Seattle Brain Salon Abstracts

April 19, 2012 
Humans’ capacity for language: Neuro-learning
Patricia Kuhl, Ph.D., The Bezos Family Foundation Endowed Chair in Early Childhood Learning, Co-Director, Institute for Learning & Brain Science, University of Washington
Humans’ capacity for language has puzzled scholars for centuries, from the earliest philosophers of mind, to biologists, neuroscientists, and more recently engineers and computer scientists who want machines to crack the speech code. The study of infants’ brains is radically changing traditional views on how humans master language. I will describe a model of the earliest phases of language acquisition that addresses the nature-nurture debate by delineating the contributions of both to the acquisition of human speech. The model’s unique solution lies in the tenet that infants’ computational skills are fundamentally coupled to, in fact ‘gated’ by, the social brain. This work is leading to biomarkers that may allow early diagnosis of developmental disabilities such as autism and dyslexia. The model also suggests new technologies that will improve second language learning beyond the ‘critical period’ for language. Understanding language may provide a key to understanding the human mind.
Origins of the human mind in the infant brain
Ghislaine Dehaene, Ph.D., Research Scientist, INSERM-CEA Cognitive Neuroimaging Unit, NeuroSpin Center, Paris
Recognizing other fellow humans, understanding what they are saying, telling stories, walking, running, singing, figuring out space and number, recognizing 3D objects, etc., all essential competencies are the consequences of complex learning which begins from the first days of life on. Thanks to the development of noninvasive brain-imaging techniques, we can now open the black box of the child’s brain and mind. These new tools reveal an early and complex organization surprisingly close to the adults’ organization. This cerebral architecture has been selected through evolution as the most efficient to help human infants to pick up the correct cues in the environment. These breakthroughs into the early learning brain are crucial to understand what are the decisive properties of the human brain allowing the development of high cognitive functions such as language, mathematics, music, etc., but also to take in charge pathologies specific to childhood such as developmental deficits (e.g. dyslexia, dyscalculia, etc.), neurological and psychiatric diseases (e.g. cerebral palsy, epilepsy, mental retardation, autism, etc.). Our hope is that a better understanding of the structural and functional properties of the developmental brain would help to better define the neuronal algorithms sustaining human thoughts. I will illustrate these ideas through examples drawn from language learning.
Wiring the infant to learn from culture
Andrew Meltzoff, Ph.D., Job and Gertrud Tamaki Endowed Chair, Co-Director, Institute for Learning & Brain Science, University of Washington
Humans have a long period of infantile immaturity compared to other animals. This immaturity has co-evolved with two powerful social learning mechanisms—imitation and gaze following of others—which are operative in human infants, rare in the animal kingdom, and impaired in autism. Newborn humans imitate facial gestures they have never seen themselves make, suggesting a primitive self-concept allowing them to cross-modally map gestures they see another perform and their own unseen acts. The newborns’ coding of others as “like me” provides a felt connection to people that undergirds emotional learning, the rapid acquisition of social roles, and a sense of identification with one’s own tribe or in-group. I will explore the psychological origins of the “Like-Me” mechanism, from its first manifestations in childhood.
Children as scientists: Why children are so smart and why teaching (sometimes) can make them so dumb
Alison Gopnik, Ph.D. Professor of Psychology, UC Berkeley
I'll review recent work exploring the computations that young children's brains unconsciously perform and that allow them to learn so much so quickly. I'll describe recent work showing that even the youngest children use some of the same learning techniques as both scientists themselves and mating learning algorithms. In particular children use statistics, experimentation, and Bayesian inference to infer causal models of the world. I'll also describe work showing how learning from a teacher can facilitate, or impair, children's natural learning abilities.
Understanding others
Giacomo Rizzolatti, Ph.D. Professor of Human Physiology, Università degli Studi di Parma
Humans’ life is characterized by complex social interchanges. A striking feature of these interchanges is that others’ behavior is usually interpreted as a mark of something as insubstantial as mental activity. Actions of others are clues to their intentions just waiting to be captured. Philosophers typically claim that the capability to understand others is based on the observers’ capability to infer others’ internal mental states and to ascribe them a causal role in generating the observed behavior. In our daily life, however, we often make sense of our conspecifics behavior without resorting to inferential processes, relying instead on a direct understanding of what they do. Neurobiological evidence for a direct understanding of others has been provided by the discovery of specific classes of neurons (mirror neurons) discharging both when people perform a given motor act and when they observe someone else performing a similar motor act. There is a fundamental difference between inferential and direct understanding of others. In fact, only in the latter case the understanding is “empathic”: the other people are “like me.” Social (and medical) implications of the dichotomy between inferential and empathic understanding will be discussed.
Using lesions and single-unit recordings to understand human cognition
Ralph Adolphs, Ph.D. Bren Professor of Psychology and Neuroscience, Director, Caltech Imaging Center, California Institute of Technology
The vast majority of cognitive neuroscience studies in humans use a single source of neurobiological data, fMRI, with well-known limitations. Yet opportunities exist for obtaining insight into human cognition through invasive recordings directly made from the brain, or through causal inferences from studies of lesion patients. What is lacking thus far is an infrastructure for scaling up such relatively rare sources of data to larger samples, for making them more readily available to researchers, and for applying them to state-of-the-art experimental protocols. I will provide an overview of efforts my own laboratory has made in this direction. We have been conducting studies in up to 500 lesion patients, permitting strong statistical analyses of how damage to specific structures results in impairments in specific functions. And we have been conducting single-unit recordings in neurosurgical patients, using many of the same tasks. Together, the findings have provided unique insights into human social cognition, and point to ways of tackling major unsolved questions in the future.
Modeling and decoding the human brain
Jack Gallant, Ph.D. Professor of Psychology, UC Berkeley
How is information represented in the human brain? The best current technology for exploring these issues is functional MRI, but typical fMRI studies provide only coarse localization information. My lab has pursued a system identification approach for collecting and modeling fMRI data that reveals the representational principles of the human brain in much greater detail than was possible previously. In typical experiments we first record brain activity evoked by complex stimuli such as natural movies. We then estimate quantitative encoding models that describe the link between the stimuli and evoked brain activity. Projecting the encoding models onto flattened brain maps reveals how structural and semantic information are represented across the cortical surface. The patterns revealed by this approach are generally consistent with the coarse parcellations provided by previous techniques, but they provide much more detailed information about representation and they extend well beyond areas identified using standard methods. Furthermore, examination of the encoding models reveals what specific information is represented within each visual area, and these representations are transformed across the brain. Finally, one additional benefit of our approach is that estimated encoding models can be easily converted into decoding models. These decoding models recover both the structural and semantic information in natural movies, even from slow hemodynamic signals. This approach has many potential applications for modeling, understanding and decoding other brain systems, such as those involved in language, navigation, decision making, planning and so on.
NeuroFuture: How cognitive neuroscience may enhance human flourishing
John Gabrieli, Ph.D. Grover Hermann Professor in Health Sciences and Technology and Cognitive Neuroscience, Director, Athinoula A. Martinos Imaging Center at the McGovern Institute, Massachusetts Institute of Technology
Over the last 20 years, our knowledge about the functional organization of the human brain has grown remarkably. The combination of experimental analyses of the consequences of focal brain injures (lesions) and non-invasive neuroimaging methods, such as functional magnetic resonance imaging (fMRI), has produced many insights about how the human brain makes possible the human mind. Further, whereas lesion studies have provided compelling evidence about universal aspects of brain design, neuroimaging studies have allowed for the discovery of how functional brain designs vary across humans (neurodiversity and neuroindividuality). Neuroimaging, however, has had little impact to date on humanitarian or practical goals of cognitive neuroscience, such as helping people with developmental or neuropsychiatric disorders, or enhancing education and job performance. I will describe recent research efforts indicating that neuroimaging may offer significant evidence, not otherwise available, for enhancing practices that promote mental health and educational outcomes by predicting individuals’ future behaviors and future responses to interventions.
The Allen Human Brain Atlas
Allan Jones, Ph.D. Chief Executive Officer, Allen Institute for Brain Science
The ALLEN Human Brain Atlasis a publicly available online resource of gene expression information in human brain. Comprising multiple datasets from various projects characterizing gene expression in human tissue, one key component is an ‘all genes, all structures’ survey that generated genome-wide microarray-based gene expression profiles in human brain with accompanying anatomic and histology data. The data, freely available to any researcher worldwide via easy to use web tools, provides an unprecedented look at the functional organization of the brain from a molecular viewpoint, showing striking similarities across brains from different individuals and providing the first links of brain function to molecular function across the whole brain.
Towards a dynamical neuroscience
Michael Gazzaniga, Ph.D. Professor of Psychology, Director, SAGE Center for the Study of Mind, UC Santa Barbara
I will review split-brain studies that have lead me to change my long term view on how to best understand mind/brain interactions. Overall, the view is consistent with the idea that complex neural systems, like all complex information processing system, are highly modular.  At the same time, how the modules come to interact and produce unitary goals is the great unknown. In this process the importance of self-cuing cannot be overestimated. It is demonstrably evident in the human neurologic patient and especially in patients with hemispheric disconnection. When viewed in the context of modularity, it may provide insights into how a highly parallel and distributed brain coordinates its activities to produce a unitary output. Gaining a full understanding of cueing mechanism, will require shifting gears away from standard linear models and adopting a more dynamical system view of brain function.
Signatures of conscious processing in the human brain
Stanislas Dehaene, Ph.D. Professor of Experimental Cognitive Psychology, College de France, Director, INSERM-CEA Cognitive Neuroimaging Unit, NeuroSpin Center, Paris
Understanding how brain activity leads to a conscious experience remains a major experimental challenge. I will describe a series of experiments that probe the signatures of conscious processing. In these experiments, my colleagues and I ask whether a specific type of brain activity can be detected when a person suddenly becomes aware of a piece of information. We create minimal contrasts whereby the very same visual stimulus is sometimes undetected, and sometimes consciously seen. We then use time-resolved methods of electro- and magnetoencephalography to follow the time course of brain activity. The results show that conscious access is easily detectable as a global burst of late synchronized activity (a cortical “ignition”), distributed through many cortical areas. We propose a theory of a global neuronal workspace, according to which what we experience as consciousness is the global availability of information in a large-scale network of pyramidal neurons with long-distance axons. This knowledge is now being applied to the monitoring of conscious states in non-communicating patients. Using real-time signal processing techniques, we believe that a few minutes of testing with simple sounds and two recording electrodes might suffice to determine whether a person is conscious.
Project MindScope: Teasing Apart the Circuits of the Cerebral Cortex
Christof Koch, Ph.D. Chief Scientific Officer, Allen Institute for Brain Science
The Allen Institute for Brain Science is initiating a ten year project to study the principles by which information is encoded, transformed and represented in the mammalian cerebral cortex and related structures. The Institute will build a series of brain observatories to identify, record and intervene in the neuronal networks underlying visually guided behaviors in the mouse, including visual perception, decision making and consciousness. This is a large-scale, in-house team effort to synthesize anatomical, physiological and theoretical knowledge. It has the potential to revolutionize our understanding of the mammalian brain, the most complex piece of organized matter in the known universe. The fruits of this cerebroscope will be freely available to the public.
Share this