![[This is a yellowbar divider.]](http://www.sccn.ucsd.edu/~scott/icons/yellowbar.gif){kind=icon}
Event-related spectral perturbations
Author(s): Scott Makeig(1,2) and Tzyy-Ping Jung(1,3)
Institution(s): (1)Naval Health Research Center, San Diego, CA; (2)Institute for Neural Computation, University of California San Diego, and (3)Computational Neurobiology Laboratory, The Salk Institute
Address: La Jolla CA 92037
Email: smakeig@ucsd.edu
Phone: (858) 453-4100 x1455
FAX: (858) 587-0417
Abstract:
From Berger's very first EEG reports, nonstationarity in the spectral character of the EEG, as well as apparent consistency in the EEG spectral changes accompanying changes in wakefulness and attention were noted to be outstanding traits of EEG signals. In 1977, Pfurtscheller described a method of averaging the time course of event-related reductions in power in a narrow frequency-band and called the phenomena event-related desynchronization (ERD). During the last decade, increasing observations of subcortical brain centers which modulate spontaneous and event-related cortical activity, and suggestions that subcortical and cortical oscillations may serve to transiently synchronize or 'bind' activity in separate brain areas have fueled interest in observing the dynamics of EEG/MEG modulation and synchronization processes noninvasively in humans. Makeig (1993) demonstrated that averaging broad band time-frequency transforms of event-related EEG epochs in an auditory attention experiment revealed precisely-timed, both narrow- and wide-band perturbations in the power spectrum of the ongoing EEG, phenomena called event-related spectral perturbations (ERSPs). ERSPs need not fit into traditionally-defined frequency bands, and often exceed the durations and latencies of supra-DC components of event-related potentials. Time-frequency averaging may be accomplished by several methods (FFT, wavelet and Wigner transforms, matching pursuit, etc.) Both electric and magnetic spectral perturbations have definitely been shown to depend on cognitive processing as well as on stimulus character, and may represent both transient amplitude modulation of ongoing brain activity, or transient changes in synchronization of activity in previously desynchronous brain areas.
Event-related gamma band activity indicates binding in Hebbian cortical networks
Author(s): N. Birbaumer, F. Pulvermueller, W. Lutzenberger & H. Preissl
Institution(s): University of Tuebingen
Department: Institute of Medical Psychology and Behavioral Neurobiology
Address: Gartenstr. 29, D-72074 Tuebingen, Germany
Email: niels.birbaumer@mailserv.uni-tuebingen.de
Phone: (49)7071 294219
FAX: (49)7071 295956
Abstract:
In a series of experiments with meaningful and non-meaningful visual and verbal stimulus material the hypothesis of a functional role of high frequency brain oscillations for the construction of meaning was tested. EEG-frequency bands were analyzed in 10 Hz wide windows between 10 and 80 Hz in the time-frequency domain as described by S. Makeig. In the visual modality, an irregular pattern of horizontal lines in the four visual fields changed into an orderly Gestalt-like pattern (waterfall appearance). Over occipital cortices only, spectral power in the 35-45 Hz frequency band increased significantly during the gestalt-like patterns. A similar effect was seen in the 25-35 Hz range over left perisylvian cortices for meaningful words in comparison to meaningless pseudowords. These data support a Hebbian model of Gestalt processing in the brain as deduced from animal experiments with simultaneously moving visual stimuli (Singer 1995). Supported by the German Research Society (DFG).
Characterization of P3 using matching pursuit
Author(s):
B. Shen(1) & John J. B. Allen(2)
Institution(s): (1) New Jersey Neuroscience Institute, JFK medical center (2) Psychology Department, University of Arizona
Address: 65 James Str. P.O. Box 3059, Edison NJ08818-3059
Email: bshen@jfk.hbocvan.com
Phone: (908)321-7000x2551
FAX: (908)632-1584
Abstract:
To enhance the signal-to-noise ratio, event-related potentials (ERPs) are often averaged over many trials of repeated stimuli. However, this approach presumes that cognitive processes remain stationary over those trials, which may not be valid in many experimental paradigms. Traditional single-trial methods are heuristic and have failed to extract information from noisy signals reliably. A recently developed technique, matching pursuit (Mallat and Zhang, 1993), offers adaptive time-frequency decomposition and has been shown to be effective in hippocampal EEG analysis (Shen et al., 1995) and ERP studies (Allen et al., 1995). In this study, we use matching pursuit analysis to study the statistics and the time frequency character of the classic P3 components of the ERPs in a memory-assessment task. The P3 was mathematically modeled by an adaptive Gaussian waveform while oscillatory wavelets were screened out. This method offered a better measure of P3s because its signal-to-noise ratio of a single-trial ERP was larger. The statistics of the amplitudes, the latencies and the durations of single-trial P3s were more informative than the averaged ERP amplitudes. The correlation between the P3s and other time-frequency components suggested that there existed a distinct time-frequency character of the P3 components. This approach provides a new method for examining ERP data, and a means of accessing the non-stationarity of cognitive processes.
Stimulus induced coherence:
a measure for interareal synchronization
Author(s):
Astrid von Stein
Institution(s): The Neurosciences Institute, San Diego, CA, Institute of Neurophysiology, University of Vienna, Austria The Neurosciences Institute, 10640 John Jay Hopkins Drive San Diego CA 92121
Email: astein@nsi.edu
Fax: (619) 626-2099
Abstract:
A paradigm shift has occurred in neurobiology. The old concept of single cell coding is being replaced by coding in DISTRIBUTED NEURONAL CELL ENSEMBLES. Results from cat intracortical recordings have shown that synchronization among the members of a cell ensemble play a major role in this process. This makes it desirable to measure synchronization also in humans. In my talk I will show that synchronization of neuronal cell ensembles can be detected in the frequency components of human scalp EEG. According to our results, spectral power in the high beta and in lower frequency ranges reflects the degree of synchronization between columns within a cortical area (von Stein et al. 1995). Furthermore coherence, the normalized crosspower between two cortical signals, gives us the possibility to determine synchronization BETWEEN DIFFERENT CORTICAL AREAS. These interareal synchronizations are of specific importance for higher cortical processes where neuronal ensembles are thought to extend over large distances. I will present data on local and interareal synchronization during auditory and visual perception and during several higher cognitive tasks. We present stimuli lasting for 1-2 s and measure the averaged power/ crosspower spectra over 30-50 trials. We find significant task-specific increases of coherence between visual cortex and associative cortex during visual perception, between several sites in parietal cortex of both hemispheres during spatial imagery tasks, between temporal and parietal cortex during feature integration, and more. Changes occur in single frequency ranges or in combinations of frequency ranges. These induced interareal synchronizations are usually reproducible within single subjects and reveal significant consistencies for groups of subjects. Thus, EEG coherence analysis seems to be an adequate tool for measuring synchronization within extended neuronal ensembles directly in humans.
Event-related power shifts in the theta and alpha bands
during encoding and retrieval
Author(s): Wolfgang Klimesch
Institution(s): Department of Physiological Psychology, University of Salzburg,
Address: Hellbrunnerstr. 34, A-5020 Salzburg, Austria
Email: Wolfgang.Klimesch@sbg.ac.at
Fax: 0043 662 8044 5126