1. State-of-the-Art EEG Signal Processing
Techniques in the Diagnosis and
Treatment of Neurological Diseases
Md. Kafiul Islam, PhD, SMIEEE
Assistant Professor, Dept. of EEE, IUB
Lab Head, Biomedical Instrumentation and Signal Processing Lab
Web: https://www.researchgate.net/profile/Md_Kafiul_Islam
Email: kafiul_islam@iub.edu.bd
Dr. Md. Kafiul Islam
1
2. About Us!
Biomedical Instrumentation and Signal
Processing Lab
Featured Projects:
1. EOG based Computer Mouse Cursor Control for Physically Challenged
People
2. Design and Development of a EOG-based System to Control Electric
Wheelchair for People Suffering from Quadriplegia/Quadriparesis
3. Design and Implementation of a Low-cost EMG Signal Recorder for
Application in Prosthetic Arm Control for Developing Countries.
4. Development of a Low Cost and Portable ECG Monitoring System for
Rural/Remote Areas of Bangladesh
5. Emotion Classification using Wireless EEG Devices
6. Motion Artifact Removal from Ambulatory EEG
Dr. Md. Kafiul Islam
2
3. Motivation
Patient Monitoring (critical care)
ICU, Coma, NICU
Disease Detection/Diagnosis
Brain (EEG)
Preventive Healthcare (Prediction)
Brain-Machine Interfacing (BMI)
BCI, Prosthetesis
Controlling wheelchair, prosthetic arm
Treatment and Rehabilitation
Basic understanding of the brain’s physiology and function
Patient › EEG Signals › Processing › Decision2
Dr. Md. Kafiul Islam
3
[2] https://www.embs.org/about-biomedical-engineering/our-areas-of-research/biomedical-signal-processing/
4. Digital Signal Analysis and Processing
Analysis
Measurement of signal properties
Looking at the signals in other than time domain (frequency, wavelet, etc.)
Processing
Remove unwanted noise/interference/artifacts (Filtering)
Improve signal quality (SNR)
Extract specific or critical information
Detect useful events
Dr. Md. Kafiul Islam
4
5. Challenges with EEG Signals3
Very low SNR
Non-stationarity and non-linearity
properties
Diverse types of noise
Noise and Artifacts overlap both in
Spectral and Temporal Domain
False alarms and errors can
mislead the diagnosis, treatment or
control of any external devices
Traditional Signal processing
techniques fail
Dr. Md. Kafiul Islam
5
[3] Biomedical Instrumentation by Ákos Jobbágy and Sándor Varga
6. Brain Recordings!
Non-Invasive
EEG (electrical) and MEG (magnetic)
Invasive
Sub-scalp EEG (S-EEG)
ECoG / iEEG
Intracortical Implant
Better signal quality but …!?
Motivation to record brain signals:
How brain works?
Information processing
Memory formation
Understanding neurological disorders
& providing treatment
Dr. Md. Kafiul Islam
6
7. What is EEG?
Recording of the brain's spontaneous electrical activity over a period of time
by placing flat metal discs (electrodes) attached to the scalp.
Measures voltage fluctuations resulting from ionic current within the neurons
of the brain.
Scalp-EEG: Most popular brain recording technique
Low cost
Non-invasive
Portable
Reasonable temporal resolution
The first human EEG recording obtained by Hans Berger in 1924. The
upper tracing is EEG, and the lower is a 10 Hz timing signal.
Electro-encephalo-gram Electrical-Brain-Picture
Dr. Md. Kafiul Islam
7
8. EEG Acquisition
Standard 10-20 Electrode Montage Acquisition involves
Electrodes
Amplification (IA)
Filtering (Active)
Digitizing (ADC)
Storage
Challenge: To detect signal as
low as 1 µV
Dr. Md. Kafiul Islam
8
9. Delta: (< 4 Hz)
◦ adult slow-wave sleep
◦ in babies
◦ continuous-attention tasks
Theta: (4 - 8 Hz)
◦ higher in young children
◦ drowsiness in adults and teens
◦ idling
Alpha: (8 - 13 Hz)
◦ relaxed/reflecting
◦ closing the eyes
Beta: (14 - 30 Hz)
◦ active thinking, focus, high alert, anxious
Gamma: (> 30 Hz)
◦ Displays during cross-modal sensory processing
Mu: (7.5 - 12.5 Hz)
◦ Shows rest-state motor neurons
Gamma
•Rhythmic: EEG activity consisting in waves of approximately constant frequency.
•Arrhythmic: EEG activity in which no stable rhythms are present.
•Dysrhythmic: Rhythms and/or patterns of EEG activity that characteristically appear in patient
groups or rarely or seen in healthy subjects.
Dr. Md. Kafiul Islam
9
10. Applications of EEG
Epilepsy Diagnosis: Seizure Detection Most
Common
Also to diagnose Alzheimer’s disease,
Parkinson’s, Sleep
disorders, Coma, Encephalopathy, brain injury,
and brain death
Brain-Computer Interface (BCI) / Neural
Prosthesis
Basic Neuroscience Research
Cognitive science, cognitive psychology, and
psychophysiological research.
Dr. Md. Kafiul Islam
10
11. Applications of EEG
(Epilepsy Diagnosis: Seizure Detection)
Dr. Md. Kafiul Islam
11
Typical waveforms recorded during seizures in
an animal model of absence epilepsy caused
by the chemo convulsant pentyleneterzol (PTZ).
The anterior thalamus is a sub-cortical
element that has shown to have strong
association with cortex during these seizures.
The spike and wave shape is clearly evident in
anterior thalamic and cortical derivations
12. Applications of EEG
(Alzheimer’s Disease Diagnosis)
Fiscon, Giulia, et al. "Combining EEG signal processing with supervised methods for Alzheimer’s
patients classification." BMC medical informatics and decision making 18.1 (2018): 35.
12
14. Applications of EEG
(Diagnosis of Encephalopathy)
Dr. Md. Kafiul Islam
14
Jacob, Jisu Elsa, Gopakumar Kuttappan Nair, Thomas Iype, and Ajith Cherian. "Diagnosis of encephalopathy based on energies of
EEG subbands using discrete wavelet transform and support vector machine." Neurology research international 2018 (2018).
15. Common Signal Analysis & Processing Techniques
Fourier Transform / Power Spectral Density (PSD/Pwelch)
Filtering (FIR/IIR: Notch, LPF, BPF, HPF)
Advanced Techniques
Adaptive filtering (use of reference channel)
A-priori user input required
Wiener/Kalman/Particle Filtering
Blind Source Separation
ICA/CCA/MCA/SVD
Time-series Analysis
Wavelet transform / Spectrogram (STFT)
Empirical Techniques (Non-linear & non-stationary)
HHT/EMD/EEMD
Dr. Md. Kafiul Islam
15
16. Power Spectral Density, PSD (Pwelch)
Dr. Md. Kafiul Islam
16
Effect of Notch Filtering on Time Domain Signal
Effect of Notch Filtering on PSD
18. Adaptive Filtering (EEG Artifact Removal6)
Dr. Md. Kafiul Islam
18
[5] http://www.jscholaronline.org/articles/JBER/Signal-Processing.pdf
[6] https://www.sciencedirect.com/science/article/pii/S098770531630199X
Reference ECG
Artifactual EEG
Clean EEG after ECG pulses are removed
19. Independent Component Analysis (ICA)
(for EOG and EMG artifact removal from EEG7)
Dr. Md. Kafiul Islam
19
[7] https://sccn.ucsd.edu/~jung/Site/EEG_artifact_removal.html
Recorded Multi-channel EEG Signals
Artifacts Contaminated EEG ( 𝑿)
After applying ICA, Independent
Components are separated and
Identified ( 𝑺′)
Artifactual ICs are removed
followed by Inverse ICA to get
Artifacts Corrected EEG
ICA
Un-mixing 𝑾
Inverse ICA
Mixing 𝑨
Blink Artifact
EMG Artifact
20. 20
Wavelet Transform
(A Multi-resolution Analysis)
Split Up the Signal into a Bunch of Signals
Representing the Same Signal, but all Corresponding to Different
Frequency Bands
Only Providing What Frequency Bands Exists at What Time Intervals
Presented By Md Kafiul Islam
Translation
(The location of
the window)
Scale
Mother Wavelet
dt
s
t
tx
s
ss xx
*1
,,CWT
Translation
(The location of
the window)
Scale
Mother Wavelet
From http://www.cerm.unifi.it/EUcourse2001/Gunther_lecturenotes.pdf, p.10
Wavelet
Small wave
Means the window function is of finite length
Mother Wavelet
A prototype for generating the other window functions
All the used windows are its dilated or compressed
and shifted versions
Scale S>1: dilate the signal
S<1: compress the signal
21. Wavelet Transform on EEG Signals
(for separating different EEG sub-band rhythms8)
Dr. Md. Kafiul Islam
21
[8] http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7856467
22. Empirical Mode Decomposition (EMD) on EEG Signals
(for detecting Eye-blink artifacts9)
Dr. Md. Kafiul Islam
22
[9] https://www.researchgate.net/publication/228524659_Real_time_characterisation_of_neural_signals_with_application_to_neonatal_monitoring/figures?lo=1
23. Artifacts & Interferences in EEG10
Dr. Md. Kafiul Islam
23
[10] https://emedicine.medscape.com/article/1140247-overview
24. Artifacts Found in EEG* (1)
Horizontal Eye
Movement (T1-T2)
and
Blink Artifacts
(Frontal channels)
EOG
*Chang, B. S., Schachter, S. C., & Schomer, D. L. (2005). Atlas of ambulatory EEG.
Amsterdam ; London: Amsterdam ; London : Elsevier Academic Press, c2005.
Dr. Md. Kafiul Islam
24
25. Artifacts Found in EEG (2)
Rhythmic eye flutter can occur at quite high frequencies, in this case 5–8 Hz, and
trigger automated seizure detection algorithms.
This artifact can be distinguished from ictal events in different ways, including the
lack of evolution in frequency or amplitude over time.
Eye Flutter Artifact
recorded by Seizure
Detection Algorithm
Dr. Md. Kafiul Islam
25
26. Artifacts Found in EEG (3)
Electrode Tapping
Artifact
With ambulatory EEG recording, there may be periods in which the patient
is scratching or pressing on the scalp electrodes.
Dr. Md. Kafiul Islam
26
27. Artifacts Found in EEG (4)
Jaw-Clenching
Artifact
Periods of jaw clenching and biting during ambulatory EEG recording can be associated with artifact.
This activity is often most prominent over the temporal regions, presumably due to temporalis or
masseter muscle contraction.
Dr. Md. Kafiul Islam
27
28. Artifacts Found in EEG (5)
Chewing Artifact
Chewing artifact can often have rhythmic features and at times may resemble ictal activity.
This artifacts are characterized by the rhythmic appearance of muscle artifact centered mostly
over the temporal regions bilaterally.
Dr. Md. Kafiul Islam
28
29. Artifacts Found in EEG (6)
Chewing Artifact
Recorded by
Spike Detection
Algorithm
Chewing artifact can have rhythmic and sharp features, and can thus be picked up by
automated detection algorithms.
Here, the spike detections at 23:53:23 and at 23:53:34 capture high frequency sharp waveforms
that are artifactual in nature, in association with a nighttime snack.
Dr. Md. Kafiul Islam
29
30. Artifacts Found in EEG (7)
Dry Electrode Artifact
One disadvantage of ambulatory EEG monitoring is the inability of technologists to check on
recording quality frequently in real time and to respond with appropriate technical adjustments as
needed throughout the day, as would typically occur in an inpatient monitoring unit.
Here, a dry electrode at F8 causes an artifact seen throughout this excerpt.
Dr. Md. Kafiul Islam
30
31. Artifacts Found in EEG (8)
Forehead Rubbing
Artifact
As with many other patient movement artifacts, rhythmic rubbing artifact can be a potentially
confusing finding on ambulatory EEG monitoring, particularly in the absence of concomitant video
recording. Here, rhythmic rubbing of the forehead produces an artifact in the anterior channels,
extending from the beginning of this excerpt through 20:50:16.
Differentiation of such artifact from an ictal event, based on the lack of evolution of the artifact
in frequency or amplitude, is critical.
Dr. Md. Kafiul Islam
31
32. Artifacts Found in EEG (9)
Pulse Artifact
Pulse artifact is caused by the rhythmic movement of scalp electrodes, usually over the
temporal regions, by the pulsations of blood through vessels close to the skin, such as the
superficial temporal artery.
It is to be distinguished from the more common electrocardiographic (EKG) artifact in that pulse
artifact is typically not as sharply contoured, resembles a slow wave, and follows the QRS complex
rather than being simultaneous to it.
Here, pulse artifact is seen in channels 9 and 10 over the left temporal region throughout this
entire excerpt.
Dr. Md. Kafiul Islam
32
33. Purpose:
Diagnosis/Detection of Epilepsy Seizure by Long-term EEG Monitoring (up to 72 hours)
Early warning of seizures (prediction) onset in order to stop seizure
Special Features
Continuous monitoring
Wireless EEG (up to 10 m or so)
Usually Dry electrode
Few no. of channels
typically 4-8 channels
Patient can lead normal daily life
Except few tasks (e.g. taking bath)
Ambulatory EEG
34. Head Movement ( < 10 Hz )
Shake
Nod
Roll
Hand Movement
Eye Movement
Eyebrow Movement
Eye blinking
Horizontal and Vertical Eye Movement
Extra-ocular muscle activity
Jaw Clenching/Eating/Drinking
Walking
Running
Sleeping
Drowsiness, Light Sleep, Deep Sleep, REM
Talking
Working on PC
Watching TV
And Many More!
Movements Associated with Ambulatory EEG
Consequence:
Too Many Motion Artifacts!
Challenge: Differentiate
between artifacts, epileptic
seizures and normal EEG activities!
How …!???
35. Dr. Md. Kafiul Islam
35
Summery of Existing Artifact Removal Methods
Single artifact type
Reference channel
Mostly general purpose
Often Manual or Semi-automatic
Often suitable for Multi channel
Real-time/Online processing capability
Not enough quantitative evaluation
Often after-effects not reported
Lack of adequate dataset used
Often hybrid methods (wICA, EEMD-CCA, etc.)
Presented By Md Kafiul Islam
36. EEG Signal Classification techniques
Dr. Md. Kafiul Islam
36
Supervised Classification
SVM
(Better for binary classification)
ANN
(Suitable for multi-classes)
Deep CNN
Preprocessing may be ignored
Feature Extraction may not needed
KNN
Un-Supervised
Clustering (K-means)
37. Domain
Time
Frequency
Wavelet
Statistical Features
Entropy
ApEn, MSE, CMSE, MMSE, etc.
Kurtosis
Skew
RMS/Variance/S.D.
Max/Min/Mean
Maxima/Minima
Features for Epileptic Seizure Detection
(Example of Feature Extraction)
High Frequency Oscillations
(HFO)
80 Hz – 200 Hz
Ripple
200 Hz – 600 Hz
Typical Scalp EEG!
0.05Hz – 128 Hz
Seizure Appears in Scalp EEG
0.5 Hz - 29 Hz [1,2]
1. J. Gotman, “Automatic recognition of epileptic seizures in the EEG,”
Electoencephalogr Clin. Neurophysiol., vol. 54, pp. 530–540, 1982
2. M. E. Saab and J. Gotman, “A system to detect the onset of epileptic
seizures in scalp EEG,” Clin. Neurophysiol., vol. 116, pp. 427–442,
2005.
38. EEG Features before and after Artifact Removal
Features Extracted:
(i) Entropy (ii) Kurtosis (iii) Line Length (iv) Peak
(v) NEO (vi) Variance (vii) FFT (viii) FFT Peak
Note: The features between seizure and non-seizure data are more separable after artifact removal which
suggests that it increases the detection rate and minimizes false alarms (false alarms are due to artifacts).
39. Dr. Md. Kafiul Islam
39
Effect of Artifact Removal from EEG (Epilepsy Diagnosis)
Presented By Md Kafiul Islam
False alarms improvement
40. Dr. Md. Kafiul Islam
40
BCI Performance ImprovementSNDR Improvement
Effect of Artifact Removal from EEG (BCI)
41. EEGLAB: Open Source MATLAB toolbox for processing continuous &
event-related EEG
Dr. Md. Kafiul Islam
41
EEGLAB Features
•Graphic user interface
•Multi-format data importing
•High-density data scrolling
•Interactive plotting functions
•Semi-automated artifact removal
•ICA & time/frequency transforms
•Event & channel location handling
•Forward/inverse head/source modeling
•Defined EEG & STUDY data structures
• Over 100 advanced plug-in/extensions
https://sccn.ucsd.edu/eeglab/index.php
44. Future Challenges & Recommendations
Challenges
Ambulatory Recording ->
motion artifacts
Real-time processing ->
computational complexity
On-chip on-the-Fly Decision
Making
Big Data storage and use for
future prediction (preventive
healthcare)
Dr. Md. Kafiul Islam
44
Recommendations
Application Specific Models
Artifact Specific
Standard performance evaluation
Ground truth data
Considering ML & AI (Deep CNN)