|
Spatio-temporal EEG data processing under MATLAB
 What is g.EEGtoolbox for g.BSanalyze?
g.EEGtoolbox is an interactive environment for spatio-temporal
EEG data processing.
Specialized functions for EEG artifact detection and
data quality computations assure highest-quality results. Adaptive
Filtering or bandpower analyses allow to extract EEG signal parameters.
The parameter extraction section includes also the extraction of Hjorth
and Barlow parameters.
Signal similarities in the time domain are determined by
cross-correlation methods. In the frequency domain coherence
analysis yield linear couplings as function of frequency. Signal Averaging
methods allow to determine phase-locked Evoked-Potentials. A section
with sophisticated functions for Event-Related Desynchronization and
Event-Related Synchronization analyses allow to investigate stimulus- but
non-phase locked EEG phenomena.
The graphical user interface for g.EEGtoolbox enables
you to explore EEG data sets and to extract relevant information that can
not be seen from raw data only. Results of multi-channel computations
are visualized in the result2D tool allowing also topographical
arangement of the results. You can load and save your preferred
processing steps as a script program and automatically process your data in
batch mode. Toolbox Highlights
- Data quality
- Power spectrum estimation and comparison with significance analysis
- Bandpower estimation, Hjorth Parameter estimation, Barlow Parameter
estimation
- Cross correlation based template matching
- Wavelet Analysis
- Electrode montages definition and topographical mapping
- Event-related coherence / Coherence analysis
- Event-Related Desynchronization (ERD) / Event-Related Synchronization (ERS)
analysis
- Time-Frequency ERD/ERS analysis
Example: Determination of Event-Related Synchronization (ERD)
 Power Spectral Density estimation (Fourier
Transformation)
The first
step in ERD quantification is to identify the subject's specific most
reactive frequency bands.EEG during right hand and foot movement imagery was
recorded over electrode positions C3 and Cz. The type of movement imagery
was given via experimental instructions on the computer screen. A total of
160 imaginations of movements were performed (80 hand and 80 foot
movements).
The analyze menu Spectrum allows to compute
the power spectral density (PSD) of your data. PSD is a measure of how
power in a signal changes as a function of frequency. The spectral
analysis detects periodic oscillations (amplitude and frequency) and has
been employed in a great variety of signal processing applications.
g.BSanalyze allows not only to compute the
PSD for specific time segments, but also to statistically compare
the PSD between data segments. Hence differences in the PDS between
reference periods versus active periods can be found easily.
The figure to the right displays the PSDs for an EEG signal in the
reference period (color coded blue) and active period (color coded green).
At a frequency of about 17 Hz there is a significant amplitude difference
for the 2 period. The graph at the top (color coded violet) displays the
amplitude difference. The dashed lines represent the 95% significance
level. Values exceeding the dashed lines indicate
significant differences.
 Quantification of ERD as function of time
The second step is to quantify ERD within the most
reactive frequency bands as function of time. Hence one can observe time
evolution of ERD over different brain areas. Different options exist to express Event-Related Synchronization and
Event-Related Synchronization.
(i) ERD/ERS values can be expressed as absolute power values varying
over time or
(ii) ERD/ERS values can be related to a so-called reference period.
Then time varying ERD/ERS values are given in percentage values indicating
the relative power decrease or increase with respect to the reference
period.
The figure to the left displays percentage values of ERD/ERS over time.
The reference period is indicated as a horizontal red line. After second 2
ERD values decrease to about -50%. Between second 5 and 8 an ERS can be
observed with a bandpower increase of up to 300 %.
 Time-Frequency distribution of ERD/ERS
g.EEGtoolbox allows for a third
option in calculating ERD/ERS:
A novel time-frequency computation and representation allows to conveniently
identify the most reactive frequency bands and the related channels.
The analyze menu ERDmaps
allows to calculate ERD and ERS for multiple frequency bands and to present
the results in a single colormap per channel.
ERDmaps can be calculated without a significance statistics (fast) or
with a bootstrap algorithm. When using the bootstrap algorithm it is
possible to define a significance level. Then, ERD and ERS are only
displayed if they are within the specific confidence range.
 Example: Topographical Mapping of results
The result of spatio-temporal computations can be displayed in the 2D result presentation tool g.Result2D.
After defining the corresponding electrode montage for the
EEG experiment in the montage creator g.MONcreator, including
electrode co-ordinates and electrode names, computational results can be
displayed topographically.
The figure to the left displays the average and standard
deviation for a 60 channel ERD/ERS experiment.
Topographical plots of grand averages and standard
deviations are a very useful tool to discover artifacts in the data.
If technical artifacts (e.g. cross-talk of trigger channels)
can not be ruled out, then such artifact can be detected and made visible by
simple averaging across trials.
Package includes
- Software modules
- help
manual
- hardlock
Technical Requirements
MATLAB, g.BSanalyze base versionCopyright
© g.tec
|
