. Conceptual knowledge will be assessed in an imitation task that requires abstract inferences about relations. Word knowledge will be measured by verbally asking children to select objects specified by attribute or relation, e.g. "Find the big one", "Which two are the same." Additional experiments investigate the perception of relations in children and adults in order to understand the development of the mechanisms that underly the developmental growth in relational knowledge.

The ability to compare objects according to discrete dimensions such as size, color or shape seems basic to human intelligence. However, developmental research indicates that knowledge of dimensions is complex and emerges gradually throughout the preschool and early school years. This reserach investigates a componential model in which an understanding of identity, wholistic similarity, attribute, dimension, direction, intensive order, and continuity comprises the structural core of dimensional knowledge. The proposed project tests the developmental independence of these components and their order of acquisition. Performance in verbal and nonverbal tasks will be compared to provide a complete description of the development of dimensional concepts. Certain aspects of this growth may reflect experience with objects, other aspects may reflect experience with language. Children aged 2-8 will be the primary participants. In addition language-disordered children and adults will participate in some studies. The experiments of Project I compare children's conceptual knowledge of relations and their acquisitions of the words that refer to relations. Conceptual knowledge will be assessed in an immitation task that requires abstract inferences about relations. Word knowledge will be measured by verbally asking children to select objects specified by attribute or relation, e.g. "Find the big one", "Which two are the same." Project II examines children's knwoledge of relations via learning tasks, reasoning tasks and seriation tasks. Project III investigates the development of relational language and relational concepts in language-disordered children and in children learning a first language other than English. Project IV investigates the perception of relations in children and adults in order to understand the development of the mechanisms that underly the developmental growth in relational knowledge. Since aspects of dimensional knowledge are crucial components of mathematics and inferential reasoning, this project will contribute to educational programming for all children and special populations such as that of language-disordered children.

In their first 4 years of language learning, children acquire roughly 15000 words. Understanding how they acquire so many words so quickly is fundamental to understanding a critical aspect of human cognition: how natural predispositions interact with learning mechanisms to carve the world into the categories encoded by language. Answering this question is in turn a prerequisite for understanding disorders in language development and the commonly observed links between language disorders and other learning disabilities. In this research, we investigate a "shape bias" in children's early word learning. As first reported by Landau, Smith, & Jones (1988), 2- and 3 year-old children who are given a novel name for a novel object extend the name to other objects that are the same shape as the exemplar, regardless of variations in texture and size. We believe that the shape bias provides an important new window on the mechanisms of word acquisition and especially the interaction of perception, language, and knowledge in first language learning. The present research consists of 9 experiments directed to three objectives: (1) Tracking the developmental origins of the shape bias, in interactions between object knowledge, perception, and language. Five experiments examine the shape bias in children 16 months to 5 years of age. We are specifically interested in how the perceptual properties of rigid objects and language might interact to create and refine the shape bias. (2) Challenging the shape bias: two experiments examine the "moveability" of the shape bias in children from 24 to 48 months of age. These experiments follow up on previous findings that the shape bias is at first fragile, then strong and rigidly applied, then more flexible and differentiated. (3) The shape bias and category induction: two experiments ask whether newly-formed categories based on shape will support children's induction. Such inductions provide an important mechanism for the growth of category knowledge. In the core method, children are presented with novel 3-dimensional objects, made of wood, wire, or cloth. The object is named. The children then decide whether other objects can be called by the same name. This method duplicates the natural context of first word acquisitions.

Two- and 3- year old children are smart noun learners. Indeed, they are so skilled that they often need to hear only a single object named to correctly generalize that name to other members of the category. Young children's facility in mapping nouns to categories is particularly remarkable because nouns name objects in many different kinds of categories, categories of substances, people, animals, artifacts. This research tests the hypothesis that children learn how to learn words by learning words. The idea is that each word learned makes children ever smarter and ever faster word learners. The project tests four hpotheses: 1. Early learned nouns present the kinds of statistical regularities that could teach the organizational principles of different kinds. We will study the statistical properties of a corpus of about 300 nouns that are typical of the first nouns that children learn. 2. The statistical regularities characteristic of early noun categories are sufficient to create knowledge about kinds. We will teach the corpus of nouns to an associative learning device, a connectionist net. 3. The specific nouns known by an individual child determine how that child interprets and generalizes a newly encountered noun. We will measure the individual nouns known by children 20 to 36 months of age and their generalization of novel nouns in an artificial word learning task. We will present the connectionist network with the nouns that individual children know and attempt to model individual children's performances in artificial word learning tasks. 4. Children learning different languages exhibit different developmental trajectories. We will test hypotheses 1 through 3 in two languages - English and Japanese. Past research indicates that English- and Japanese-speaking children differ in their interpretation and generalization of novel names for novel objects. This research tests the idea that this is because children's novel noun generalizations reflect the statistical properties of the nouns that they have already learned. Children become smart learners of the specific language they are learning.

Two- and 3- year old children are smart noun learners. Indeed, they are so skilled that they often need to hear only a single object named to correctly generalize that name to other members of the category. Young children's facility in mapping nouns to categories is particularly remarkable because nouns name objects in many different kinds of categories, categories of substances, people, animals, artifacts. This research tests the hypothesis that children learn how to learn words by learning words. The idea is that each word learned makes children ever smarter and ever faster word learners. The project tests four hpotheses: 1. Early learned nouns present the kinds of statistical regularities that could teach the organizational principles of different kinds. We will study the statistical properties of a corpus of about 300 nouns that are typical of the first nouns that children learn. 2. The statistical regularities characteristic of early noun categories are sufficient to create knowledge about kinds. We will teach the corpus of nouns to an associative learning device, a connectionist net. 3. The specific nouns known by an individual child determine how that child interprets and generalizes a newly encountered noun. We will measure the individual nouns known by children 20 to 36 months of age and their generalization of novel nouns in an artificial word learning task. We will present the connectionist network with the nouns that individual children know and attempt to model individual children's performances in artificial word learning tasks. 4. Children learning different languages exhibit different developmental trajectories. We will test hypotheses 1 through 3 in two languages - English and Japanese. Past research indicates that English- and Japanese-speaking children differ in their interpretation and generalization of novel names for novel objects. This research tests the idea that this is because children's novel noun generalizations reflect the statistical properties of the nouns that they have already learned. Children become smart learners of the specific language they are learning.

? DESCRIPTION (provided by applicant): This is a basic science training grant focused on an integrative understanding of post- natal human behavioral development. The rationale for the focus on integrative training is that effective translation requires more than merely rapid movement of single-variable basic science findings to efficacy studies but a different basic science, one that embraces complex causal pathways of development, and considers processes at nested time scales and multiple levels of analysis. The training program focuses on behavioral development (and relations to brain development) because advancing research shows that post-natal behavior and experience modulates both structural and functional connectivity in the brain, tunes specialized neural systems and influences gene expression, with atypical patterns of early behavior and experiences determining the quality and opportunities of whole lifetimes. The trainees are 5 predoctoral candidates in psychology, cognitive science, and neuroscience and 3 postdoctoral fellows from various fields interested in developmental process. The training program for predoctoral trainees is 5 years (with 2 years supported by the training grant and 3 years by the participating PhD programs); the training program for postdoctoral trainees is typically 2 years. The training program emphasizes the use of cross-level methods to study the same problem, basic science that can link to translation, the collection and analysis of large data sets, open data and data sharing, and the ethical conduct of research.

[unreadable] DESCRIPTION (provided by applicant): Human categories are organized into different kinds, for example, living and nonliving things, animates and inanimates, objects and substances. These differences are evident in how people reason about different kinds and in their neural organization of categories. This research investigates the developmental origins of these distinctions. The studies test the hypothesis that the organization of categories into different kinds is the learned consequence of correlations among the perceptual properties of things, lexical category structure, and language. To test this hypothesis the research proposes 11 studies investigating the statistical regularities presented by the first 300 nouns that children learn. The regularities in category structure, their relation to perceptual cues, to events, and to the verbs with they co-occur will be studied in two languages -- English and Japanese. By comparing the structures presented by two languages we can tease apart the relevant contributions of language and perceptual correlations. We model the processes through which these regularities may organize categories into kinds with neural networks. From these neural network simulations, we derive predictions that are then tested in behavioral studies with children learning English and children learning Japanese as their only language. The children range in age from 18 months to 4 years old. This research contributes to our understanding of the role of language in cognitive development and in so doing has direct implications for child health in the areas of language delay, environmental retardation, and other areas in which language learning is in some way compromised. In addition, the research seeks a fine grained specification of the learning mechanisms involved which have general significance beyond the role of language in early cognitive development. [unreadable] [unreadable]

DESCRIPTION (provided by applicant): Development in one domain may have profound effects on developmental progress in other domains. This research examines how very young children's learning their first noun categories -categories such as house, chair, dog, and spoon - may play a formational role in organizing processes of visual object recognition, and how in turn these processes may feedback onto and influence the nature and speed and of early lexical learning. Further, the work examines the relation between these developmental achievements and symbolic play. Symbolic play has been used as am important marker of language delay but the mechanistic nature of the link between symbolic play and early language development has not been understood. This proposed research tests the idea that the link is through the effects of lexical learning on visual object recognition. The experiments include studies of typically developing children from 15 to five years of age and also studies of children (so-called late talkers) whose progress in early word learning is (at least initially) slower than their peers and who often (even after seeming to catch up in language) show learning deficits in school. The studies include observational studies, experimental studies directed to detailing the specific cognitive processes involved in lexical learning and object recognition, and training studies that seek to understand the causal mechanisms of change by experimentally inducing change in the laboratory.

DESCRIPTION (provided by applicant): Development is about creating something more from something less -a limb or heart or brain from a mass of identical cells, a walking talking toddler from a helpless crying infant, a child who reasons about time, number and space from one whose judgments are tightly tied to the immediate perceptual input. Advancing knowledge suggests that these remarkable consequences are the product of a self-organizing system emergent in nested processes and mutually interacting mechanisms over many time scales and levels of analysis. If we are to understand developmental mechanisms, then, scientists must be trained to integrate processes of change at different levels of analysis - from the brain to behavior - and at different times scales --from the milleseconds of neuronal activity to the seconds and minutes of tasks to the days, months and years of learning and growth. Accordingly, funds are requested to support 5 pre-doctoral and 4 post-doctoral trainees in a strongly collaborative and interdisciplinary community with a shared vision of an integrative approach to both training and research. The Training faculty forms a close-knit community from 5 units at Indiana University (Psychology, Speech and Hearing Sciences, Optometry, Computer Science, Kinesiology). Training is accomplished through course work, research in the laboratories of the faculty, and specialized training seminars, workshops, and colloquia. All trainees are required to work in at least two laboratories and must attend the on-going integrative DTG seminar. We require active participation in two laboratories so that trainees may go beyond their primary mentor in synthesizing methods and advances at different levels of analyses and times scales, and in so doing be prepared to be at the forefront of where developmental science is going. In brief, this is a forward-looking training program built on a strong and continuing base of theoretically driven, highly interdisciplinary, and collaborative research among faculty and trainees, and one that has, to date, been highly successful in producing exciting young scientists ready to make their mark.

The goal of this research program is to understand the acquisition of motor skills in infancy and early childhood. Specifically, this research is concerned with the dual-nature of skill: how children learn to become both stable in their performance and flexible in their abilities to engage in new tasks. The studies are informed by a dynamic systems perspective and address skill development in terms of dynamics tested and coupled over levels and time scales. This theoretical approach focuses on the multiple, interacting factors that lead to movement. Here we especially consider how perception, motor memory, and task intersect in producing movements of eyes and head, and manual actions of varying complexities. Studies look at changes in real and developmental time and include empirical research, model simulation and robotics. Specific Aim 1 is based on a new model of infant visual habituation. In six experiments, we manipulate the metric parameters contributing to infants' development of visual attention. In Specific Aim 2 we study the dynamics of action memories for simple reaching with hands and with feed in infants and toddlers, and how those memories interact with visual attention. Specific Aim 3 is to study the dynamics of complex motor actions. Infants and toddlers are tested with toys of varying complexity to understand how recent and longer-term experience affect their tendencies to repeat or switch to new actions. In Specific Aim 4 we implement these theoretical models in an autonomous robot to further explore mechanisms of real time behavior and its developmental course. Specific Aim 5 is to continue experiments on infant limb perturbations currently funded and Specific Aim 6 to write a book on development of embodied cognition based on these and previous experimental and modeling work. These studies impact upon our understanding of basic mechanisms of skill development and continue to contribute new theories of clinical intervention in pediatric physical and occupational therapy.

[unreadable] DESCRIPTION (provided by applicant): Understanding the regularities in the learning environment, how they differ among individuals, and how individuals through their own behaviors create those differences are essential to understanding typical and atypical courses of development, to developing remediation and intervention procedures. The proposed research develops a new method to study the dynamic visual environment of infants and toddlers, from the first-person view. Progress has been limited by several technical problems including: (1) recording devices that will be tolerated by young children and (2) the number of independent degrees of freedom in eye, head, and body movements. We believe we can solve these technical problems. The research has three specific aims within the goal of studying active vision in children 10 to 36 months of age: (1) to develop a system through which one can measure the dynamic first-person visual field (via a small camera on the head of the child), the direction of eye gaze in that field (via additional cameras and software that track eye position), and also head, hand and body movements (via position sensors placed on the head, shoulders and hands); (2) to determine the spatial and temporal resolution of the head-camera image relative to the child's perceptual field; (3) to demonstrate the utility and functionality of the system in two different task contexts - mother-infant toy play at a table and free play with ambulatory movements. This new method promises fundamental new insights about the learning environment and the development of dynamic visual processes in the context of a 3-dimensional physical and social world, the context in which children build social, language and cognitive skills. The proposed new method has significant biomedical relevance for the early detection of attentional disorders, for studying their cascading consequences in social, cognitive and linguistic development, and for developing new therapies and interventions that generalize to the complex natural environment [unreadable] [unreadable] [unreadable]

This Integrative Graduate Education and Research Traineeship (IGERT) award supports a training program on the dynamics of brain-body-environment interaction in behavior and cognition at Indiana University. The purpose of the training program is to create a new kind of scientist with expertise in both the experimental and theoretical tools necessary to analyze intelligence as an emergent property of a complex dynamic system. The training program includes new courses, a professional development seminar, a colloquium series that provides opportunities for extended interactions between students and top researchers, research internships, and opportunities for international collaboration. The program also includes a detailed assessment plan and a summer program for undergraduates from underrepresented groups run in partnership with Indiana University Northwest, approximately 80% of whose students come from minority, first-generation college, female or low socioeconomic categories. Broader impacts of this program include recruiting new students from underrepresented groups into cognitive science and providing graduate students with the opportunity to participate in cutting-edge multidisciplinary research. All materials produced by this program will be made freely available on the web. More generally, this program will foster new kinds of discourse between the various disciplines that make up cognitive science. Finally, by placing cognition within its proper embodied and situated context, the proposed training program may impact how society fosters and measures cognitive ability. 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.

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).

Children begin to comprehend words at 9 months. They say their first word at around 12 months. The pace of vocabulary learning then accelerates so that by 24 to 30 months, children add words at the staggering rate of 5 to 9 new words per day. There have been many studies focusing on documenting developmental progress in early language acquisition, and most theories of learning derived from those studies have focused on macro level descriptions that sound like explanations, such as "the mother tried to elicit the child's attention by waving the toy." These descriptions may capture higher-level human behaviors, but they fall short of a mechanistic account of how word learning works in real time. Toddlers learn words through millisecond by millisecond, second by second, and minute by minute events that are generated by actively engaging in the world, with objects, and with their social partners. But very little is known about how any of this works in real time and in the cluttered context of the real world interactions of toddlers and parents, contexts typically characterized by many interesting objects, shifts in attention by each participant, and goals (beyond teaching and learning words). In light of this, the series of experiments in this project will provide a systematic study of child-parent interaction and learning as coupled complex systems. The child's actions (head and eye movements, hand movements, picking up objects) create within the child dynamic dependencies of looking, seeing, touching and feeling. Each moment of perceptual and motor activity by the learner determines the next -- a head turn determines what is seen next, which may determine what is reached for and brought close to the eyes, which selects and generates the next view. Thus, the learner is a dynamic complex system. But the toddler is not alone when learning new words. Instead, a mature partner -- who is also a complex multimodal system -- offers words, gestures and actions. Critically, the streams of touches, sights and sounds from two participants are closely coupled, with one agent shaping the experiences and behaviors of the other. The study will measure the dynamic multimodal behavioral patterns within and across social partners as children and parents actively engage with and talk about objects in everyday contexts. The project will collect multiple streams of high-resolution, high-quality video and speech data from both participants. The dense and rich streams of multimodal data are useful only to the degree that one can find meaningful patterns in those dynamic streams that bring new insights into real-time learning events. To this end, the project will develop new methods of data analysis, visualization and data mining to quantify fine-grained behavioral patterns within an individual's cognitive, perceptual and motor systems and across social partners. This constitutes a significant advance in theoretical approaches to early word learning and one that also has broad applications. Measuring interaction patterns within and between complex systems is a critical problem across science -- from cells, to brains, to coupled physical systems, to human-computer interaction, to groups of animals, to teams of people. Thus, this research will bring new methods and analytic tools for measuring the information in coupled interactive systems.

Understanding learning mechanisms in the context of a dynamic, everyday learning environment is essential to understanding typical development, individual differences, and atypical development. Designing effective procedures to benefit children with developmental delays requires a principled understanding of that dynamic environment as it relates to the cognitive learning system. Thus, the work will provide scientists, educators, and parents with an understanding of children's early cognitive processes and general principles to facilitate child-parent social interaction and early language learning. Moreover, building anthropomorphic machines that can acquire language automatically may be best accomplished by emulating how toddlers learn language. Artificial intelligence systems with human-like language skills have important utilities in real-world applications. Finally, this approach is methodologically novel. Not only will it provide new findings, but the research will be a proving ground for the development and invention of these new techniques -- techniques that may be applied in many different domains of social and behavioral studies, such as typical and atypical cognitive development, collaboration and joint problem solving, and adult social interactions.

DESCRIPTION (provided by applicant): The visual world is cluttered with many targets and many distractions. Learning requires selecting and stabilizing attention on just some of that information. The proposed research examines the idea that for toddlers, effective visual attention is fundamentally a sensory-motor process involving external bodily actions. The proposed research addresses a gap in current understanding of how sensory- motor behavior supports effective visual attention, a critical gap given the evidence implicating atypical attention in many developmental disorders and the well established co-morbidity of atypical sensory-motor patterns in these same developmental disorders The proposed experiments measure multiple sensory and motor streams: head and hand movements, eye-gaze direction, and the first- person head-centered view as 12 to 24 month old children act on and learn about objects. The studies will provide a fine-grained description of the dynamics of visual attention, of how children's own actions help to sustain and stabilize visual attention, and the role of these processes in learning about objects.

DESCRIPTION (provided by applicant): Language and visual object recognition are two domains of human intelligence that have wide impact on all cognitive systems. Previous research suggests that attention to object shape is a strong predictor of early noun learning and that young children who are delayed in early language learning show attention to shape in noun learning tasks and also deficits in the visual recognition of common objects. These previous findings suggest developmental connections between early lexical development and visual object recognition. The overall objective of this application is to characterize this relation in 18 to 30 month old whose early language is progressing at normative rates and also in children whose early noun learning is progressing more slowly and who are risk for later language learning and language processing deficits. The rationale for doing this research is that it will lead to a deeper understanding of the interdependencies between early language learning, language delay, and developmental changes in visual object recognition. The studies will provide new developmental benchmark measure of visual object recognition and a fine-grained description of the development of visual object recognition in relation to early noun learning in typically developing and language delayed children. PUBLIC HEALTH RELEVANCE: Relevance Children who are slow to learn and use language are at risk for significant difficulties in language learning, in language processing, and in later school learning. The proposed research focuses on the earliest noticeable language delays in 18- to 30-month old children and the inter-relations between early noun learning and developmental changes in visual object recognition and methods for enhancing language learning. The work is relevant to public health in that it may aid in early diagnosis and early remediation of these difficulties.

Human visual object recognition is fast and robust. People can recognize a large number of visual objects in complex scenes, from varied views, and in less than optimal circumstances. This ability underlies many advanced human skills, including tool use, reading, and navigation. Artificial intelligence devices do not yet approach the level of skill of everyday human object recognition. This project will address one gap in current knowledge, an understanding of the visual experiences that allow skilled object recognition to develop, by capturing and analyzing the visual experiences of 1- to 2-year-old toddlers. This is a key period for understanding human visual object recognition because it is the time when toddlers learn a large number of object categories, when they learn the names for those objects, and when they instrumentally act on and use objects as tools. Two-year-old children, unlike computer vision systems, rapidly learn to recognize many visual objects. This project seeks to understand how the training experiences (everyday object viewing) of toddlers may be optimal for building robust visual object recognition. The project aims to (1) understand the visual and statistical regularities in 1- to 2-year-old children's experiences of common objects (e.g., cups, chairs, trucks, dogs) and (2) determine whether a training regimen like that experienced by human toddlers supports visual object recognition by state-of-the art machine vision.

Considerable progress in understanding adult vision has been made by studying the visual statistics of "natural scenes." However, there is concern about possible artifacts in these scenes because they typically photographs taken by adults and thus potentially biased by the already developed mature visual system that holds the camera and frames the pictures. Also, photographed scenes differ systematically from the scenes sampled by people as they move about and act in the world. Accordingly, there is increased interest in egocentric views collected from body-worn cameras, the method used in the present work. Toddlers will wear lightweight head cameras as they go about their daily activities, allowing the investigators to capture the objects the toddlers see and the perspectives and contexts in which they see them. The research will analyze the frequency, views, visual properties, and range of seen objects for the first 100 object names normatively learned by young children, providing a first description of the early learning environment for human visual object recognition. These toddler-perspective scenes will be used as inputs to machine learning models to better understand how the visual information in the scenes supports and constrains the development of visual object recognition. Machine-learning experiments will determine which properties and statistical regularities are most critical for learning to recognize common object categories in multiple scene contexts. Data collected will be shared through Databrary, an open data library for developmental science.

The Discovery Research K-12 program (DRK-12) seeks to significantly enhance the learning and teaching of science, technology, engineering and mathematics (STEM) by preK-12 students and teachers through research and development of innovative resources, models and tools. Projects in the DRK-12 program build on fundamental research in STEM education and prior research and development efforts that provide theoretical and empirical justification for proposed projects.

Many children have trouble understanding what multidigit numbers mean and this can lead to long-term problems in mathematics. Poor place value understanding plagues children from all socioeconomic backgrounds and is not limited to those with learning disabilities. This problem is widespread and pervasive; indeed, perhaps more widespread than currently understood because some tests may overestimate what children know. The critical skill needed for long-term success, and one not always measured, is called decomposition. It involves knowing how to break a multi-digit number down into its components by place (ones, tens, hundreds) and interpret its meaning (e.g., 642 = 6 hundreds, 4 tens, and 2 ones). This project will target that crucial skill and test new ways of teaching place value based on principles of analogical reasoning, gleaned from decades of cognitive science research. These new approaches will make place value more transparent by highlighting and aligning its structure across spoken number names, written numerals, and sets of objects. By leveraging these powerful analogical learning mechanisms, it may be possible to teach place value earlier than is typical, so the project will target K-1 students. If successful, this approach could head off the misconceptions that are currently common among older children. The project will also track children over time, using tests that measure decomposition and other place value concepts to see how they interrelate. The project's activities have been designed to be inexpensive and fit into everyday educational practice, so that the results may be easily implemented by teachers.

The project will test new instructional approaches designed to help K-1 students comprehend place value. The project is innovative in that it will emphasize the underlying relational structure of place value symbols, and target this structure with instructional materials and techniques drawn from the structure mapping literature. Its theory of action is that instruction which scaffolds structure mapping for place value will better prepare children to face the challenges of advanced operations, such as multi-digit calculation. The project specifically targets decomposition, as this has been identified as a particular stumbling block for children. Also, decomposition skill has been linked to better long-term mathematics outcomes. There will be three studies. Study 1 will track the development of place value understanding from kindergarten to 2nd grade, using some measures that clearly require decomposition, and others that are in wide use but may not require decomposition, such as number line estimation and magnitude judgments. The study will show how these measures are related over developmental time, as well as testing whether a firm understanding of decomposition predicts later mathematics learning. Studies 2 and 3 will use a pretest-training-posttest design to test the efficacy of six structure mapping activities for place value in K-1 students. The training in Study 2 will be focused on one activity for three weeks, whereas the training in Study 3 will include all six activities presented over an entire semester.

This project uses methods from artificial intelligence (AI) to better understand how people learn visuospatial reasoning skills like mental rotation, which are a critical ingredient in the development of strong math and science abilities. In particular, this project proposes a new approach to quantify the learning value contained in different visual experiences, using wearable cameras combined with a new AI system that learns visuospatial reasoning skills from video examples. Results from this project will not only advance the state of the art in AI but also will enable researchers to measure how valuable different real-world visual experiences are in helping people to learn visuospatial reasoning skills. For example, certain types of object play activities might be particularly valuable for helping a child to learn certain visuospatial reasoning skills. Ultimately, this new measurement approach could be used to identify early signs of visuospatial reasoning difficulties in children and could also help in the design of new visuospatial training interventions to boost children?s early math and science development.

The core scientific question that this project aims to answer is: How are visuospatial reasoning skills learned from first-person visual experiences? This question will be answered through computational experiments with a new AI system---the Mental Imagery Engine (MIME)---that learns visuospatial reasoning skills, like mental rotation, from video examples. Training data will include first-person, wearable-camera videos from two different settings that are both important for human learning: unstructured object manipulation by infants and visuospatial training interventions designed for children. Results from experiments with the MIME AI system will advance the state of the art in both AI and the science of human learning by helping to explain how visuospatial reasoning skills can be learned from visual experiences, and, in particular, how having different kinds of visual experiences can affect the quality of a person?s learning outcomes in different ways.

Project Summary/Abstract: In the first year after birth, human sensitivity develops markedly to the fundamental features on which all higher- level vision depends: contrast, spatial scale, edge orientations, chromatic content. It is well known that this development is highly dependent on visual experience because disruptions in experience have significant and, in some cases, permanent consequences for vision from sensation to cognition. The field does not have, however, an empirical characterization of the low-level feature statistics of typical infant visual experience. This gap is critical because emerging studies of higher-level content indicate these statistical properties change with development and are dependent on the infant?s own changing internal visual biases and behaviors (eye movements, head movements, other body movements). These factors play a direct role in selecting and organizing the spatial structure of images projected to the eye. This project will collect and analyze the first- person visual experiences of 200 infants (50 each) at 2-3, 5-6, 8-9, and 11-12 months of age, plus a sample of 20 infants tested at all of those ages. The core hypothesis is that the statistics change systematically in a developmentally consistent sequence in the everyday lives of infants. The experiences are collected by infant head cameras worn for hours in the home and precision measures of eye and head movements in the laboratory. Analyses will quantify the spatial organization of fundamental low-level features in the collected images as a function of age, posture, activity, and specific contents. The project will also characterize the influence of refractive error and front-end visual immaturities on the images. The research will determine how infants? behaviors influence the spatial organization of visual features in the input by analyzing the motion patterns in the at-home head-camera images and through direct measures of eye and head motion patterns in the laboratory. The research will provide the first characterization of the natural visual statistics of infant experience in the first year after birth and is expected to reveal specific developmental risk-points in those expected visual statistics.