ArticlePDF Available

Real-time EEG processing based on Wavelet Transformation

Authors:
Thomas Malina
Thomas Malina
A. Folkers
A. Folkers
A. Folkers
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Ulrich G. Hofmann at University Medical Center Freiburg
Ulrich G. Hofmann
Download full-text PDF
Read full-text
Download citation

Abstract and Figures

We report on a novel data acquisition system, as part of a project to measure feedbackcoupled ERP signals, which uses wavelet transformation to decompose EEG signals in realtime in their respective energy bands. Due to constraints in DSP-based computing power, we have to settle for the less than optimal decomposition by short wavelets, but nevertheless achieve satisfactory discrimination power for clinical applications.
Figures - uploaded by Ulrich G. Hofmann
Author content
All figure content in this area was uploaded by Ulrich G. Hofmann
Content may be subject to copyright.
Content uploaded by Ulrich G. Hofmann
Author content
All content in this area was uploaded by Ulrich G. Hofmann on Dec 14, 2012
Content may be subject to copyright.
Real-time EEG processing based on Wavelet Transformation
T. Malina, A. Folkers and U.G. Hofmann
Institute for Signal Processing, Medical University of Lübeck, 23569 Lübeck, Germany
[malina, folkers, hofmann]@isip.mu-luebeck.de
Abstract: We report on a novel data acquisition
system, as part of a project to measure feedback-
coupled ERP signals, which uses wavelet
transformation to decompose EEG signals in real-
time in their respective energy bands. Due to
constraints in DSP-based computing power, we have
to settle for the less than optimal decomposition by
short wavelets, but nevertheless achieve satisfactory
discrimination power for clinical applications.
Keywords: Real-time wavelet transform, EEG band
discrimination, digital signal processor
Introduction
Long going research into the possibilities of brain-
computer-interfaces [2] for locked-in patients points
currently at the utilisation of evoked potentials instead
of slow brain waves to command input devices [1].
However, physiological research at our university [4],
strongly suggests that the macroscopic electric (i.e.
EEG) state of a subject's brain influences the responses
to defined stimuli, manifested in higher amplitudes of
evoked potential and thus influencing all control
attempts.
In order to investigate this evidence in realtime, we
built a data acquisition system able to record raw EEG
signals from up to 32 scalp electrodes from standard
amplifier configurations over a prolonged period of
time. The design is meant to evaluate the incoming
signals according to user defined criteria and in
succession trigger the stimulus to evoke an ERP by
either of auditory, visual or sensory stimuli [7]. Of
particular interest are EEG recordings in a Virtual
Reality car simulator, currently under construction,
where the visual stimulus consist of the break lights of
preceeding cars.
Figure 1: Sketch of low-cost experimental setup, to record
EEG/ERPs while driving in a virtual environment.
Materials and Methods
In order to satisfy the need for a cost-effective system,
we rely on a Windows-PC and Texas Instruments (TI,
Dallas, TX, USA) digital signal processors of type
C6701 to record and digitize signals, amplified by any
available, analog EEG-amplifier. A single DSP is
integrated by Innovative Integration, Inc. (II, Thousand
Oaks, CA) onto a M67 PCI-card, which carries in turn
two OMNIBUS cards (AD 16) with 16 analog/digital
converters each, without multiplexing.
Programming is done with TI’s “Code Composer
Studio” and Borland's "C++ Builder". The interrupt
driven, DSP-BIOS II based, raw data recording routines
utilize less than 5% of the DSP's computing power at a
sampling rate of 5 kSamples/sec. The A/D conversion
itself may run with up to 50 kSamples/sec.
The GUI-controlled program contains the following
display panels: The command panel contains the
command and editing functions necessary for each
experiment and known from a tape recorder: Run, Stop,
Record, Replay, Forward, Back. It furthermore toggles
the available online toolpanels and provides information
on the hardware status and opportunity to take notes,
which are all stored together with incoming data.
12th Nordic Baltic Conference on Biomedical Engineering and Medical Physics, Reykjavik, June 2002
A traditional multi-channel polygraph plot (blue-on-
white) displays scalp electrodes amplitudes and trigger
channels. A spectrogram window may be opened and
shows the Fourier transform of one arbitrary channel.
An array potential display illustrates the ongoing
activity on scalp electrodes by color-coding and
logarithmically interpolating their potentials. This panel
may be replaced by a 2D-current source density display
[6]. Another panel shows an average of one channel
based on a trigger signal on another channel (online
ERP-display). The most important panel is the energy
decomposition display, which utilizes an online
wavelet decomposition scheme proposed by [5] and
implemented with a fast lifting scheme described
elsewhere [3].
Figure 2: EEG-band decomposition panel. Left for a 20Hz,
right for a 5 Hz sinusoidal signal. From top to bottom are
γ, β α, θ, and δ components and the original signal over time
displayed. The bottom most bar graph shows the energy at any
given moment.
Numerical tests and simulations of algorithms were
performed under Matlab (The Mathworks Inc., Natick,
MA, USA) and then implemented in C on the DSP.
Results and Discussion
Despite our progress in implementing fast wavelet
transform (WT) algorithms, we have to pay attention to
the limited computing power and signal delays and thus
cannot use high-order wavelets like Daubechie's No 8
(db8), but instead have to settle for smaller ones like
Daubechie's No 3 (db3). Figure 3 illustrates the price in
discriminatory power coming with this limitation:
From top to bottom are the δ, θ, α, β and γ bands
sketched in rows. Each row contains (from top)
measured selectivity with a db3 and simulated
selectivity with the same db3 and the db8 wavelet.
Selectivity is gray scale coded as percentage of the
whole signal's energy. It is clearly visible, that the use of
db8 leads to a clearer discrimination of components
(sharper gray bars), however at higher computing costs
and signal delays.
Figure 3: Frequency discrimination achieved with WT (see
text for further description)
Although our online decomposition seems to work
satisfactory, a real feedback coupled ERP-experiment
with triggering a VR-stimulus on reaching a predefined
threshold e.g. in the γ-band, is still pending, but under
construction.
REFERENCES
[1] Bayliss, J., Ph.D. Thesis, 2001, University of
Rochester.
[2] Craelius, W., The Bionic Man: Restoring Mobility.
Science, 295(5557): (2002) 1018-1021.
[3] Folkers, A., F. Mösch, et al. Realtime bioelectrical
data acquisition and processing from 128 channels
utilizing the Wavelet-Transformation. submitted to
CNS*02 (2002). Chicago: Elsevier.
[4] Rahn, E. and E. Basar, Prestimulus EEG-activity
strongly influences the auditory evoked vertex response:
A new method for selective averaging. Intern. J.
Neurosc, 69: (1993) 207-220.
[5] Rosso, O.A., S. Blanco, et al., Wavelet entropy: a
new tool for analysis of short duration brain electrical
signals. J. Neurosci. Methods, 105(1): (2001) 65-75.
[6] Scherg, M., From EEG source localization to source
imaging. Acta Neurol. Scand., Supp. 152: (1994) 29-30.
[7] Windhorst, U. and H. Johansson, Modern
Techniques in Neuroscience Research. 1 ed. (Springer,
1999, Berlin)
... The popular alternative today is Wavelet Transform, because this decomposition method is well suited to non-stationary signals [25]. It was applied to extract band activity [27] and to analyze sleep spindles. In a study using spectral features derived from wavelet analysis and classified by a neural network, the tested accuracy was 77.6% [28]. ...
... Sensitivity wake = T P T P + F N = 74.2% (26) Specif icity wake = T N T N + F P = 94.3% (27) Negative and Positive Predictive Values are an alternative pair to sensitivity and specificity. ...
... 27: ROCs of MBA to select settings for differentiating Awake using the STD2 formulation. ...
View
... The amplitude, frequency, and characteristics of an EEG signal change from one state to the other, and with age. An EEG is decomposed mainly into 5 subbands: delta (0-4 Hz), occurring during deep sleep, during childhood, and in serious organic brain diseases; theta (4)(5)(6)(7)(8), occurring in childhood, during emotional stress; alpha (8)(9)(10)(11)(12)(13), occurring in a normal person in an awake state; beta (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30), affected by mental activity; and gamma (above 30 Hz) [1,2]. Instead of studying the complete EEG, EEG subband-related information yields more accurate information about the underlying neuronal activities [3]. ...
... DWT is used for the decomposition. It works as a filter to divide the signal into bands at each specific level known as detailed and approximate components [17,18]. The output CD is known as the detailed component and output CA is known as the approximate component. ...
Statistical features based comparision of analysis and synthesis of normal and epileptic electroencephalogram for various wavelets
Article
Full-text available
  • Jun 2016
Background: Electroencephalogram (EEG) is an electrical signal in micro volts captured non-invasively from brain, which provides important and unique information about the brain. Frequency of EEG signal lies between 0-100 Hz. Decomposition of EEG signal into various bands such as alpha, beta, delta, theta and gamma is essential in seizure related studies. EEG plays a key role in diagnosis of epileptic seizures and neurological disorders. In this paper multiple wavelets families for decomposition and reconstruction are explored and are compared based on risk functions and reconstruction measures. Aim: While dealing with the wavelets it’s a difficult task to choose the correct/accurate wavelet for the given bio-signal analysis. Various statistical properties were studied by the authors to check the suitability of various wavelets for normal and diseased EEG signal decomposition and reconstruction. Material and Methods: The methodology was applied to three groups (63 subjects) consisting both the genders and aged between 01-80 years: 1) Normal healthy subjects 2) Patients with focal seizures 3) Patients with generalised seizures. Results and Conclusion: Our result shows that 'Haar' and 'Bior3.7' wavelets are more suitable for normal as well as diseased EEG signals, as mean square error (MSE), mean approximate error (MAE) and percent root mean square difference (PRMSD) of these wavelets is much lesser than other wavelets. Signal to error ratio (SER) for Haar and Bior3.7 was much higher than any other wavelet studied.
View
... Discrete wavelet transform (DWT) with various types of wavelets are used for the decompositionanalysis. It is a filter that divide the biosignal into bands at each level known as detailed component and approximate component [7,8]. ...
Qualitative Analysis of Brain and Heart Signals using DWT
Article
Full-text available
  • Nov 2019
Bio-signal processing is a widely carried out by the researchers for better understating of complex biological processes. Instead of studying complete biosignal; a decomposed biosignals gives better and more accurate information about the various dynamics involved in the process. Wavelet is one of the powerful transform which is applicable to time and frequency domain. While dealing with various types of physiological signals it is a tedious task to choose the correct or accurate wavelet for the given biosignal analysis. Electrical brain and heart signal along with peripheral pressure signal for 120 subjects is studied by the authors. Authors have checked various quality metrics decide suitability of various Wavelets for EEG, ECG, PP and PPG signal decomposition and reconstruction. The methodology was applied to normal as well as diseased subjects. Our results based on performance parameters like Mean Square Error, Mean Approximate Error, Signal to Error Ratio, PRMSD shows that orthogonal and biorthogonal wavelets are more suitable for bio-signal decomposition and reconstruction. This shows that selection of wavelet should not always be based on similarity between the mother wavelet and the nature of bio-signal.
View
... The signals have to be amplified (gain of one million) for processing the signal using any digital signal processing system. In order to analysis and to adopt EEG signals in real time, the researchers have developed a data acquisition system to record a raw EEG signals up to 32 scalp electrodes over a long period of time with aid of robust amplifier configurations for different stimuli like auditory, visual, and sensory (Malina et al., 2002). For real time system applications, these amplified signals are converted by the National Instruments (NI) 6052E analoguedigital converter card for reading the data in American Standard Code for Information Interchange (ASCII) format. ...
Development of EEG data acquisition device by using single board computer
Article
Full-text available
  • Jan 2013
Electroencephalogram (EEG) plays a vital role in several medical diagnosis (brain tumour detection, Alzheimer disease, epilepsy, etc.), engineering applications (emotion detection, drowsiness detection, stress assessment, etc.) and others. However, the cost of the EEG device is usually high (< $20,000) and it is still a challenging issue for many consumers. The cost of the EEG data acquisition device is purely depends on number of EEG channels, mode of signal transmission (wired/wireless), data resolution, software portability, etc. In recent years, the development of handy embedded systems opens a gateway to design and develop the data acquisition devices at a cheaper cost. In this work, we have used SBC TS 7800 (500 MHz, 128 MB RAM) for designing the EEG wired data acquisition device. This embedded system is used to acquire the data from the 32 EEG channels and to save the them in American Standard Code for Information Interchange (ASCII) values in either .txt or xls format for further research investigations. All the EEG channels are made up of Ag/Ag-Cl and have an impedance of 10 kΩ. This complete system is operating in Linux platform, and programs are developed using C and JAVA programming languages.
View
La Complejidad Neuronal. Su aplicación al problema del Alzhaimer
Book
Full-text available
  • Jul 2007
En el presente trabajo se realiza un breve estudio acerca del análisis de la electroencefalografía y neurofisiología del cerebro humano. Luego se procede a desarrollar la teoría correspondiente al procesamiento de señales, teniendo en cuenta las propiedades y características particulares de las señales electroencefalográficas (EEG). Además, debido a la compleja naturaleza de estas señaales, se utilizan los métodos de análisis no lineal para comprender eficientemente el funcionamiento y desarrollo del cerebro humano; de ahí la importancia que el trabajo le concede al estudio de una serie de técnicas no lineales de la complejidad, las cuales han tenido resultados relevantes para el caso de la enfermedad del Alzheimer. Una vez estudiadas dichas técnicas, se procedió a su programación en el Software MATLAB versión 7.0, para facilitar la interpretación de las mismas por los especialistas clínicos. In the present work a brief study of the analysis of the electroencephalography and neurophysiology of the human brain is carried out. The corresponding theory of the signal processing is then developed, taking into account the properties and particular characteristics of the EEG signals. Furthermore, due to the complex nature of these signals, to comprehend eficiently the working and development of the human brain the use of non-linear analysis methods is shown; therefore the importance brought to the study of several complexity non-linear techniques which has obtained relevant results applied to the research of the Alzheimer disease. Once studied those techniques, it is then programmed by means of the MATLAB 7.0, to easy the their interpretation by clinical specialists.
View
Pervasive real-time biomedical monitoring system
Conference Paper
  • Jan 2009
With the tremendous advancements in low cost, power efficient hardware and the recent interest in biomedical embedded systems, numerous traditional biomedical systems can be replaced with smaller embedded systems that do real-time analysis to provide bio-feedback to the users. This paper presents a prototype of an embedded system which is capable of real-time data collection, using analog and digital sensors and processing, to compute physiological variables and metrics. These metrics in turn can be used to determine information about the user's general well being. The sensors provide motion, brain wave activity (EEG) and blood oxygenation (SpO2) information. The system presented automatically computes the application specific metrics and indicates the results of the detection scheme to the user and to a monitoring base station. The metrics being used have been validated using raw data from patients suffering epileptic seizures and from past research. The paper also deals with application scenarios for such systems and architecture for an FPGA based implementation is discussed.
View
Pervasive real-time biomedical monitoring system
Article
  • Dec 2009
With the tremendous advancements in low cost, power efficient hardware and the recent interest in biomedical embedded systems, numerous traditional biomedical systems can be replaced with smaller embedded systems that do real-time analysis to provide bio-feedback to the users. This paper presents a prototype of an embedded system which is capable of real-time data collection, using analog and digital sensors and processing, to compute physiological variables and metrics. These metrics in turn can be used to determine information about the user's general well being. The sensors provide motion, brain wave activity (EEG) and blood oxygenation (SpO2) information. The system presented automatically computes the application specific metrics and indicates the results of the detection scheme to the user and to a monitoring base station. The metrics being used have been validated using raw data from patients suffering epileptic seizures and from past research. The paper also deals with application scenarios for such systems and architecture for an FPGA based implementation is discussed.
View
Helmet Integrated Nanosensors, Signal Processing and Wireless Real Time Data Communication for Monitoring Blast Exposure to Battlefield Personnel
Article
Full-text available
DISTRIBUTION STATEMENT: Approved for public release; distribution unlimited The views, opinions and/or findings contained in this report are those of the author(s) and should not be construed as an official Department of the Army position, policy or decision unless so designated by other documentation.
View
Recognition of cognitive states in EEG signals.
Article
Full-text available
EEG signal processing allows interpreting, analyzing, studying, researching and experiencing the brain electrical activity in response to different cognitive processes, effects of drugs or drugs, the study of neurological or psychiatric diseases, among others. This thesis is oriented to the recognition of brain patterns to classify cognitive states using the EEG signals recorded from subjects performing mental tasks. Also, we expect to collect as many patterns as possible for each cognitive status and to seek parameters that provide more information, examine the major bands and all brain electrodes available in the database. The methodology used includes the registration of five cognitive tasks analyzed with three different approaches: analysis of long-range temporal-correlations with the Detrended Fluctuations Analysis (DFA) algorithm, the power analysis of brain signals using the Wavelet Transform and finally the study of phased looked brain using the Hilbert transform. The approaches used for this research report excellent results for differentiating the cognitive tasks studied, showing that the use of information from all the electrodes and their main brain bands contribute positively. Also, one can recognize and identify which parameters produce more information for this analysis. O processamento de sinais EEG permite interpretar, analisar, estudar, pesquisar e experimentar a atividade elétrica do cérebro como resposta para diferentes processos cognitivos, efeitos de drogas ou fármacos, estudo de doenças psiquiátricas ou neurológicas, entre outras. Esta dissertação é orientada ao reconhecimento de padrões cerebrais que permitam classificar estados cognitivos mediante os sinais de EEG registrados em sujeitos realizando tarefas programadas. Ademais espera-se obter a maior quantidade de padrões para cada estado cognitivo e procurar os parâmetros que oferecem maior informação, analisando as principais bandas cerebrais e todos os eletrodos disponíveis na base de dados. A metodologia usada compreende o registro de cinco tarefas cognitivas analisadas com três abordagens diferentes: análises de longe-range tenporal correlations com o algoritmo de Detrended Fluctuations Analysis (DFA), análise da potência dos sinais cerebrais utilizando a Transformada Ondeleta e finalmente o estudo da sincronia cerebral usando a Transformada de Hilbert. Conclui-se que as abordagens utilizadas nesta dissertação reportam alentadores resultados para diferenciar as tarefas cognitivas estudadas, demonstrando que a utilização da informação de todos os eletrodos e de suas principais bandas cerebrais contribuem de forma positiva. Também se consegue reconhecer e identificar quais parâmetros produzem maior informação para esta análise.
View
Reliable health monitoring: a commercial off-the-shelf and a field programmable hardware approach
Article
  • May 2010
With the tremendous advancements in low cost, power-efficient hardware and the recent interest in biomedical embedded systems, numerous traditional biomedical systems can be replaced with smaller and faster embedded systems that perform real-time analysis to provide bio-feedback to the users. This thesis takes a look at two hardware implementations – one using commercial off-the-shelf (COTS) components and the other using field programmable logic. The focus of the design was to ensure a portable, inexpensive, power-efficient and robust device that could perform analysis of physiological signals, which would in turn help alert the user in the event of an abnormality. The COTS hardware implementation provided the framework using a microcontroller as the processing element for a reliable health monitoring device with a seizure detection directly embedded in it. The field programmable gate array (FPGA) platform based implementation was proposed and simulated to overcome the two disadvantages of the COTS approach – the inability to support customization of the device to suit the end-user’s monitoring requirements and complex detection schemes requiring significant processing capability. The FPGA platform was simulated first as a standalone module and later as part of an SoC design. The novel algorithm included a feature extraction phase and a machine learning based seizure detection phase. Simulation based testing of the device showed a detection accuracy of 99.2 %.
View
Wavelet entropy: A new tool for analysis of short duration brain electrical signals
Article
Full-text available
  • Jan 2001
Since traditional electrical brain signal analysis is mostly qualitative, the development of new quantitative methods is crucial for restricting the subjectivity in the study of brain signals. These methods are particularly fruitful when they are strongly correlated with intuitive physical concepts that allow a better understanding of brain dynamics. Here, new method based on orthogonal discrete wavelet transform (ODWT) is applied. It takes as a basic element the ODWT of the EEG signal, and defines the relative wavelet energy, the wavelet entropy (WE) and the relative wavelet entropy (RWE). The relative wavelet energy provides information about the relative energy associated with different frequency bands present in the EEG and their corresponding degree of importance. The WE carries information about the degree of order/disorder associated with a multi-frequency signal response, and the RWE measures the degree of similarity between different segments of the signal. In addition, the time evolution of the WE is calculated to give information about the dynamics in the EEG records. Within this framework, the major objective of the present work was to characterize in a quantitative way functional dynamics of order/disorder microstates in short duration EEG signals. For that aim, spontaneous EEG signals under different physiological conditions were analyzed. Further, specific quantifiers were derived to characterize how stimulus affects electrical events in terms of frequency synchronization (tuning) in the event related potentials.
View
Wavelet entropy: A new tool for analysis of short duration brain electrical signals
Article
Full-text available
  • Feb 2001
  • J NEUROSCI METH
Since traditional electrical brain signal analysis is mostly qualitative, the development of new quantitative methods is crucial for restricting the subjectivity in the study of brain signals. These methods are particularly fruitful when they are strongly correlated with intuitive physical concepts that allow a better understanding of brain dynamics. Here, new method based on orthogonal discrete wavelet transform (ODWT) is applied. It takes as a basic element the ODWT of the EEG signal, and defines the relative wavelet energy, the wavelet entropy (WE) and the relative wavelet entropy (RWE). The relative wavelet energy provides information about the relative energy associated with different frequency bands present in the EEG and their corresponding degree of importance. The WE carries information about the degree of order/disorder associated with a multi-frequency signal response, and the RWE measures the degree of similarity between different segments of the signal. In addition, the time evolution of the WE is calculated to give information about the dynamics in the EEG records. Within this framework, the major objective of the present work was to characterize in a quantitative way functional dynamics of order/disorder microstates in short duration EEG signals. For that aim, spontaneous EEG signals under different physiological conditions were analyzed. Further, specific quantifiers were derived to characterize how stimulus affects electrical events in terms of frequency synchronization (tuning) in the event related potentials.
View
The Bionic Man: Restoring Mobility
Article
Full-text available
  • Mar 2002
  • SCIENCE
Bionics engineers are making increasingly bold and successful use of their tools to restore mobility to persons with missing or nonfunctional limbs. These tools include the latest materials, minielectronics and megacomputers, advanced robotic mechanisms, and algorithms. With crucial help from their pioneering users, they are learning how and where the residual sensorimotor system can be tapped in order to transmit its intents to replacement or reactivated body parts.
View
Realtime bioelectrical data acquisition and processing from 128 channels utilizing the wavelet-transformation
Article
  • Jun 2003
  • NEUROCOMPUTING
We propose a versatile signal processing and analysis framework for bioelectrical data, and in 9 particular for neural recordings and EEG. Within this framework the signal is decomposed into subbands using fast wavelet transform algorithms, executed in real-time on a current digital sig-11 nal processors hardware platform. The decomposition is used to perform various processing and analysis tasks. Besides fast implementation of high, band, low pass ÿlters, the decomposition is 13 used for denoising and lossy, as well as lossless compression. Furthermore speciÿc electrophys-iologic analysis tasks like spike detection and sorting are performed within this decomposition 15 scheme. c 2002 Published by Elsevier Science B.V. 1. Overview 19 Recording neural activity from a high number of neurons is a key issue in understand-ing how the brain works. Within the project VSAMUEL we developed successfully 21 a versatile data acquisition system based on DSP boards [3]. The system is used for continuous neural data acquisiton in vivo or in vitro with a high channel count (up 23 to 128 channels) at sampling rate F = 50 kHz with a precision of 16 bits per sample.
View
Prestimulus Eeg-Activity Strongly Influences the Auditory Evoked Vertex Response: A New Method for Selective Averaging
Article
  • Jul 1993
A new algorithm for averaging of evoked potentials (EPs) is introduced and compared with the conventional method. The root mean square (rms) values of the alpha and theta band EEG activities 1-s preceding stimulus presentation were computed: stimuli were blocked during phases of high prestimulus activity, thus evoking auditory responses selectively during periods of low activity. Ten volunteers were stimulated with tone bursts of 2000 Hz, 80 dB and 1-s duration. The main results were: (1) The averaged EPs showed an increase in amplitude of nearly 40% in comparison to experiments with conventional stimulation. This increase was in the range of 30% when an ISI correction was taken into account. (2) Single trials tended to form patterns of damped oscillations of e.g., 7-8 Hz, thus revealing a high correlation to the averaged EP. In alpha or theta band contingent experiments, correlation coefficients between single trials and averaged EPs were significantly higher (p < .01; p < .05).
View
From EEG source localization to source imaging
Article
  • Feb 1994
  • Acta Neurol Scand Suppl
A new functional imaging technique, "FOCUS", has been developed to transform the traditional scalp EEG into an image of source activities. The image is based on multiple spatio-temporal dipole models and consists of gross spatial patterns and source waveforms reflecting the estimated activities of the different brain regions. The application of the 'FOCUS' technique to the EEG in temporal lobe epilepsy revealed the presence of different activities at the basal and lateral aspects of the temporal lobe. The source waveforms showed propagation patterns consistent with subdural recordings which were not recognizable in the scalp EEG.
View
M o d e r n Techniques in Neuroscience Research
  • Jan 1999
  • U Windhorst
  • H Johansson
Windhorst, U. and H. Johansson, M o d e r n Techniques in Neuroscience Research. 1 ed. (Springer, 1999, Berlin)