A N N O U C E M E N T O F A S Y M P O S I U M
Frontier Science in EEG Symposium:
C O N T I N U O U S W A V E F O R M A N A L Y S I S
For scientists in the field of signal analysis
in both biomedical and physical sciences.
SATURDAY, OCTOBER 9, 1993
COURSE REGISTRATION INFORMATION AND CONTACT PEOPLE ARE LISTED AT THE END OF
* To be held in conjunction with
* the annual meeting of the American EEG Society
* Oct. 11 - 13, 1993
* call 203-243-3977 about regist. for the Am. EEG Soc. meeting
[this will not register you for the symposium - see below]
NEW ORLEANS MARRIOTT
NEW ORLEANS, LOUISIANA
University of Pittsburgh Medical Center
Center for Continuing Education in the Health Sciences
Supported by an Educational Grant from:
Parke-Davis, Div. of Warner-Lambert
Electroencephalography (EEG) is the study of the electrical activity of the
brain. The field of EEG includes the technology to record these electrical
signals, the science to analyze them and the expertise to apply them to
patient care. This symposium will explore the scientific frontiers related
to EEG, presenting the latest research and thought with this year's topic
being continuous waveform analysis. As advances in science and technology
often involve collaboration among scientists from different fields, we are
bringing together a diverse group of investigators, many from areas not
conventionally associated with EEG, to actively encourage multidisciplinary
research in EEG and foster new ideas.
WHO SHOULD ATTEND
This conference is designed for anyone dealing with signal analysis,
especially neurologists, neurophysiologists, electroencephalographers,
mathematicians, physicists, computer scientists, astronomers, and engineers.
Richard M. Dasheiff, M.D.
Univ. Pittsburgh Epilepsy Center
email rmd at med.pitt.edu
Diana Joan Major-Vincent, Ph.D.
University of Pittsburgh
email dvi at med.pitt.eduvincent at pittvms.bitnet
John K. Chapin, Ph.D.
Dept. Physiology & Biophysics
D. Kent Cullers, Ph.D.
Lester Ingber, Ph.D.
Lester Ingber Research
Ben Jansen, Ph.D.
Department of Electrical Engineering
University of Houston
B. H. Juang, Ph.D.
Supervisor, Speech Research Department
AT&T Bell Labs
Periklis Ktonas, Ph.D.
Department of Electrical Engineering
University of Houston
Fernando Lopes da Silva, M.D., Ph.D.
Institute of Neurobiology
Paul L. Nunez, Ph.D.
Brain Physics Group
Department of Biomedical Engineering
Tulane University School of Engineering
Edward T. Olsen, Ph.D.
Robert D. Sidman, Ph.D.
Department of Mathematics
University of Southwestern Louisiana
M. Victor Wickerhauser, Ph.D.
[ Conflict of Interest Disclosure Statement ]
[ In accordance with the policies on disclosure of the Accreditation Council ]
[ for Continuing Medical Education and the Faculty Advisory Committee for ]
[ Continuing Education in the Health Sciences, University of Pittsburgh, ]
[ presenters for this program have identified no personal relationships which,]
[ in the context of their topics, could be perceived as real or apparent ]
[ conflict of interest. ]
PROGRAM for SATURDAY OCTOBER 9, 1993
7:00 - 7:30am - Registration with continental breakfast
7:45am - Introduction: Richard Dasheiff,M.D. Univ. of Pittsburgh
Focus is on continuous, time-series, alternating current/signals.
Standing potentials (DC EEG), Evoked Potentials,
Non-Stationary Processes, Epilepsy, Intensive care unit (ICU)
monitoring, etc., will be topics for future symposia.
7:55am - Moderator: Diana Major-Vincent, Ph.D. Univ. of Pittsburgh
8:00am - Biophysics of the EEG and Neural networks: Fernando Lopes da
Silva, M.D., Ph.D. Univ. of Amsterdam
This talk will address the relationship between dynamic aspects
of the biological generators of the EEG and the structural constraints
imposed by the anatomy of the brain (a non-homogeneous, irregularly
shaped structure) and of the surrounding tissues. Topics include the
effect these biological structures have on signal degradation, the
differences in signals from intracranial and extracranial recording
(spatial and temporal dispersion, S/N, voltage loss, extraneous
noise), and how these affect recording and modeling of the EEG.
8:40am - History of source localization and bioelectric imaging:
Robert Sidman, Ph.D. Univ. of Southwestern Louisiana
What should a mathematical model of the EEG do? An early model,
"nonlinear oscillations and the EEG", Dewan's 1964 paper may have been
chaos 30 years before its time. Source localization in the time
domain (dipole localization methods) offers a parsimonious model
requiring few assumptions, but has limitations. Imaging or
topographical mapping methods in the spatial domain such as CIT and
the Spline-Laplacian are also presented with a discussion of the
mathematical ideas that led to CIT. Many of the methods for analyzing
the EEG have their roots in a few seminal research papers. I will
discuss these papers and their relationship to "modern" techniques.
9:20am - Computer-based EEG pattern recognition: Periklis Ktonas,
Ph.D. Univ. of Houston
Computer-based recognition of specific EEG signals (e.g.
spindles, spikes) in multichannnel EEG records has relied on
mathematical techniques (e.g. spectral analysis, "optimum" filtering)
and on heuristic "mimetic" approaches, including expert system-based
methods. Recently, nonlinear discriminant analysis methodologies
under the broad term "neural networks" have been tried as well. This
presentation will focus on the pitfalls, achievements and overall
usefulness of the above techniques.
10:00 - 10:20am Coffee Break
10:30am - Spatial Analysis of EEG: Paul Nunez, Ph.D. Tulane Univ.
The field of scalp recorded EEG spans about nine orders of
magnitude of temporal scale (Ktonas, Dasheiff), but spatial resolution
is limited to several cm's (Sidman). Neocortical dynamics evidently
involves neural interactions at multiple spatial scales (Lopes da
Silva), with effects that cross hierarchical levels from microscopic
to the macroscopic scale of scalp EEG (Ingber). Modern measures of
dynamic function based on a small number of EEG channels may result in
a distorted view of this dynamics (Jansen, Ingber, Juang).
The severe limitations of conventional EEG and new methods which
increase spatial resolution by about a factor of three are reviewed.
The ways in which this vast amount of new information may be used in
clinical and cognitive studies is considered. The implications of
recording strategy for various estimates of dynamic function are
11:10am - Chaos and Quantitative EEG Analysis: Ben H. Jansen, Ph.D
Univ. of Houston
Support for a nonlinear and/or chaotic nature of the human
electroencephalogram (EEG) is reviewed. It is argued that dimension
calculations and other classical methods to assess chaos are not
reliable in the case of EEG data because it is impossible to ascertain
that the system (i.e., the brain) is time-invariant and that it
displays asymptotic behavior during the interval of observation.
Despite the lack of hard scientific evidence, there are ample reasons
to conclude that the EEG is produced by a nonlinear, possibly chaotic,
(deterministic) system. The implications of this change from a
(stochastic) to a deterministic viewpoint of EEG generation on the
practice of visual and quantitative EEG interpretation is discussed.
It is argued that conventional EEG analysis techniques will have to be
revised and that previously reported findings need to be reinterpreted
in the light of the chaotic or nonlinear nature of the EEG.
11:50am - 12:20pm Panel Discussion
12:20 - 1:20pm Lunch (included in tuition)
1:20pm - Wavelets, Adapted Waveforms, and Denoising: Victor
Wickerhauser, Ph.D. Washington Univ.
The goal is to describe some new libraries of waveforms well-
adapted to various numerical analysis and signal processing tasks.
The main point of this presentation is that by expanding a signal in a
library of waveforms which are well-localized in both time and
frequency, one can achieve both understanding of structure and
efficiency in computation. The talk will briefly cover the properties
of the new "wavelet" and "localized trigonometric" libraries. The
main focus will be applications of such libraries to the analysis of
complicated transient signals: a feature recognition algorithm based
on fast approximate principal factor analysis, a feature extraction
and data compression algorithm for acoustic signals which uses best-
adapted time-frequency decompositions, and a wavelet-based denoising
algorithm for passive sonar and music. These signals share many of
the same features as EEG traces, and the algorithms are directly
useful for that application as well.
2:00pm - Multiple Scales of EEG: Lester Ingber, Ph.D. Lester Ingber