Conference Speakers
Center for Molecular and Behavioral Neuroscience, Rutgers
University
Oscillations Organize Cell Assemblies
Abstract:† How does the
brain orchestrate perceptions, thoughts and actions from the spiking activity
of its neurons?† Previous single neuron recording
research has regarded spike pattern variability as noise that should be
averaged out to reveal the brain's representation of invariant input.† The alternatively view is that variability
of spikes is centrally coordinated and that this brain-generated ensemble
pattern in cortical structures is a potential source of cognition.† Large-scale recordings from neuronal
ensembles now offer opportunities for challenging and testing these competing
theoretical frames.† A postulated
signature of the cell assembly is that its participants show a higher
probability of spiking together than with members of other assemblies, even in
the absence of external inputs.†
Interactions among parallel-recorded hippocampal neurons revealed a
consistent temporal structure beyond that predicted from the environmental
inputs.† We find that prediction of
spike times of hippocampal pyramidal neurons is improved using the spike times
of simultaneously recorded neurons, over prediction from the animal's
trajectory in space, or a spatially-dependent theta phase modulation.† Thus, we suggest that the assembly
organization arises from the internal dynamics of neuronal circuits, and
reflects the operation of non-sensory cognitive phenomena.† Importantly, the time window within which spike
times were best predicted from simultaneous peer activity is 10-30ms,
suggesting that cell assemblies are synchronized at this timescale.† Because this temporal window matches the
period of the gamma oscillation and the time window for synaptic plasticity, we
suggest that cooperative activity at this timescale is optimal for information
transmission and storage in cortical circuits.†
Altering synaptic weights within the hippocampal network by LTP,
assembly membership could be altered.†
We suggest that assembly-based approach provides an insight into
centrally-organized (cognitive) events without reference to introspection.
Harris, K. D., Csicsvari, J., Hirase, H., Dragoi, G.. and Buzs·ki,
G. Organization of cell assemblies in the hippocampus. Nature 424:552-556,
2003.
Dragoi G, Harris KD, Buzsaki G.†
Place representation within hippocampal networks is modified by
long-term potentiation. Neuron 39:843-853, 2003.
Buzsaki, G. and Draguhn, A. Neuronal oscillations in cortical
networks. Science 304: 1926-1929, 2004.
Bio:† Gyˆrgy Buzs·ki earned an
M.D from the University of Pecs, Hungary (1974), and a Ph.D. from the Academy
of Sciences in Budapest (1984).† He was
a postdoc at the University of Texas, San Antonio and at the University of
Western Ontario, London, Canada.† He was
an Assistant Professor at the University of Pecs, a Visiting Associate
Professor at the University of Lund, Sweden, and a J.D. French Foundation
Fellow and Associate Professor in-Residence at University of California, San
Diego. Since 1990, he has been a Professor at the Center for Molecular and
Behavioral Neuroscience at Rutgers University.†
Dr. Buzs·ki is on the Editorial Boards of Neuron; Neuroscience (Section
Editor, 2000-2004); Journal of Neuroscience; Neurophysiology; Behavioral Brain
Research; Epilepsia (1999-2004); Hippocampus; Thalamus; Neuroscience
Communications; Behavioral Neuroscience (1992-1996); and Restorative Neurology
and Neuroscience.† Dr.† Buzs·ki is an ISI Highly Cited researcher, a
Fellow of† the Neurosciences† Research Program, an Elected† Fellow of†
the† AAAS, a Foreign member† of the†
Hungarian Academy Sciences, and a†
Fogarty International Senior Fellow.†
He received the Krieg Cortical Discoverer Award, the first Pierre Gloor
Award, and an Excellence in Research Award from Rutgers University.
Web Address: †http://osiris.rutgers.edu/fronthigh/indexhigh.html
*************************
Google
Petabyte
Processing Made Easy
Abstract:† Having to process petabytes of data sounds hard, but
it's easier than you
may think, and more affordable too.† In
this talk I'll describe how
we deal with large amounts of data at Google.†
It starts with compute
clusters that are optimized for throughput rather than peak performance, and a software layer that turns these
clusters of relatively
unreliable machines into a reliable computing platform.† To simplify storage management, the Google File System
(GFS) organizes a sea
of local IDE disks into a convenient and reliable file system optimized for very large files.† To simplify programming, the
MapReduce framework allows programmers
to focus on just the transformations
they want to accomplish, freeing them from worrying about parallelization, scalability, and machine
failures.† And finally, to allow for effective sharing of a large
clusters, a work queue
management system schedules the combined workloads onto the
available servers (yes, batch
computing is back!).† I hope that this talk will encourage you to demand
similar facilities for
your own work--$1000 buys you roughly two Terabytes of raw disk
space these days, so why shouldn't
*you* have a few? † Similarly,
PC-based servers provide substantial
performance for little money.†
All that's missing is a few
pieces of software to turn this cheap hardware into a reliable tool for researchers who
need to process large
datasets. If you encourage the cluster software community to
provide this software, it might just happen.
Bio:† Urs joined Google as its first VP Engineering after having been
an Associate Professor of CS at UCSB and currently serves as the VP of Operations
and as a Google Fellow.† In his previous
life he contributed to
Sunís Hotspot JVM which is based in part on compiler research he
performed at Stanford and Sun Labs as
part of the Self project. With a good search engine you can find out much more about
him.
Web Address:† http://www.cs.ucsb.edu/~urs/
*************************
Laboratory of
Sensory Neuroscience, Rockefeller University
Making an
Effort to Listen: Mechanical Amplification by Myosin Molecules and Ion Channels
in Hair Cells of the Inner Ear
Abstract:† Uniquely among vertebrate sensory receptors,
the hair cell of the internal ear amplifies its inputs.† An active process in auditory organs
increases responsiveness to sound by over one-hundredfold and sharpens
frequency selectivity.† Two epiphenomena
arise from the active process, nonlinearity in transduction and spontaneous
otoacoustic emissions (SOAEs).† Although
changes in the length of outer hair cells are thought to mediate amplification
in the mammalian cochlea, the auditory receptor organs of non-mammalian
tetrapods, which lack electromotile hair cells, display essentially identical
sensitivity, tuning, nonlinearity, and SOAEs.†
The active process necessary to explain the properties of hearing in
these animals may therefore constitute part or all of the mammalian cochlear
amplifier as well.† When bathed in a low
Ca2+ saline solution resembling endolymph, a hair bundle from the sacculus of
the bullfrog's internal ear undergoes spontaneous oscillations of approximately
±20 nm.† The frequency of oscillation
increases with the load applied to the hair bundle by a flexible stimulus
fiber; the range of 5 100 Hz corresponds well to the characteristic frequencies
of afferent neurons innervating the sacculus.†
Application of the fluctuation-dissipation theorem, which relates a
system's mechanical responsiveness to stimulation to its reaction to thermal
noise, confirms that spontaneous oscillations involve energy expenditure by the
hair bundle.† These oscillations may
therefore supply the energy requisite for the production of SOAEs.† When a sinusoidal mechanical input as small
as ±1 nm is applied by a flexible stimulus fiber, the bundle's movement is
entrained if the frequency lies near that of unstimulated oscillation.† As judged by the amplitude of the response,
the bundle appreciably amplifies its input.†
Moreover, the bundle exhibits power gain: an active energy source in the
bundle is required to counter the power dissipated by viscous drag and that
abstracted by the stimulus fiber.† As
the amplitude of stimulation is increased, the response grows as the one-third
power of the input.† This relation,
which resembles that for basilar-membrane motion in the mammalian cochlea, is
anticipated for an amplificatory process poised near a Hopf bifurcation.† Active hair-bundle motility therefore
constitutes the active process of hair cells in the bullfrog's sacculus.† Displacement-clamp measurements reveal that
a hair bundle displays negative slope stiffness over a range of positions
subtending roughly ±10 nm.† This
phenomenon stems from the concerted gating of transduction channels: as each
channel opens or closes, the change in gating-spring force encourages other
channels to do likewise.† Two processes
power active hair-bundle motility, whether spontaneous or stimulus-evoked.† An adaptation processed mediated by myosin
Ic attempts to situate the hair bundle in the negative-stiffness region; the
bundle then lunges across this unstable domain, producing spontaneous
oscillations.† Stimuli trigger the
bundle's progression through the same trajectory, thereby entraining amplified
bundle movements.† Active bundle
movements are also driven by rapid, Ca2+ dependent reclosure of transduction
channels, a phenomenon capable of mediating oscillation even at kilohertz
frequencies.† Because the nonlinearity
responsible for negative bundle stiffness, mechanical adaptation, and Ca2+
dependent channel reclosure appear to occur universally, the same mechanisms
may power the active processes of hair cells in general, including those of the
mammalian cochlea.
Bio:† Jim Hudspeth conducted his undergraduate
studies at Harvard College and received his PhD and MD degrees from Harvard
Medical School.† Following postdoctoral
work at the Karolinska Hospital in Stockholm, he served on the faculties of the
California Institute of Technology, the University of California, San
Francisco, and the University of Texas Southwestern Medical Center.† After joining Howard Hughes Medical Institute,
Jim moved to The Rockefeller University, where he is F. M. Kirby Professor and
Head of the Laboratory of Sensory Neuroscience.† Dr. Hudspeth conducts research on hair cells, the sensory
receptors of the inner ear.† His work
has been recognized by the Spencer Award from Columbia University, the Lamport
Award from the New York Academy of Sciences, the Cole Award from the
Biophysical Society, the Dana Award from the Charles A. Dana Foundation, the
Rosenstiel Award from Brandeis University, the Hugh Knowles Prize from
Northwestern University, the Award of Merit of the Association for Research in
Otolaryngology, and the Ralph W. Gerard Prize from the Society for
Neuroscience.† Jim is a member of the
National Academy of Sciences and the American Academy of Arts and Sciences.
Web Address:† http://www.rockefeller.edu/research/abstract.php?id=67
*************************
Computer Science
Division, University of California at Berkeley
Games,
Algorithims and Networks
Abstract:† The Internet is the first
computational artifact that was not designed, but emerged from the interaction
of many agents.† As a result, it is the
first subject in Computer Science that must be approached with the same puzzled
humility with which other sciences approach the universe, the cell, the brain,
the market.† Since the Internet is both
the result and the theater of the interaction of selfish agents, mathematical
economics should be relevant to its study.†
Conversely, the Internet has enabled new kinds of interactions and
markets, and in this sense it is a challenge to that field.† In recent years we have seen an increasingly
active interface, motivated and informed by the advent of the Internet, between
the theory of games on the one hand, the theory of algorithms and complexity on
the other, and both with networking.†
Even though this corpus of research is already quite extensive and
diverse, one can identify in it at least three important themes:† Understanding the computational complexity
of the fundamental computational problems associated with games, such as
finding Nash and other equilibria; this quest is arguably of fundamental value
to game theory, and not just the result of the professional bias of a few
computer scientists.† There is also the
field of algorithmic mechanism design, striving to devise computationally
efficient methods for designing games whose equilibria are precisely the
socially desirable outcomes (for example, that the person who has the highest
personal appreciation for the item being auctioned actually wins the
auction).† Finally we have an
ever-expanding family of problems collectively given the playful name "the
price of anarchy," studying how much worse a system emerging from the
spontaneous interaction of a group of selfish agents can be when compared with
the ideal optimum design.† This talk
will review recent results and open problems in these areas.
This
research is supported by NSF ITR grant CCR-0121555 and a grant from Microsoft
Research.
Bio:† Christos H. Papadimitriou
studied in Athens Polytechnic (BS in EE 1972) and Princeton (MS in EE, 1974 and
Ph.D. in EECS, 1976).† Since then, he has
taught at Harvard, MIT, Athens Polytechnic, Stanford, and the University of
California, San Diego.† He came to
Berkeley in January 1996 (but he was there also in 1978 as a Miller fellow).† He is interested in the theory of algorithms
and complexity, and its applications to databases, optimization, AI, and game
theory. He has written these books: Elements of the Theory of Computation
(Prentice-Hall 1982, with Harry Lewis, second edition, September 1997);
Combinatorial Optimization: Algorithms and Complexity (Prentice-Hall 1982, with
Ken Steiglitz; second edition by Dover, 1998); The Theory of Database
Concurrency Control (CS Press 1988); Computational Complexity (Addison Wesley,
1994); Turing (a Novel about Computation), MIT Press, 2003, published in Greek
as Turing's smile, and a book of essays (in Greek).
He is
a member of the American Academy of Arts and Sciences and of the National
Academy of Engineering of the USA. He is an ACM Fellow and recipient of the
Knuth Prize; he holds honorary doctorates from ETH, University of Macedonia and
University of Athens.
Web
address: http://www.cs.berkeley.edu/~christos/
*************************
Alessandro Vespignani
School of Informatics,
Department of Physics and Center for Biocomplexity, Indiana University
Evolution and Structure of the
Internet
Abstract: †The talk will present an overview of the large-scale
topological and dynamical properties of real Internet maps.† First recent developments in the methodology
used to obtain large scale maps of the Internet at the router and autonomous
system level is reviewed.† Then the
statistical features and regularities observed in the large scale structure of
the Internet and the importance of the dynamics in the formulation of adequate
models are presented.† Finally the
various results and models will be scrutinized in the light of a statistical
theory that considers the map's incompleteness due to measurement biases.
Bio:†
Alessandro Vespignani, Professor of
Informatics, Complex Systems, Simulation and Modeling, Bioinformatics.† Dr. Vespignani, joined the Indiana
University School of Informatics faculty in 2004, is also Adjunct Professor of
Physics, and a member of the Core Faculty in Cognitive Science, and affiliated
with the Biocomplexity Institute.†
Vespignani obtained his Ph.D. at the University of Rome ìLa Sapienza.î† After holding research positions at Yale
University and Leiden University, he joined the condensed matter research group
at the International Center for Theoretical Physics (UNESCO) in Trieste.† Before joining Indiana University,
Vespignani was a member of the French National Council for Scientific Research
carrying out research and teaching activities at the Laboratoire de Physique
Theorique of the University of Paris-Sud.†
He has authored more than 100 scientific papers on the properties and
characterization of non-equilibrium phenomena, critical phase transitions and
complex systems.† Recently Vespignani's
research activity is focused on the interdisciplinary application of
statistical physics and numerical simulation methods in the analysis of epidemic
and spreading phenomena and the study of biological, social and technological
networks.† He was the advisor of several
graduate and undergraduate theses and organizer of international conferences
and schools.† Vespignani is author,
together with R. Pastor-Satorras, of the book Evolution and Structure of the
Internet, published by Cambridge University Press.† He was among the five scientists nominated
for the Wired Magazine Rave Award in science for 2004.
Web Address: †http://www.informatics.indiana.edu/people/profiles.asp?u=alexv