Electroencephalography II Laboratory
Contents
1. 0 0 200 0 200 0 Aini I A ny uy A i y Aha MA Samp le Rate Hz ON i Wapa w l NAW age y Ww Yon d A Wy Pia hy y re 200 0 0 0 4A i yl i iT wy Fai Wy Wine hes W BioRadio oaa af M i yy i Information A j n F Ny l Y A i i M Awl A co A WAT 193 hy l Ke V pow La WN y Y 4 2 91 Time Seconds 6 Save approximately 30 seconds of data to a file with the subject sitting relaxed and fe looking at the screen Name the file EEGrelaxed Now save approximately 30 seconds of data to a file with the subject sitting relaxed and looking at the screen However the subject should make a few rapid blinks approximately every few seconds during the data recording Name this file EEGblinks 2006 Cleveland Medical Devices Inc Cleveland OH 7 Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0 Clevel_abs E Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Electroencephalography II Laboratory 8 Now save approximately 30 seconds of data to a file with the subject sitting relaxed and looking at the screen However the subject should shake their head a few times during the data recording Name this file EEGmotion 9 Now click on the tab labeled Sample EEG From the drop down menu select each type of EEG signal from the2. OH 11 Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0
3. This allows someone to discover what frequencies are occurring at particular points of time instead of just examining frequency components of the signal as a whole This method is particularly useful for non stationary signals 7 Discuss the tradeoffs between using each type of JFTA method to analyze a biomedical signal Are there particular reasons to use a specific type of method Each algorithm has a unique impact on the time and frequency resolutions that can be achieved cross term interference and computational speed The Gabor algorithm expands the signal into a set of weighted frequency modulated Gaussian functions and is the inverse of the STFT It has good time and frequency resolution and the speed of the algorithm is moderate The Wigner Ville distribution gives high time and frequency resolution but suffers from cross term interference This algorithm is computationally fast and does not suffer from windowing effects The Cohen s class algorithms such as the Choi Willimas and Cone shaped distributions act to reduce this crossterm interference Both algorithms provide good time and frequency response but computationally they are very slow The appropriate method to use is based on the particular application at hand 8 What effect does the window size have on a JTFA plot using the STFT method What impact does it have on the resolution of the plot The size of the discrete intervals determines the time accuracy In other words
4. petit mal attack can initiate a grand mal attack Focal epilepsy usually results from localized organic lesions or abnormalities such as scar tissue These abnormalities can cause rapid neuronal discharge that spread to adjacent regions in a wave like fashion When the wave reaches the motor cortex the person experiences Jacksonian epilepsy where muscles contract progressively down the body Multi channel EEG recordings can be used to identify the location of the earliest component of the seizure activity Such a trigger zone may be removable by surgery to eliminate the seizure disorder Researchers are developing electronic devices to automatically detect these abnormal bursts or spikes of electrical activity in the brain If the onset of a seizure could be detected early enough then it may be possible to intervene by administering drugs or electrical stimulation to the brain before an attack occurs A problem with this technique however is that these rhythmic spikes occur quite often in normal EEG s as well 2006 Cleveland Medical Devices Inc Cleveland OH 2 Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0 Clevel_abs E amp 8 Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Electroencephalography II Laboratory Seizure Activity Certain patterns in the EEG signal indicate seizure activity These patterns include spikes sh
5. the smaller the discrete block of time is the better the time resolution However there is a tradeoff to this method The frequency resolution is inversely proportional to the time resolution This is known as the window effect The smaller the discrete interval of time the less resolution of frequency is provided 2006 Cleveland Medical Devices Inc Cleveland OH 10 Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0 Clevel_abs E amp 8 Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Electroencephalography II Laboratory References l Guyton and Hall Textbook of Medical Physiology g h Edition Saunders Philadelphia 1996 Kandel ER Schwartz JH Jessel TM Essentials of Neuroscience and Behavior 1995 Kooi Kenneth A Fundamentals of Electroencephalography Harper amp Row Hagerstown MD 2nd Ed 1978 National Instruments Corporation Signal Processing Toolset User s Manual December 2002 Edition Goldberger AL Amaral LAN Glass L Hausdorff JM Ivanov PCh Mark RG Mietus JE Moody GB Peng CK Stanley HE PhysioBank PhysioToolkit and PhysioNet Components of a New Research Resource for Complex Physiologic Signals Circulation 101 23 e215 e220 Circulation Electronic Pages http circ ahajournals org cgi content full 101 23 e215 2000 June 13 2006 Cleveland Medical Devices Inc Cleveland
6. NCSE exists However in general episodes must be continuous or recurrent for longer than 30 minutes without improvement in the patient s clinical state Visual Evoked Potentials While it is not a clinical disorder EEG patterns can also be created by visual stimuli Visual evoked potentials are the components of the EEG that arise in response to a visual stimulus EEG is recorded while patients are subjected to visual stimuli Visual evoked can be used to examine the visual pathway Abnormalities can indicate lesions of the optical nerve or multiple sclerosis Electroencephalography Analysis Methods As explained above and as you will see in lab course software examples there are different types of EEG patterns that can be produced by the brain based on electrode location environmental inputs and clinical conditions One method for diagnosing abnormalities is through a visual inspection of the EEG signals by a trained clinician However a trained physician is not always readily available to review these recordings Therefore it may be advantageous to use a 2006 Cleveland Medical Devices Inc Cleveland OH 3 Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0 Clevel_abs E 8 Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Electroencephalography II Laboratory computer based program to quantitatively assess the EEG signa
7. OH 4 Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0 ClevelLabs E amp 8 Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Electroencephalography II Laboratory response but computationally they are very slow The appropriate method to use is based on the particular application at hand Experimental Methods Experimental Setup During this laboratory session you will record four channels of EEG Two channels will be recorded from the frontal region and two channels will be recorded from the occipital region You should view the setup movie included with the CleveLabs software before beginning the experimental setup for this laboratory session l Figure 2 Two of the six gold cup electrodes will be placed at O1 and O2 to monitor EEG If possible the subject for this laboratory should be a person with shorter hair The subject also should have their scalp free of any types of hair gel You will need seven gold cup electrodes for this laboratory Gold cup electrodes will be placed at locations Ol O2 Fpl and Fp2 Figs2 4 to measure EEG on each mastoid as references and at FpZ middle of the forehead for the ground The mastoid processes A1 and A2 are the bony structures that you can feel behind the ears Fig 3 Before applying electrodes to the subject it is first important to properly prepare and clean the electr
8. View Top Down View OQ V Q amp FpZ Fo 1 O j pes Al lt Jumper p S o1 A2 O2 BioRadio CleveMed Figure 4 Laboratory EEGI Setup 2006 Cleveland Medical Devices Inc Cleveland OH Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0 Clevel_abs E 8 Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Electroencephalography II Laboratory Procedure and Data Collection Run the CleveLabs Course software Log in and select the Electroencephalography IT laboratory session under the Advanced Physiology subheading and click on the Begin Lab button Turn the BioRadio ON Your BioRadio should be programmed to the LabEEGII configuration Click on the Raw EEG data Tab and then on the green Start button You should begin to see four channels of EEG scrolling across the screen Set the time scale to show a 2 second window of data Click on Screen Capture to capture a picture of this to your report This may not look like EEG yet because you are not filtering out high frequency noise that may be contaminating the signal such as 60Hz noise gt EEG Il Laboratory Electroencephalography II Laboratory Data Collection Interval ms 100 i Background Anatomy Setup Sample EEG Raw EEG Data Spectral Analysis Time Frequency Analysis BioRadio Raw EEG Data
9. Clevel_abs 1 4 Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Klectroencephalography II Laboratory 2006 Cleveland Medical Devices Inc Cleveland OH Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0 Clevelabs E amp Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Electroencephalography H Laboratory Introduction Several neurological disorders exist that can have an impact on brain function Often these disorders can be examined by reviewing the electroencephalograph or EEG signal Quantitative features of the EEG signal in both the temporal and spectral domain can be used for diagnosis One important clinical disorder that can be diagnosed using the EEG signal is seizures Seizures and status epilepticus can be potentially life threatening conditions For convulsive seizures visual observation is usually used in the emergency department setting to establish the diagnosis However non convulsive seizures NCS may produce altered mental state or behavior with only subtle ambiguous or absent motor components Typically the a clinical features include coma psychomotor retardation drowsiness inattentiveness aphasia limb posturing abnormal eye movements cortical blindness automatisms delusions paranoia and agitation In unresponsive patients physicians typically l
10. arp waves and spike and wave discharges Spikes and sharp waves occur during seizures resulting from partial epilepsy Partial epilepsy means that there is increased neuronal activity in only part of the brain Generalized epilepsy on the other hand is indicated by spike and wave discharges These discharges are seen throughout the brain Grand Mal Grand mal epilepsy is a condition that is a type of generalized epilepsy where there is extreme neural activity in all parts of the brain These discharges originate in the brainstem portion of the reticular activating system RAS A discharge transmitted to the spinal cord will result in a tonic seizure The EEG during grand mal attack shows a signal that is synchronous and that has a high voltage magnitude This signal has the same periodicity as normal alpha waves Grand mal attacks have been shown to be brought on by hyperexcitability of the neurons that make up the RAS structure or by abnormality in the local neuronal pathways Non Convulsive Seizures As described above non convulsive seizures can be particularly dangerous since little to no clinical observations can be made to diagnose the condition The abnormal electrical activity of the brain can exist and cause damage before a diagnosis is made Even more severe than NCS Non Convulsive Status Epilepticus NCSE is also not only prevalent but also more dangerous due to its longer continuous duration No universally accepted definition for
11. file Select a file from the drop down box and click on the Go button This will produce a JTFA in the window 11 Try using each of the sample data files and different JTFA methods and examine the effects on the results gt EEG II Laboratory Electroencephalography II Laboratory f z A o san M ooo Soreewcapture mAn men areena groo 5 f S Background Anatomy Setup Sample EEG Raw EEG Data Spectral Analysis Time Frequency Analysis nee EEG Samples Seizure Activity Power Spectrum Spectrogram fe 3 Samp Freq 174 Trend Level J 0 00 TURE EEG 1 1 0 0 0 ITEAN sth A Gabor spectrogram Freq Bins lt r Time Waveform 3 256 z 1 0E 3 ren EA E reani aa as a 16450 1 0E 3 oom Ca esel Plot x 2 Window 1 8E 3 i i i 0 000 5 000 10 000 15 000 20 000 23 593 2006 Cleveland Medical Devices Inc Cleveland OH Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0 ClevelLabs E amp 8 Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Electroencephalography II Laboratory Data Analysis 1 Using the post processing toolbox MATLAB or LabVIEW open the saved data files EEGrelaxed and grandmal Examine the temporal and spectral features of the signal and describe how you would generate an automated program for detecting seizure activi
12. l and classify it as a particular clinical state Several tools are available to provide quantitative analysis of temporal and spectral components of the EEG signal In addition this laboratory session will introduce a joint time frequency analysis JFTA This method allows users to analyze signals in both the time and frequency domains simultaneously Fig 4 It is a useful tool to analyze non stationary signals Essentially this method yields which frequencies are occurring at what times A graphical representation illustrates how the signal power spectrum changes over time The basic approach 1s to divide the signal into several discrete intervals that can be overlapped A Fourier transform can then be applied to each block of data to illustrate the frequency components of each block The size of the discrete intervals determines the time accuracy In other words the smaller the discrete blocks of time is the better the time resolution However there is a tradeoff to this method The frequency resolution is inversely proportional to the time resolution This is known as the window effect The smaller the discrete interval of time the less resolution of frequency is provided The algorithm described above is known as a short time Fourier transform STFT b gt EEG II Laboratory Electroencephalography II Laborato ee SAVE DATA SCREEN EEE EENE ER Data Collection Interval ms 100 Background Anatomy Setup Sample EEG Raw EEG Data S
13. ly 8 10 D of the emergency room visits involve acute cerebral dysfunction or injury or 8 10 million people Utilizing EEG recordings in the emergency room may improve clinical outcomes in a large patient population One reason for EEG screenings not being a standard of care includes the lack of appropriate EEG monitoring technology Epilepsy As stated above an important application of the EEG is to diagnose neurological disorders Epilepsy is a disease of the nervous system that is characterized by uncontrolled activity of part or all of the central nervous system There are different kinds of epilepsy such as grand mal petit mal and focal epilepsy Epileptic seizures are characterized by usually synchronized neural activity resulting in abnormally large amplitude EEG activity during an attack Grand mal epilepsy involves neuronal discharges in all areas of the brain including the cortex the deeper parts of the cerebrum and even the brain stem and thalamus Generally grand mal attacks last anywhere from a few seconds to a few minutes During grand mal the individual may lose control of visceral and motor control Petit mal involves the basic thalamocortical brain activating system Often referred to as absence epilepsy an individual suffering from a petit mal attack will lose consciousness for up to thirty seconds experience several twitch like contractions in the head and neck regions and then resume previous activities Occasionally a
14. nset suggest what factors an automated program should look for when doing automatic EEG analysis How would noise be dealt with in this automated program An automated program can perform an FFT on windowed sections of the data and find the energy in the various EEG bands When the energy in a lower frequency band overcomes some threshold then it can classify that as a seizure Or if the seizure waveforms are already known a 2006 Cleveland Medical Devices Inc Cleveland OH 9 Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0 Clevel_abs E amp Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Electroencephalography I Laboratory cross correlation method can be used to match a template with the current brain activity However the disadvantage to this is that the template may not be the same for all patients Alternatively spike detection can be done by looking for sharp transients either by wavelet decomposition or a derivative method looking at the differences in samples In order to remove noise a bandpass FIR filter can be constructed that will pass from 5 22 Hz eliminating 60 Hz noise and any DC offset from EOG 6 What are the benefits of using a JFTA method to analyze data compared to only using a spectral or temporal analysis technique Using a JFTA will illustrate how a power spectrum of a signal changes over time
15. ode sites If needed refer to previous labs for information on cleaning and preparing electrode sites Now the gold cup electrodes can be attached Generously fill a gold cup electrode with the provided conductive gel allowing some gel to fill over the top of the cup Squeeze some of the conductive gel onto a gauze pad as well Push the electrode into the gel on the gauze pad and then gently push aside the hair and place the electrode on the back of the subject s head at position Ol Repeat for the other gold cup electrode at position O2 The electrode placements for O1 and O2 are illustrated in Fig 2 2006 Cleveland Medical Devices Inc Cleveland OH 5 Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0 Clevel_abs E amp Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Electroencephalography II Laboratory x i Figure 3 The mastoid processes are the bony L structures located behind the ears Mastoid Process 3 Connect gold cup leads and jumpers to channel inputs 1 2 3 4 and the ground using the picture below as a reference Fig 4 The left side view of the head is symmetrical to the right side view as seen in the top down view UJ 5 Fz lt lt p ae ae FpZ c3 PZ FPD FP2 gt s F7 p3 Fz pa FF O O ee o ee a eo ee A oO l P P3 pz P4 O KOR Left Side
16. ook for ocular deviation or nystagmoid eye movements periodic facial or eyelid twitching or unexplained fluctuation in responsiveness However by definition there are no convulsions including twitching in a true non convulsive seizure Thus clinicians are left with only a vague hint that a seizure may be occurring but not a diagnosis The only way to definitively diagnose NCS is by EEG recordings rather than by clinical observation During this laboratory session students will record EEG from the surface of their scalp They will collect four channels of their own normal EEG signals and compare it to EEG signals in an abnormal clinical database Students will examine quantitative features of the EEG signal that can be useful in the diagnosis of neurological disorders Equipment required e CleveLabs Kit CleveLabs Course Software Seven 7 Gold cup electrodes Conductive Gel Gauze and skin preparation Microsoft Excel MATLAB or LabVIEW 2006 Cleveland Medical Devices Inc Cleveland OH 1 Property of Cleveland Medical Devices Copying and distribution prohibited CleveLabs Laboratory Course System Version 5 0 Clevel_abs E amp Laboratory Course System CleveLabs Laboratory Course System Teacher Edition Electroencephalography II Laboratory Background Clinical Disorders and EEG Responses q 4 P There are approximately 100 million visits to emergency rooms every year in the United States Of these approximate
17. pectral Analysis Time Frequency Analysis EEG Samples Grand Mal T 5 4E 0 meih a nS 4 2E 1 JTFA Method 6 2E 1 3 Gabor spectrogram 1 6E 1 Trend Level 9 0 00 e AOE i 1 0 2 Samp Freq 102 d CO mrF Freq Bins g 258 Cursor LD Zoom CC Reset 1 1 1 Plot 50 000 100 000 150 000 179 990 Window Time Var 3 164 50 f gt lt Figure 1 A joint time frequency analysis can be useful for analyzing many physiological signals Other algorithms that can be used in the JFTA method include an adaptive spectrogram a Choi Williams distribution a Cone shaped distribution Wigner Ville distribution and the Gabor spectrum Each algorithm has a unique impact on the time and frequency resolutions that can be achieved cross term interference and computational speed The Gabor algorithm expands the signal into a set of weighted frequency modulated Gaussian functions and is the inverse of the STFT It has good time and frequency resolution and the speed of the algorithm is moderate The Wigner Ville distribution gives high time and frequency resolution but suffers from cross term interference This algorithm is computationally fast and does not suffer from windowing effects The Cohen s class algorithms such as the Choi Willimas and Cone shaped distributions act to reduce this crossterm interference Both algorithms provide good time and frequency 2006 Cleveland Medical Devices Inc Cleveland
18. sample database and examine its temporal characteristics as it scrolls on the screen Each signal is approximately one minute long and will repeat when it finishes You should save each of these signals to a data files with the corresponding name for later analysis gt EEG Il Laboratory Electroencephalography II Laboratory ET MENU j Data Collection Interval ms Jio Background Anatomy Setup Sample EEG Raw EEG Data L Spectral Analysis _Time Frequency oy arene Sample EEG Waveforms EEG Samples Seizure Activity 1000 0 800 0 600 0 400 0 a 200 0 B 0S 200 0 400 0 600 0 800 0 a T 19 o0 it cimical Description ial El peu ig froi subjec te ufferi ing fro ons es in the hippocampal for This recordin ngwas s taken from the epe Epto genic zone of he icone campa alfo mato ie sh Aby se res during activity The total dur tu of the tri alis 23 65 onds Certai n pa ofa in the EG aaa al indica pa vity These pal aries include Seike shan arp waves a and ake and ae e discharges v Spikes sa sharp w sulti na from pa rtie al epilepsy Partial epilepsy means that th increased nen nal a on vay only part of the brai ns Gereral ae epilepsy 0 athe other indicated by spike and wave discharges These dischar Senora n througho ut the brai 10 Next click on the tab labeled Time Frequency Analysis This section will allow you to complete a joint time frequency analysis on the data
19. ty 2 Using the post processing toolbox MATLAB or LabVIEW open the save data files EEGrelaxed EEGblinks EEGmotion and Grandmal Analyze the signal in both the time and frequency domains Note how the intentionally generated artifacts in the subject generated data files compare to the seizure activity in the sample data files Discussion Questions 1 Why is it important to use EEG recordings to confirm non convulsive seizures Non convulsive seizures may not present clinical signs that can be observed to confirm a diagnosis such as the body shaking Therefore the abnormal electrical activity occurring in the brain needs to be confirmed by an EEG recording 2 Is it difficult to distinguish seizure activity from certain types of noise that can exist in an EEG recording If so why Noise such as motion artifact and blinking can create spike artifacts in the EEG signal This may mimic a spike that occurs in the EEG signal as a result of a seizure 3 Describe any quantitative features you found in the time domain that may be useful for quantifying specific cardiac rhythms Student responses will vary based on their observations 4 Describe any quantitative features you found in the frequency domain that may be useful for quantifying specific cardiac rhythms Student responses will vary based on their observations 5 Knowing that seizures generally involve slower brain rhythms and that spiking behavior usually precedes seizure o
Download Pdf Manuals
Related Search