Title: General Signal Processing and Machine Learning Tools for BCI Analysis (Part-1)
1General Signal Processing and Machine Learning
Tools for BCI Analysis (Part-1)
- Md. Zia Uddin
- Bio-Imaging Lab, Department of Biomedical
Engineering - Kyung Hee University
2Contents
- Abstract
- Introduction
- Spectral Filtering
- Special Filtering
- Classification
3Abstract
- This presentation is about signal processing and
machine learning techniques and their
applications to BCI. - Overview of general signal processing and
classification methods as used in single-trial
EEG analysis is given. - For further study, original publications are
encouraged.
4Why ML for BCI
- Subject wise experiments
- Subject to subject result variances for same kind
of experiments. - Session wise experiments
- Session to session huge result variability for
the same person. - Real time experiments
- The system needs to identify the subjects mental
state from single trial. - Much more complexity arises.
- Solution
- A session and user brain signature adaptable
system is necessary to overcome the subject to
subject and session to session huge variability.
5Why Preprocessing?
- Relevant information extraction is difficult
because of large dimensional data (i.e., Curse of
dimensionality). - Dimensionality has to be reduced keeping the
discriminative information and eliminating
undiscriminative information. - Most of the classification methods calculate
covariance matrix of the data for further feature
analysis. Huge covariance matrix is required in
the case of large dimensional data. - Thus, Prepropcessiong steps regarding
dimensionality reduction is required - In some cases
- A priory knowledge is used (e.g., spatial Laplace
filter at predefined scalp locations) - Automatic methods (e.g., spatial filters
determined by common spatial pattern analysis)
6Spectral Filter FIR IIR
- Common approach is to use digital frequency
filter - To consider desired frequency range
- Two sequences of poles (a) and zeroes (b) with
length na and nb are necessary that can be
calculated by Butterworth or elliptic. - The source signal x is filtered to y as
- a(1)y(t)b(1)x(t) b(2)x(t-1) ...
b( nb )x(t- nb -1) a(2)y(t-1) -...- a( na )y(t-
na -1) - Where a and na are constrained to be 1, is called
FIR filter (i.e., considering all zeros). - Advantage of FIR
- Produce steeper slopes in between pass and stop
band. - In most of the BCI applications, band pass filter
is required to consider specific frequency range.
7Spectral Filter Fourier-Based Filter
- A good alternative than FIR and IIR is to use
temporal Fourier-based filtering in BCI. - A signal switches from temporal to the spectral
domain. - The filtered signal is obtained by choosing
suitable weighting to the relevant frequency
components and applying Inverse Fourier
Transformation (IFT). - The short time window determines the frequency
resolution
8Spatial Filter Bipolar Filtering
- EEG channels are measured as voltage potential
relative to a standard reference (referential
recording). - Also, it is possible to record all the channels
as voltage difference between the electrode
pairs. - From referential EEG, bipolar channels can be
obtained by subtracting the respective channels - FC4-CP4(FC4-ref) - (CP4-ref)FC4ref -CP4ref
- Reduces the effect of local smearing by computing
local gradient. - Focuses on the local activity while contributions
of more distant sources are attenuated
9Spatial Filter Common Average Reference
- The mean of all EEG channels are subtracted from
each channel to get the common average reference
signals. - Reduces the influence of far field sources but
may introduce some undesired spatial smearing - Artifacts of one channel may spread to all other
channels.
10Spatial Filter Laplace Filtering
- More localize filter can be obtained through
this. - Laplace signals are obtained by subtracting the
average of surrounding electrodes from each
individual channel. - C4Lap C4ref- ¼(C2ref C6ref FC4ref CP4ref)
- The choice of surrounding channels determine the
characteristics of the filter. - Usually, small Laplacians are used (as example
given above). - Large Laplacians use neighbors at 20 distance
as defined in international 10-20 system.
11Spatial Filter Principle Component Analysis(1)
- Represent multidimensional data with fewer number
of variables retaining main features of the data.
- It is inevitable that by reducing dimensionality
some features of the data will be lost. It is
hoped that these lost features are comparable
with the noise and they do not tell much about
underlying population. - The method PCA tries to project multidimensional
data to a lower dimensional space retaining as
much as possible variability of the data. - Its simplicity makes it very popular. But care
should be taken in applications. First it should
be analyzed if this technique can be applied.
12Spatial Filter Principle Component Analysis(2)
- Orthogonal directions of greatest variance in
data - Projections along PC1 discriminate the data most
along anyone axis - First principal component is the direction of
greatest variability (covariance) in the data. - Second is the next orthogonal (uncorrelated)
direction of greatest variability - So first remove all the variability along the
first component, and then find the next direction
of greatest variability - And so on
13Spatial Filter Principle Component Analysis(3)
- We can ignore the components of lesser
significance. - We do lose some information, but if the
eigenvalues are small, we dont lose much - n dimensions in original data
- calculate n eigenvectors and eigenvalues
- choose only the first p eigenvectors, based on
their eigenvalues - final data set has only p dimensions
Basic Steps
Eigenplot
14Visual Evoked Potential Extraction from Single
Trial EEG signals using PCA filtering(1)
- Problem Definition
- Remove the noise to get VEP in the single trial
29 channels EEG data without ensemble averaging - Technique adopted to solve the Problem
- Selection of principal components as basis for
the reconstruction of signal - Methodology
- Given signal is divided into an ensemble of
signals, for each channel - An ensemble average for each channel is obtained
as a reference - Apply PCA to find out the orthonormal
eigenvectors which are used as basis for signal
approximation - Selection of Principal components as basis by
looking at the frequency components present in
the prototype signal i.e. the averaged signal
15Visual Evoked Potential Extraction from Single
Trial EEG signals using PCA filtering(2)
Original Signal
Single epoch after PCA filtering
Reconstructed epoch stacks
16Spatial Filter Independent Component Analysis(1)
- Basically ICA is applied for Blind Source
Separation (BSS) - Assume an observation (signal) is a linear mix of
unknown independent source signals - The mixing (not the signals) is stationary
- We have as many observations as unknown sources
- To find sources in observations
- Need to define a suitable measure of independence
- For example - the cocktail party problem
(sources are speakers) Find Z - Formal Statement
- N independent sources Zmn ( M xN )
- linear square mixing Ann ( N xN )
- produces a set of observations Xmn ( M xN )
- .. XT AZT
17Spatial Filter Independent Component Analysis(2)
- demix observations XT ( N xM ) into YT WXT
YT ( N xM ) ? ZT W ( N xN )
? A-1 - How do we recover the independent sources?
- (We are trying to estimate W ? A-1 )
- . We require a measure of independence!
18Spatial Filter Independent Component Analysis(3)
- The source signals are mixed by random non
orthogonal matrix - JADE algorithm was applied to demix the signals
- After reordering and scaling, the demixed signals
are very similar to sources. - PCA would fail here as the mixed signals are not
orthogonal to each other, which is the key
assumption of PCA. - Other ICA algorithms
- Infomax
- FastICA
19Spatial Filter Independent Component
Analysis(4) Applying ICA to single-trial EEG
epochs
- Data Collection
- EEG data were recorded from 31 scalp electrodes
- 29 placed at locations based on a modified
International 10-20 system - one placed below the right eye (VEOG),
- one placed at the left outer canthus (HEOG). All
- 31 channels were referred to the right mastoid
and were digitally sampled for analysis at 256
Hz with a 0.01- to 100-Hz analog bandpass plus a
50-Hz lowpass filter. - Subjects participated in a 2-hour visual spatial
selective attention task in which they were
instructed to attend to filled circles flashed
in random order in five locations.
- Component IC1, generated by blinks
- IC4 generated by temporal muscle activity.
20Spatial Filter Independent Component
Analysis(5) Applying ICA to single-trial EEG
epochs (2)
- The scalp maps and power spectra of the 31
independent components derived from target
response epochs from a 32-year-old autistic
subject. - Blink and eye movement artifact components (IC1
and IC9) had a typical strong low frequency
peak. - Temporal muscle artifact components (i.e., ICs
14, 22, 27, and 29) had characteristic focal
optima at temporal sites and power plateaus at 20
Hz and higher.
21Conclusion
- Next class
- more classification techniques and some
practical examples.
22