Zhong-Lin Lu - US grants

Collaborative Research: Mechanisms of Perceptual Learning

Adult humans can show large improvements in the performance of even the simplest perceptual tasks as the result of training and/or practice. These improvements have been observed in tasks in virtually every sensory modality. This project investigates the brain mechanisms of perceptual learning and the circumstances under which perceptual learning occurs. The findings will help us understand the adaptive nature of human behavior, and may generate new computational and training principles for performance optimization in particular task environments. Our research applies a powerful method for identifying and characterizing the mechanism of perceptual learning in visual tasks. The method adds systematically increasing amounts of external noise - random visual noise (similar to random TV noise) to the visual stimulus and observes the effect on a perceptual task as perceptual learning takes place under different training protocols. Performance in clear and in noisy visual task environments can be modeled quantitatively to identify three mechanisms of perceptual learning. Each mechanism has a "signature". (E.g., modification of the observer's perceptual template through training only affects performance at high levels of external visual noise, where the external noise is large enough to be the limiting factor. Experts can eliminate external noise more efficiently.) We investigate the mechanisms of perceptual learning in a wide variety of perceptual tasks including discrimination and identification or classification of both simple and complex visual patterns. Combining our methods with transfer manipulations that test for critical properties of learning will validate the measurements and provide information about the level of learning. The proposed work will improve our empirical and theoretical understanding of the nature of perceptual learning. Characterization of perceptual learning mechanisms is necessary to a full understanding of the adaptive nature of the adult human brain. Our theory and methods provide a basis for developing adaptive models of the human brain and may contribute to the development of efficient training procedures in applied settings.

With support from a National Science Major Research Instrumentation Award, Professor Irving Biederman and his colleagues at the University of Southern California will purchase a state of the art three Tesla functional Magnetic Resonance Imaging (fMRI) system for the scientific investigation of how cognitive, emotional, perceptual, memory, linguistic, and motor capacities emerge from activity of the human brain.

Joining Professor Biederman and his Co-PIs Z.-L. Lu, L. Itti, A. Raine, and M. Arbib as users of the fMRI system will be members of a variety of academic units including the Neuroscience Program, the Departments of Psychology, Computer Science, Biology, Gerontology, Biomedical Engineering, Kinesiology, Electrical Engineering, and the House Ear Institute, Currently the community of interested users includes approximately 30 faculty and over 100 graduate and post-doctoral students. This on-campus facility will not only allow these research programs to proceed but will provide the capability for the development of imaging expertise within this community. The magnet will be available to researchers from other institutions as well.

The ability to probe the activity-not just the structure-of the intact human brain has been one of the great methodological advances of neuroscience in the past decade. The instrument will provide high-resolution images of brain structures but its primary use will be to assess functioning of the brain as subjects experience various stimuli or perform various tasks while the system measures neural activity at specific brain loci in the order of a few millimeters. Among the first of the research projects that will be launched once the system is installed is one focusing on regions of the prefrontal cortex known to modulate restraint and an appreciation of the consequences of one's own actions for individuals with and without a propensity for impulsive violence. Other studies are designed to understand how an image of a scene, never perceived previously, could be comprehended in a fraction of a section. Another will assess whether brain-produced opiates in areas that mediate comprehension provide the perceptual and cognitive pleasure associated with novel but interpretable experiences. Another study is motivated by the finding that neurons in monkey cortex involved in the production of certain motor movements, such as grasping, also fire when the monkey views the grasp of another organism. This research will evaluate whether such "mirror" neurons might be the core imitative capacity fundamental to the evolution of language. Still another investigation will focus on where and how "episodic memory"-the mental diaries of our lives-are produced and stored in the brain.

Plans for the operation of the magnet, to be housed in the Dana and David Dornsife Cognitive Neuroscience Imaging Center, will include instructional courses designed to give hands-on training and research experience to undergraduate as well as graduate students.
Special outreach programs are designed to involve qualified high school students from the local community as part of an effort to provide opportunities for underrepresented minorities to be counted among the next generation of scientists advancing our knowledge of cognitive and behavioral neuroscience.

Learning to read one or more foreign languages is essential for social and economic success in this era of globalization. Decades of research have provided much insight into the cognitive and neural mechanisms of reading, especially native-language reading. Less is known, however, about how the brain associates the words we see with their pronunciations when we learn a new language. With support from the National Science Foundation, Dr. Chuansheng Chen (University of California, Irvine) and Dr. Zhong-Lin Lu (University of Southern California), will probe this issue through an innovative design with an artificial language based on written Korean. This collaboration, which also includes researchers at Beijing Normal University, will use functional magnetic resonance imaging (fMRI) to map the neural networks of two types of reading: "sounding it out" (known as "assembled phonology") or memorizing the pronunciation of the whole word without sounding out syllable by syllable (known as "addressed phonology"). First, the (most likely distinct) neural networks for the two types of reading will be identified. Second, how these neural networks differ across people speaking different native languages (e.g., alphabetical English vs. logographic Chinese) will be studied. Third, behavioral, neuroanatomical, and neurofunctional measures will be used to predict individual differences in learning to read a new language.

This project will provide solid empirical evidence for the fundamental neural bases of learning to read in a second language. Results from this study will greatly advance our understanding of the neural basis of reading, with major implications for developing more efficient second language learning programs and reading difficulty interventions. The proposed study will also provide many training opportunities for both US and Chinese graduate students and junior researchers to gain experience in international collaboration.

Cognitive science is the interdisciplinary study of mind and intelligence. Since its birth around 1960, cognitive science has produced an extensive body of principles, representations, and algorithms, as well as many successful applications. This research discipline is advancing rapidly in much of the world, with some surprising exceptions, including China. A workshop will be held in China to report and discuss cognitive science research on topics ranging from cellular circuitry to computational cognition. The workshop has several related aims: (1) promote interchanges within and between countries across multiple sub-disciplines of cognitive science that investigate cognition at different levels and using different tools, (2) disseminate quantitative methods, successes, and the interdisciplinary nature of cognitive science, and (3) establish channels for collaborative research and initiate exchange programs for training the next generation of cognitive scientists. This will be the first workshop/conference on cognitive science in China that brings researchers from a broad spectrum of distinct but connected sub-disciplines together to seek cross-disciplinary interactions and collaborations and joint training programs for graduate students.

DESCRIPTION (provided by applicant): Amblyopia is the leading cause of visual impairment in children and affects approximately 3-5% of the population worldwide. The disorder causes a syndrome of monocular and binocular deficits that persist after optical correction, in the absence of observable ocular pathology. Current treatment for amblyopia is unsuccessful in 25-50% of cases and rarely restores normal binocular vision. The broad range of deficits observed in amblyopia, its different etiologies, and the individual differences in responsiveness to treatment, suggest that multiple functional defects underlie amblyopic visual impairment. This proposal aims to isolate the core deficits that underlie the spectrum of amblyopic visual impairment, develop novel testing methods that efficiently characterize such core deficits, and evaluate how these deficits are affected by current and emerging therapies. Though the list of core deficits we propose to study may not be exhaustive, and other deficits may contribute to amblyopia, developing a framework that improves methods for monocular and binocular visual assessment will facilitate diagnosis, treatment and management of amblyopia and other visual impairments. In Aim 1, we will develop and test novel methods for rapidly assessing contrast sensitivity, binocular interaction, and spatial distortion in normal and amblyopic vision at the individual leve. In Aim 2, we will quantify the dependence of broad visual pathologies observed in amblyopia on core deficits in contrast sensitivity, binocular combination, and spatial distortion. In Aim 3, we will evaluate how conventional and emerging amblyopia treatments remediate each core deficit. The overall goal of the proposed research is therefore to develop and translate state-of-the-art psychophysical methods to clinical applications. Accomplishing the proposed aims will provide new and improved methods for (1) quantifying normal and impaired monocular and binocular vision, (2) analyzing and diagnosing visual impairment, and (3) implementing customizable rehabilitation strategies for visual impairment. The eventual development of comprehensive rehabilitative therapies for amblyopia could be achieved by targeting the core deficits. Each aim addresses a goal of the NEI's National Plan for Eye and Vision Research.

Retinotopy is the mapping of visual inputs from the retina to neurons in the brain. A retinotopic map is a visual of a particular occurrence of neuron activity taking place on a specific location in the brain. By analyzing the stimulus-referred functional magnetic resonance imaging (fMRI) response, retinotopic maps of the human visual cortex are generated. It has been hypothesized that human retinotopic maps are conformal mappings, but to date no theoretical models have been developed to quantify these maps. This project uses conformal geometry and fMRI data to study retinotopic maps in an attempt to model visual cortical organizations in the brain. This project will: (i) compute the intrinsic geometrical features that will determine and/or validate the conformality in the human retinotopy; (ii) model the relationships between the retinotopic maps in extrastriate visual areas and those in the primary visual cortex; (iii) develop methods to quantify retinotopic maps of individual subjects. The mathematical models applied in this project combine tools from topology, conformal geometry, complex analysis, optimization, and quasiconformal Teichmüller theory.

Visual processing areas have been estimated to occupy more than half of the total surface of the primate neocortex. However, an understanding of visual cortical organizations still remains elusive due to their biological complexity. Retinotopic mapping of the human visual cortex, therefore, can be an important tool for studying the brain's circuitry. This project will produce theoretically sound and practically efficient methods for quantifying retinotopic maps. These methods will lead to non-invasive biomarkers of visual functions and may lead to cures for visual deficits. The computational theories and algorithms developed in this project will have applications in other fields, including computer vision, computer graphics, sensor networks, and geometric modeling. This project contributes to the BRAIN Initiative by increasing our knowledge of visual cortical organizations in the human brain and by developing laboratory infrastructure for knowledge discovery from brain imaging data. This project will create an interdisciplinary environment for graduate and undergraduate students and will develop interdisciplinary courses at the interface of mathematics and neuroscience.

Understanding the complex determinants of individual health and wellbeing is critical for the promotion and maintenance of a healthy world population. Wellbeing may be understood not only as the absence of physical and mental illness but also as the quality of life and optimal functioning of individuals. It is well known that individuals vary tremendously in terms of cognitive abilities and dispositions, as seen from performance on high-order cognitive tasks, decision-making preferences, and emotional competencies. However, the neural underpinnings of much of this variability are poorly understood: It is unclear how individual differences in brain structure and function across tasks and processes are linked to abilities and competencies. This project explores a mathematical and computational framework for investigating a large-sample neuroimaging and behavioral dataset in order to improve our understanding of individual differences in cognitive performance. An ultimate goal of the project is to predict individual cognitive performance in novel, real-world situations based on observed (past) behavioral and neuroimaging data and contribute to the understanding of cognitive health and wellbeing of individuals. The project will also offer many training opportunities for the next generation of scientists.

The technical approach will build on and integrate recent advances in cognitive science, neuroscience, statistics, and machine learning. Statistical models will integrate data from both brain imaging and behavioral tests to generate predictions that otherwise may not be possible with a single source of data. The research will go beyond establishing and explaining individual differences to predicting individual cognitive performance in a variety of tasks.

Project Summary/Abstract: Research in perceptual learning has demonstrated a remarkable ability of training or practice to enhance perception in the adult human. The last thirty years have yielded many important findings about how people learn, what limits transfer, how generalization can be improved, how to model learning, and the nature of visual plasticity. At the same time, learning and transfer have been measured at a relatively coarse scale that leads to relatively inaccurate measures of learning in individuals, which could be very important to choosing adapted training options. Related issues of estimation have also limited the types of training protocols that have been studied. The objective of this research is to use innovative new adaptive performance assessment (based on Bayesian principles) to provide unbiased and high precision estimates of learning in individuals. We also use computational neural network models to generate predictions about more complicated training regimens that are then tested experimentally. We develop a framework for searching among these predictions computationally to identify better (optimized) training methods. The long-term goal is to develop efficient new assessments of learning and transfer and the modeling techniques that may then be applied to improve clinical applications, rehabilitation, and perceptual expertise identified as key aspects of the NEI mission.

PROJECT SUMMARY / ABSTRACT As of 2015 there were 940 million people with some degree of visual impairment in the world. Visual impairments generate considerable economic burden for the society. The World Health Organization estimates that 80% of visual impairments are either preventable or curable with treatment. Noninvasive imaging techniques have been used extensively by eye specialists for diagnosis and treatment of visual disorders and imaging is one of the priorities in the six core program areas of the National Eye Institute. As a noninvasive high spatial resolution technique for measuring brain activities, functional magnetic resonance imaging (fMRI) has provided a wealth of data on visual cortical organizations. Although numerous studies have been devoted to discovering and validating different retinotopic maps in the human visual system, limited progress has been made in developing software tools that fully consider the intrinsic geometrical features of the underlying cortical structures, enforce diffeomorphic mapping when constructing retinotopic maps and atlases, and integrate both individual and population statistics for more robust data analysis. In preliminary work, we have developed a complete and invertible description of retinotopic maps (U.S. Patent Application Nos. 16/230,284 and 63/004,721, supported by NSF collaborative research awards DMS-1413417 and DMS-1412722). This project will continue developing and applying novel quasiconformal geometry and hierarchical Bayesian modeling (HBM) algorithms to retinotopy data obtained from the Human Connectome Project (HCP), the largest high resolution retinotopy dataset to date. We hypothesize that, by combining Beltrami smoothing, quasiconformal mapping and HBM, the proposed approach will reduce manual annotation work and maximize the statistical power of retinotopic mapping techniques. The project aims to: (1) Develop computational methods to effectively smooth retinotopic maps across multiple visual areas based on Beltrami smoothing. With Beltrami descriptions, the proposed method will simultaneously smooth eccentricity and polar angle retinotopy data in V1, V2 and V3, while preserving the underlying topological continuity; (2) Develop computational methods to effectively register retinotopic maps of multiple visual areas across subjects with quasiconformal mapping. Unlike previous work that relied on either structural MRI (sMRI) or fMRI data only, the proposed method will simultaneously register both sMRI and fMRI data from multiple visual areas across subjects and ensure diffeomorphism; (3) Develop an HBM of the retinotopic maps to capture the hierarchy at both the individual and group levels. The proposed HBM will help overcome measurement noise, reveal both population properties and individual differences, and offer unprecedented accuracy on retinotopic map analysis; (4) Develop and disseminate software tools and atlases of human retinotopic maps. The developed open-source software tools can be extended to analyze data from patients with not only visual impairment but many other neurological and psychiatric disorders.