<%BANNER%>

Asymmetries in EEG Signal Properties in Those with Temporal Lobe Epilepsy and Psychogenic Non-Epileptic Seizures

Permanent Link: http://ufdc.ufl.edu/UFE0044757/00001

Material Information

Title: Asymmetries in EEG Signal Properties in Those with Temporal Lobe Epilepsy and Psychogenic Non-Epileptic Seizures
Physical Description: 1 online resource (47 p.)
Language: english
Creator: Skinner, Holly J
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: asymmetry -- eeg -- epilepsy -- non-epileptic -- psychogenic -- seizures
Clinical Investigation (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Epileptic seizures can appear clinically similar to psychogenic non-epileptic events (PNEE), which can lead to erroneous diagnosis and treatment. Differentiation of these two conditions often requires multi-day, in-patient video electroencephalogram (VEEG) monitoring to record seizures or seizure-like PNEE. Also, brief runs of “epileptiform” activity may be recorded during the times when seizures are not occurring; these periods of time are referred to as the interictal periods. However, if seizures and/or interictal epileptiform abnormalities are not noted during the VEEG monitoring, then a diagnosis cannot be made. Therefore, we developed a hypothesis-driven approach to distinguish those with epileptic seizures from those with PNEE by quantitative analysis of brief epochs of electroencephalogram (EEG) in the interictal period, when no epileptiform activity was present on visual inspection. Our goal in this hypothesis testing study was to investigate whether differences in measures of EEG signal inter-hemisphere asymmetry (IHA) exist between patients with a common form of epilepsy, temporal lobe epilepsy (TLE), and patients with PNEE. Interictal EEG epochs (10 seconds each) were sampled from VEEG recordings obtained from 62 patients. A total of 620 epochs in the relaxed, awake state were collected from TLE and PNEE patient groups. Within each EEG sample epoch, we calculated the signal regularity using the pattern-match regularity statistic (PMRS), and amplitude variation (AV). These calculations were performed in the F8, T4, F7, and T3 EEG channels utilizing a non-overlapping 5.12 second computation window. IHA values were then calculated as the absolute difference between left (F7 and T3) and right (F8 and T4) channels, with respect to PMRS and AV values, respectively. We found that IHA of the PMRS from the temporal electrodes is significantly larger in patients with TLE than those with PNEE (p=0.0182). These results suggest measureable characteristics of the interictal EEG may be useful in distinguishing patients with TLE from those with PNEE.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Holly J Skinner.
Thesis: Thesis (M.S.)--University of Florida, 2012.
Local: Adviser: Stechmiller, Joyce K.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044757:00001

Permanent Link: http://ufdc.ufl.edu/UFE0044757/00001

Material Information

Title: Asymmetries in EEG Signal Properties in Those with Temporal Lobe Epilepsy and Psychogenic Non-Epileptic Seizures
Physical Description: 1 online resource (47 p.)
Language: english
Creator: Skinner, Holly J
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2012

Subjects

Subjects / Keywords: asymmetry -- eeg -- epilepsy -- non-epileptic -- psychogenic -- seizures
Clinical Investigation (IDP) -- Dissertations, Academic -- UF
Genre: Medical Sciences thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Epileptic seizures can appear clinically similar to psychogenic non-epileptic events (PNEE), which can lead to erroneous diagnosis and treatment. Differentiation of these two conditions often requires multi-day, in-patient video electroencephalogram (VEEG) monitoring to record seizures or seizure-like PNEE. Also, brief runs of “epileptiform” activity may be recorded during the times when seizures are not occurring; these periods of time are referred to as the interictal periods. However, if seizures and/or interictal epileptiform abnormalities are not noted during the VEEG monitoring, then a diagnosis cannot be made. Therefore, we developed a hypothesis-driven approach to distinguish those with epileptic seizures from those with PNEE by quantitative analysis of brief epochs of electroencephalogram (EEG) in the interictal period, when no epileptiform activity was present on visual inspection. Our goal in this hypothesis testing study was to investigate whether differences in measures of EEG signal inter-hemisphere asymmetry (IHA) exist between patients with a common form of epilepsy, temporal lobe epilepsy (TLE), and patients with PNEE. Interictal EEG epochs (10 seconds each) were sampled from VEEG recordings obtained from 62 patients. A total of 620 epochs in the relaxed, awake state were collected from TLE and PNEE patient groups. Within each EEG sample epoch, we calculated the signal regularity using the pattern-match regularity statistic (PMRS), and amplitude variation (AV). These calculations were performed in the F8, T4, F7, and T3 EEG channels utilizing a non-overlapping 5.12 second computation window. IHA values were then calculated as the absolute difference between left (F7 and T3) and right (F8 and T4) channels, with respect to PMRS and AV values, respectively. We found that IHA of the PMRS from the temporal electrodes is significantly larger in patients with TLE than those with PNEE (p=0.0182). These results suggest measureable characteristics of the interictal EEG may be useful in distinguishing patients with TLE from those with PNEE.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Holly J Skinner.
Thesis: Thesis (M.S.)--University of Florida, 2012.
Local: Adviser: Stechmiller, Joyce K.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2012
System ID: UFE0044757:00001


This item has the following downloads:


Full Text

PAGE 1

1 ASYMMETRIES IN EEG S IGNAL PROPERTIES IN THOSE WITH TEMPORAL LOBE EPILEPSY AND PSYCHOG ENIC NON EPILEPTIC SEIZURES By HOLLY S KINNER A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2012

PAGE 2

2 2012 Holly Skinner

PAGE 3

3 To my family and friends

PAGE 4

4 ACKNOWLEDGMENTS I would like to first thank my family and friends. My dad mom step mom, and grandmother have always encouraged me to learn at every opportunity. I thank my boyfriend, Dewayne and his family for welcoming me into their lives during these last few years while pursuing this degree. Next, I would like to thank all of my ment ors at the University of Florida (UF) and the Malcom Randall Veterans Affairs Medical Center (MR VAMC ) especially my primary mentor these last three years, J. Chris Sackellares MD He has taught me a great deal about EEG, epilepsy, research, a s well as, advised me on many aspects of my career. Also, I thank Stephan Eisenschenk MD, who was my clinical neurophysiology fellowship director and epilepsy division chief He has been supportive of me in pursuing this degree, and always willing to of fer his expertise to improve my research design s I am grateful to the team at Optima Neuroscience, and in particular Deng Shan Shiau PhD. Dr. Shiau has been instrumental to helping me collect and analyze the data for this project. Next, I would like to thank David FitzGerald MD, Stephen Nadeau MD, and Bruce Crosson PhD, all researchers affiliated with the Brain Rehabilitation Research Center at the MR VAMC. Each of these researchers helped me to apply for a VA career development award over the las t few years I am grateful for the experience in applying for external research funding. Also, I would like to sincerely thank my thesis committee chair, Joyce Stechmiller, PhD and committee member, Steven Roper MD. I greatly appreciate the

PAGE 5

5 time they devoted to th e defense, and ideas about alte rative conclusions to be drawn from the results Finally, I would like to thank the administrators, program coordinator Eve Johnson, and teaching faculty of the UF Advanced Post graduate Program for Clinical I nvestigation (APPCI) at UF. I gained and improved upon many research skills because of the opportunity to participate in this program. I plan to use these skills for years to come. Over the long term, I hope to make the UF APPCI and my mentors proud of t heir investment in my research future.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ............................. 9 ABSTRACT ................................ ................................ ................................ ................... 10 CHAPTER 1 AN INTRODUCTION TO ELECTROENCEPHALOGRAPHY ................................ 12 A Brief History of Electroencephalography ................................ ............................. 12 Indications for EEG ................................ ................................ ................................ 13 Technical Considerations in EEG Recording ................................ .......................... 15 Montages ................................ ................................ ................................ ................ 16 Neurophysiologic Basis of EEG Activity ................................ ................................ .. 17 Visual Analysis of EEG ................................ ................................ ........................... 17 Cellular Substrates of Brain Rhythms ................................ ................................ ..... 19 2 EPILEPSY AND NON EPILEPTIC EVENTS ................................ .......................... 25 Introduction to Epilepsy ................................ ................................ ........................... 25 Classification of Epilepsy ................................ ................................ .................. 26 Temporal Lobe Epilepsy ................................ ................................ ......................... 27 Seizur e Semiology in TLE ................................ ................................ ................ 28 Electroencephalographic Characteristics of TLE ................................ .............. 29 Psychogenic Non epileptic Events ................................ ................................ .......... 29 Clinical Manifestations of PNEE ................................ ................................ ....... 30 Etiologies of PNEE ................................ ................................ ........................... 31 Differentiating Epilepsy from PNEE ................................ ................................ ........ 31 3 MEASURING INTRAHEMISPHERIC EEG ASYMMETRY IN TLE AND PNEE ...... 33 4 STUDY DESIGN AND OUTCOME ................................ ................................ ......... 36 Me thods ................................ ................................ ................................ .................. 36 Results and Conclusion ................................ ................................ .......................... 38 LIST OF REFERENCES ................................ ................................ ............................... 44 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 47

PAGE 7

7 LIST OF TABLES Table page 1 1 Clinical indications for obtaining an EEG ................................ ............................ 20

PAGE 8

8 LIST OF FIGURES Figure page 1 1 Hans Berger ................................ ............................. 21 1 2 The Inte rnational 10 20 System for Scalp Electrode Placement. ........................ 22 1 3 Wave discharges. ................................ ................................ ............................... 23 1 4 Rhythmic and sharply contoured, 7 8 Hertz theta activity at the onset of a focal seizure localizing to the left temporal lobe. ................................ ................ 24 4 1 Bi variate ana lysis of diagnosis and gender ................................ ....................... 40 4 2 Bi variate analysis for diagnosis and age (in years). ................................ ......... 41 4 3 Distributions of the interhemispheric asymmetry. ................................ ............... 41 4 4 Bi variate analysis of the interhemispheric differences in the PMRS=pattern matched regularity statistic. ................................ ................................ ............... 42 4 5 Bi variate analysis of the interhemispheri c differences in the AV=amplitude variation. ................................ ................................ ................................ ............. 43

PAGE 9

9 LIST OF ABBREVIATION S ADC Analog to digital converter AED Anti epileptic drug ANOVA Analysis of variance AV Amplitude variation CPS Complex partial seizure DC Direct current EEG Electroencephalogram EMU Epilepsy monitoring unit Hz Hertz (cycles per second) ILAE International League Against Epilepsy IHA Interhemi spheric asymmetry IRB Institutional Review Board MR VAMC Malcom Randall Veterans Affairs Medical Center MUSC Medical University of South Carolina NTLE Neocortical temporal lobe epilepsy PLED Periodic Lateralized epileptiform discharges PNEE Psychogenic non epileptic events PMRS Pattern match regularity statistics REM Rapid eye movement (sleep) TLE Temporal lobe epilepsy UF University of Florida VEEG Video EEG

PAGE 10

10 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science ASYMMETRIES IN EEG S IGNAL PROPERTIES IN THOSE WITH TEMPORAL LOBE EPILEPSY AND PSYCHOG ENIC NON EPILEPTIC SEIZURES By Holly Skinner August 2012 Chair: Joyce Stechmiller Major: Medical Science Clinical and Translational Science Epileptic seizures can appear clinically simila r to psychogenic non epileptic events (PNEE), which can lead to erroneous diagnosis and treatment. Differentiation of these two conditions often r equires mul ti day in patient video electroencephalogram ( V EEG ) monitoring to record seizures or seizure like PNEE. Also, brief runs of may be recorded during the times when seizures are not occurring; these periods of time are referred to as the interictal periods. However if seizures and/or interictal epileptiform abnormalities are not noted during the V EEG monitoring, then a diagnos is cannot be made. Therefore, we developed a hypothesis driven approach to distinguish those with epileptic seizure s from those with PNEE by quantitative analysis of brief epochs of electroencephalogram ( EEG ) in the interictal period when no epileptiform activity was present on visual inspection Our goal in this hypothesis testing study was to investigate whether differences in measures of EEG signal inter hemisphere asymmetry (IHA) exist between patients with a common form of epilepsy, temporal lobe epi lepsy (TLE), and patients with PNEE. Interictal EEG epochs (10 seconds each) were sampled from V EEG recordings obtained from 62 patients. A total of 620 epochs in the relaxed, awake state were

PAGE 11

11 collected from TLE and PNEE patient groups. Within each EEG sample epoch, we calculated the signal regularity using the pattern mat ch regularity statistic (PMRS), and amplitude variation (AV). These calculations were performed in the F8, T4, F7, and T3 EEG channels utilizing a non overlapping 5.12 second computati on window. IHA values were then calculated as the absolute difference between left (F7 and T3) and right (F8 and T4) channels, with respect to PMRS and AV values, respectively. We found that IHA of the PMRS from the t emporal electrodes is significantly la rger in patients with TLE than those with PNEE (p=0.0182) These results suggest measureable characteristics of the interictal EEG may be useful in distinguishing patients with TLE from those with PNEE.

PAGE 12

12 CHAPTER 1 AN INTRODUCTION TO E LECTROENCEPHALOGRAPH Y A Brief History of Electroencephalography EEG is a neurophysiologic tool by which te mporal and spatial information about brain activity can be recorded EEG electrodes can record brainwaves non invasively by placing electrodes on the scalp or through invasive procedures brain activity can be recorded near or within brain tissue. However, scalp EEG recording s are the focus of the research performed in this thesis. Scalp EEG is widely used in many fields of neuroscience including neurology, psychology, sleep medicine, and neuroscience research. The first electrical brain activity was recorded in animals by English physician Richard Cato n (1842 1926) during the 1870s ( 1 ) This brain activity was recorded using a galvanometer with a beam of light cast onto a mirror to reflect a large scale on a wall. However, it was Austrian neuropsychiatrist Hans Berger (1873 1941) who was deemed the fa ther of encephalography.( 1 ) Dr. Berger was the first to record a single channel of electrical brain activity in humans, and published this recording in 1929. He first used a sting Galvanometer (Figure 1 1), and later a double coil Galvanometer.( 2 3 ) Over the next several decades, several technological advances improved the quality of EEG records. Researchers began to use oscilloscopes to observe waveforms in real time. The quality of cerebral waveforms captured was improved by the development of amplifiers and filters. Also, over time the number of EEG electrodes used to record brain activity increased. Beginning in the 1970s, mechanical apparatus used to capture and record EEG was replaced by computerized techniques.( 2 ) Until the 1990s, cente rs utilizing EEG

PAGE 13

13 relied on paper tracing and tape d video recording of EEG data. However, since that time, most centers have begun store data using digital means, which provides greater opportunity for manipulation of the data after recording. As such, so phisticated techniques for EEG analys i s have emerged.( 4 ) A standard system of electrode placement is essential for communication of EEG results between different laboratories. In 1958, Herbert Jasper proposed the 19 channel, International 10 20 system fo r electrode placement worldwide (Figure 1 2 ) Placement of electrodes in this system begins by distinguishing the sagittal anterior posterior distance between the nasion and inion, placing the first two electrodes at distances 10 percent above those bony landmarks, then placing the additional electrodes in specified locations which are at 10 to 20 percent distances from previously marked landmarks. The International 10 20 system is still in place today, and was the electrode placement used for the record ing of EEG data in this project. I ndications for EEG Around the time Hans Berger was recording the first human EEGs, his observations primarily focused on describing normal human physiology. For example, he noted that alpha waves arising from the occip ital region attenuated with eye opening. ( 1 5 ) As multi channel EEG emerged, alterations in wave morphology were noted in the area of brain lesions. As such, EEG served as a non invasive means of localizing foca l, pathologic processes in the brain. Since the arrival of neuro imaging techniques such a computed tomography and magnetic resonance imaging, EEG is not nearly so relied upon for its localizing capabilities. However, EEG remains a useful source of information that aids in diagnosis for many clinic al scenarios.

PAGE 14

14 Table 1 1 provides a full list of indications for EEG. However EEG is most often used to evaluate for the presence of epileptiform appearing activity. Epileptiform activity refers to paroxysmal, sharply contoured or rhythmic activity tha t may be seen in the setting of epilepsy. Furthermore, the EEG can help clinicians to classify the type of seizure disorder classification and localization are discusse d further in Chapter 2. Epileptiform activity may occur during the ictal or interictal period. The time during a seizure is referred to as the ictal period, and interictal period refers to the time in between seizures. During the interictal, brief epil eptiform activity, generally lasting approximately 0.5 second s to 3 seconds, may or may not be present. Interictal activity or sharps going slow wave, higher amplitude activity than the backgro und, and disruption of the background rhythm. ( 1,5 ) Examples of interictal epileptiform activity are displayed in Figure 1 3. Continuous ictal activity can be seen on an EEG throughout a seizure, and can last seconds to minutes, or occasionally even hours. This observation is true except in rare instances where seizure activity is confined to a small brain region that is not easily recorded with scalp EEG such as in seizure localized to the orbital frontal regions o r near the skull base Figure 1 4 illustrates ictal activity localizing to the left temporal lobe. However, for this research project, we specifically selected 10 second samples of EEG data where no interictal or ictal epileptiform activity was visualize d. This app roach was important because an EEG recording may fail to demonstrate ictal or interictal epileptiform activity even if a patient has e pilepsy. Increasing the time of the recording decreases the chances of missing epileptiform activity.( 6 )

PAGE 15

15 A s implied above, the d uration of an EEG recording can vary from 20 minutes to several days depending on the goals of the study. Because the timing of seizure activity is largely unpredictable, b aseline EEGs lasting 20 to 60 minutes generally capture the i nterictal period As such, t he baseline EEG is most useful for finding evidence of interictal epileptiform activity rather than ictal activity. While EEGs recordings can last from minutes t o hours to days, or even weeks, the goal of l onger recording s i s usually to capture seizures or ictal activity. To improve the chances of capturing seizures during a specific time, patients may be admitted into an epilepsy monitoring unit. ( 7,8 ) In these units, continuous EEG with video ( V EEG) is performed Also, pr ovocation techniques such as tapering seizure medications, sleep deprivation, flashing lights, and hyperventilation may be used to increase the likelihood that a patient will have a seizure. All of the EEG data for this study was obtained by analyzing rec orded V EEG data. This data was recorded in epilepsy monitoring units at the Medical University of South Carolina in Charleston, SC. Technical Considerations in EEG Recording The EEG recordings used for this project were digitally recorded EEG data acquisition begins when electrodes placed on the on the scalp after cleaning the skin to remove oils and applying an ionic solution at the electrode site. This preparation allows current to flow from the neurons, through human tissue and an elect rode wire, then into jack box This direct current (DC) signal is adjusted with filters and amplifiers Once EEG data has been captured, it is digitized by an analog to digit al converter (ADC). The ADC converts continuous information about EEG voltages into samples measured many times per second. In this study, a 256 cycles per seconds or hertz (Hz) sampling rate was used. Resolution of the EEG waveforms on a c omputer

PAGE 16

16 monitor depends on the amount of data stored in bytes and the computer monit or screen resolution. For this study, a 12 bit recording system was used. Montages Voltage refers to the electric tension or potential between two points, and how voltage changes over time is the basis of EEG recordings. Since each EEG electrode is me asuring current from a single brain region, a second comparison point must be used to measure the voltage between the two points. ( 1,5 ) The voltage between these two points is referred to a s a n EEG channel, and the two points in the channel may be adjacen t electrodes, distant electrodes, or even a ground. The configuration of how electrodes channels are viewed on paper or a computer monitor is referred to as a montage. Two main types of montages, the bipol ar montage and the referential montage, are used in EEG recordings. When bipolar montages are used, each electrode is compared to an adjacent electrode in a chain like fashion. Alternatively, referential montages produce EEG channels that compare each scalp elect rode to one or two r eferences. One advantage of using a referential montage is that all electrical amplitudes are compared to a single source. Hence, when looking at several channels, the channel with the largest amplitude waveform will be the source of that wave. ( 1 ) Also, since homologous electrodes on contralateral hemispheres (T8 on the right versus T7 on the left) will be equal distances from the reference, referential montages are ideal for assessing symmetry between the hemispheres. In our study, an average, referent ial montage was used. Voltage was measured by comparing the current in each electrode to the current at the mid way point between the Cz and Pz electrodes.

PAGE 17

17 Neurophysiologic Basis of EEG A ctivity EEG waveforms are produced by current generated by neur ons in the brain. These currents are produced by the flow of ions moving in and out of the extracellular space.( 5 ) Single cell EEG recordings have demonstrated that waveforms noted on EEG are due to post synaptic pote ntials which last 20 200 msec. ( 1 ) A n excitatory post synaptic potential (EPSP) occurs when positively charged sodium and calcium ions move into the intracellular space and the cell is depolarized. ESPS produce negative or upward deflected waveforms. Inhibitory post synaptic potentials (IP SP) occur when cells become hyperpolarized from potassium mov ing out of the cell. IPSP produce downward or positive deflections. Also, a n EEG best detects electric potentials that are a s hort distance from the scalp. As such activity from neurons in the cerebral cortex is detected better than activity from deeper brain tissue. Visual Analysis of EEG V isual interpretation of EEG i s a skill is mastered with years of experience, but the process that EEG readers or encephalographers, follow can be broken down into steps including the evaluation of EEG frequency, rhythmicity, amplitude, symmetry, and synchrony As an encephalographer scrolls through an EEG, one of the first characteristics noted is the frequency of EEG rhythms, which are measured in Hz. Brain wave activity falls into one of four frequency bands: beta (13 30 Hz), alpha (8.5 12 Hz), theta (4 8 Hz), delta (<4 Hz). Often overlapping frequencies are seen in the waveforms. The next aspect of importance is amplitude, wh ich is measured in microvolts ( V), and can range from low (0 25 V) to moderate (25 75 V) to high (>75 V). Amplitude can be influenced by several factors, such as cortical injury, extra axial fluid

PAGE 18

18 collection (hematoma or hygroma), or increased skull t hickeness, all of which can decrease the measured amplitude. Likewise, s kull defect from fracture or craniotomy will decrease resistance and increase measured amplitudes. Symmetry and synchrony are important to evaluate when determining if the left and right hemispheres of the brain are functioning similarly. If an EEG demonstrates symmetry, then equal frequencies and amplitudes are noted in the bilateral hemispheres. The EEG is synchronous if brain waves are appearing at the same points in time. Fina lly, brain activity can be quite rhythmic, generally with amplitudes waxing and waning in a clean sinusoidal pattern, or activity can be poorly sustained and non rhythmical. An additional essential sk ill in EEG interpretation is the ability to determine wake and sleep st ages These sta ges include: alert awake, relaxed awake drowsy, N1 and N2 sleep or light sleep, N3 sleep, also known as slow wave sleep or deep sleep and rapid eye movement (REM) sleep. Each stage has a characteristic pattern of bra in activity on EEG ( 5 ) Wakefulness is dominated by the presence of low amplitude beta and alpha rhythms which attenuate or decrease in amplitude with eye opening In stage N1 sleep beta and alpha range frequencies are replaced by upper range theta activity, and this is noted maximally in the occipital head regions. Continued infusion of slower theta rhythms occurs during stage N2 sleep along with the appearance of intermittent beta range sp indle activity in the central brain regions. Stage N3 is characterized by the presence of mixed low to moderate amplitude theta and high amplitude delta rhythms. Finally, REM brain activity is similar to that during wakefulness ; however, it is slightly s lower with predominantly low amplitude alpha and

PAGE 19

19 theta rhythms. the central regions. In this research project, we collected samples of EEG data in the relaxed, awake state during eye closure when alpha activity was apparent in the posterior (occipital) channels. This stage was selected because it is the easiest to recognize and would be fast to capture on a short duration EEG and our long term goal is to reduce the recording time and amount of EEG data needed to differentiate patients with epilepsy and non epileptic events. Cellular Substrates of Brain Rhythms The cellular substrates of these brain rhythms are partially understood and described.( 5 ) Diffuse delta waves seen in norma l sleep states originate from oscillations of transient calcium currents between the thalamus and cortex. Theta activity is most commonly noted during stage N1 and N2 sleep. Though the pacemaker source for theta activity remains unclear, the medial septum and its connections to the supramammillary nucleus of the hypothalamus and brainstem reticular formation are involved. Alpha activity is seen maximally in the occipital visual cortex and is referred to as the posterior dominant rhythm. This rhythm has a lso been recorded in the pulvinar and lateral geniculate nucleus of the thalamus. As such, thalamocortical linkages appear to be important in the generation of this activity, which is most prominent in the relaxed wake state with eyes closed. Finally, fa ster beta frequencies predominantly seen in the awake, alert state are thought to originate from diffuse connections made by the mesencephalic reticular formation and intralaminal nuclei of the thalamus. For the purposes of this study, our focus was on t he right left symmetry of amplitude variation and signal regularity of brain rhythms in the anterior temporal regions

PAGE 20

20 Table 1 1. Clinical i ndications for o btaining an EEG Indication Example Diagnosis of seizures disorders Prevalence of spike or sharp waves, or rhythmic delta or theta range activity Classification of seizure disorders Focal or Generalized Localization of seizure onset zones Left or right Frontal Temporal Parietal Occipital Generalized Identification of neurologic disorders with classic EEG patterns including: Herpes Simplex encephalitis PLEDS Creutzfield Jakob disease periodic frontal sharp waves Subacute sclerosing panencephalitis high voltage bifrontal spikes Confirmation that altered mental status is not due to seizure Enceph alopathy Syncope Psychogenic episode Confirmation of brain death Electro cerebral silence Prognosis in coma Based on background pattern, presence or absence of seizure activity, and reactivity to stimuli Confirmation of diagnosis of sleep disorders EEG channels used in polysomnography, and the multiple sleep latency test

PAGE 21

21 Figure 1 indicated on the figure are as follows: 1) crank, 2) marker fibers, 3) on/off switch (far r ight), 4) lens, 5) diaphragm, 6) paper box, 7) tuning fork. Reproduced with permission from Elsevier from the journal article: Gloor, P. Hans Berger and the discovery of the electroencephalogram. Electroencephalography and Clinical Neurophysiology.1969; S28:1 36.

PAGE 22

22 Figure 1 2. The International 10 20 System for Scalp Electrode Placement. Reproduced from the Wikipedia Commons freely licensed media file repository.

PAGE 23

23 A. B. Figure 1 3. Wave discharges. A.). Diffuse, anteriorly maximal 3 hertz polyspike wave discharges due to primary generalized epilepsy. B.) Right temporal sharp wave discharge followed by a run of t emporal intermittent rhythmic delta activity (TIRDA) Both sharp waves and TIRDA are commonly noted in patients with Temporal lob e epilepsy.

PAGE 24

24 Figure 1 4. Rhythmic and sharply contoured, 7 8 Hertz theta activity at the onset of a focal seizure localizing to the left temporal lobe.

PAGE 25

25 CHAPTER 2 EPILEPSY AND NON EPILEPTIC EVENTS Introduction to Epilepsy Epilepti c seizures are transient signs and symptoms du e to abnormal excessive or synch ronous neural brain activity.( 9,10 ) Likewise, e pilepsy is a brain disorder characterized by recurrent and unprovoked seizures Epilepsy is the second most common disorder of the central nervous system, after stroke, and affects about 0.4 1% of the population.( 11 ) An estimated 40 million people worldwide have epilepsy. ( 10 ) The term seizure go ing back to its origin in Greek, means An epileptic seizure is du e to abnormally excessive or synchronous neuronal activity in the brain ( 9 ) During the onset of an epileptic seizure ( i.e. the ictal period), synchronous and rhythmic discharges may originate from one part of the brain (focal or localization related seiz ures) or begin simultaneously in both sides of the hemispheres (generalized seizures). ( 10 ) A fter the onset focal seizures may remain localized within one part of the brain or propagate to the other side of the hemisphere and cause a wider range of synchr onous neuronal activity (secondarily generalized seizures). This research project focuses on differentiating a specific type of type of focal seizure disorder, temporal lobe epilepsy (TLE) from a condition called psychogenic non epileptic event (PNEE) TLE was chosen because it is the most common type of focal epilepsy seen in epilepsy monitoring units.( 12,13 ) Distinguishing epileptic seizure from PNEE can be difficult as clinically they can appear quite similar. ( 14,15 ) However, making the correct diagnosis is important because therapies for the two diagnoses differ greatly. Epileptic seizures are treated with oral anti epileptic medications, epilepsy surgeries, and special diets. Conversely,

PAGE 26

26 PNEE are treated with psy chotherapy, specifically, cognitive behavioral therapy. Psychotropic medications may be helpful as well. ( 16 ) In th is chapter, classification of epileptic seizure disorders is discussed to help the reader understand how a diagnosis of TLE is made. Next clinical and EEG features of TLE will be discussed. Finally, diagnostic challenges in distinguishing epilepsy, including TLE, from PNEE are explained. Classification of Epilepsy Epileptic seizures are classified by type fo r the purposes of formulating appropriate treatment plans and offering prognosis. The most widely used system is that proposed by the International League Against Epilepsy (ILAE) ( 9,10 ) Originally drafted in 1989, this classification system was update d in 2001 and 2010. In the most recent version, a five dimension approached is u s ed. Dimension 1 is focused on localization of the seizure onset zone based on all known clinical data, including information obtained during the history, clinical exam, EEG, and neuro imaging Despite la rge advances in neuro imaging over the last several decades VEEG remains the gold standard for determi nation of seizure onset zones. ( 8,14 ) While D imension 1 describes the brain localization of seizure activity, it does not provide information on what type of seizures are experienced. This information is covered in Dimension 2. Most commonly, the ILAE International Classification of Epileptic Seizures is used for this purpose This system was initially developed in 1964 and revised in 1981. The revised version categorize s seizures based on an electro clinical approach, meaning a combination of signs and symptoms during seizures and EEG findings. Dimension 3 of the ILAE classification system describes the etiology of the epileptic seizures Dimension 4, while not particularly

PAGE 27

27 useful for loca lization, helps to describe the severity of the condition by documenting the frequency of seizures. Finally, Dimension 5 lists related medical information that may be helpful in identifying an epileptic syndrome or s eizure onset zone. This information may include focal neurologic deficits or seizure triggers. Temporal Lobe Epilepsy As described in dimension 1 of the ILAE classification system, localization of seizure onset zone is determined by using all available da neuro imaging and EEG. Seizures may be focal or generalized in onset, and approximately 50% of epilepsies are focal in onset. The ILAE Commissio n (1989) classifies foca l epilepsies acco rding to their anatomical origin. Focal epilepsies may localize to the frontal, temporal, parietal, or occipital brain regions. Localization to the temporal region is most common, as such, the temporal lobe is co nsidered the most epileptogenic region of the brain ( 17 ) The true prevalence of temporal lobe epilepsy (TLE) is unknown However, in the setting of medically refractory epilepsy patients undergoing video EEG monitoring, approximately 2/3 of focal epilepsies localize to the temporal lobes. ( 12,13 ) Many e tiologies of TLE exist. These etiologies include : current or past ce ntral infection system infections ( herpes encephalitis, bacterial meningitis, neurocysticercosis ), t rauma brain injury that leads encepha lomalacia and/or cortical scarring, cortical developmental abnormalities, h amartomas in neurocutaneous disorders, brain tumors ( meningiomas, gliomas, gangliomas) vascular malformations ( arteriovenous malformation, cavernous angioma ), and paraneoplastic syndromes ( anti Hu or NMDA receptor antibodies ).( 17 ) Often, the cause is said to be either cryptogenic meaning the c ause is presu med but has not been identified, or idiopathic,

PAGE 28

28 which may imply a genetic predisposition. Febrile s eizures during infancy and childhood can lead to TLE later in life. Seizure Semiology in TLE TLE seizures are usually brief, lasting 2 3 min utes. These events may be preceded Olfactory auditory, gustatory and visual hallucinations may occur. ( 10,17,18 ) Patients may report distortions of sound or changes in the shape size, and distance of objects. Things may appear shrunken (micropsia) or larger (macropsia) than usual. ( 17 ) Also, those with TLE may experience an aura of vertigo. Psychic phenomena such as a feeling of dj vu or jamais vu or a sense of familiarity or unfamiliarity, are common auras in TLE. P atients may experience depersonalization ( i.e. feeling of detachment from oneself) or derealization ( i.e. surroundings appear unreal). They may also report a sense of dissociation or autoscopy, in which they report seeing their own body from outside Additionally, unexplained f ear or anxiety may precede temporal lobe seizures. Often the fear is strong, and described as a feeling of impending doom. ( 17 ) TLE aura can also be in the form of a utonomic phenomena which may include changes in heart rate, piloerection, and sweating. Patients may experience an epigastric "rising" sensation or nausea. ( 10 ) Following the aura, a temporal lobe complex partial seizure commonly begins with a motionless stare, dilated pupils, and behavioral arrest. Oral alimentary automatisms such as lip smacking, chewing and swallowing may be noted. Ipsilateral m anual automatisms or contralateral dystonic posturing of a limb also may be observed. ( 17,18 ) Patients may continue their ongoing motor activity or react to their surroun dings in a semi purposeful manner ( i.e. reactive automatisms).

PAGE 29

29 A complex partial seizure may evolve to a secondarily generalized tonic clonic seizure. Often, the documentation of a seizure notes only the generalized tonic clonic component of the seizure. A careful history from the patient or an observer is needed to elicit the partial features of either a simple seizure or a complex partial seizure before the secondarily generalized seizure is important. Patients usually experience a post ictal period of confusion, which may help distinguish TLE from PNEE as those with PNEE sometimes have an immediate return to baseline responsiveness. ( 14,19 ) In TLE, p ost ictal aphasia su ggests onset in the language dominant temporal lobe. ( 10 ) Electroencephalographic C haracteristics of TLE In temporal lobe epilepsies the interictal scalp EEG may show the follow ing: n o abnormality. unilateral of bilateral slowing of cerebral activity in the temporal EEG channels. unilateral or bilateral epileptiform spikes sharp waves and/or slow waves ( 1,10 ) During a seizure, ictal EEG activity may begin a t the time of aura onset or not until a complex partial seizure begins. Ictal activity includes: a sudden unilateral or bilateral inter ruption of background activity temporal or multilobar low amplitude fast activity, temporal or multilobar moderate am plitude rhythmic spikes, sharps, or slow waves. Psychogenic Non epileptic Events PNEE are paroxysmal changes in behavior that resemble epileptic seizures, but have no electrophysiological correlate or clinica l evidence for epilepsy. ( 16 ) Also, positive evidence for psychogenic factors that may contribute to these events are often present. Point prevalence of PNEE has been estimated to be low in the range of one person per 30,000 50,000. The incidence rate is equivalent to 4% of that of

PAGE 30

30 epilepsy. ( 20 ) None the less, approximately 25 30% of patients undergoing in patient video EEG monitoring for medically refractory epilepsy have PNEE ( 16 ) A major complication to the issue is that between 5 and 40% of the patients with PNEE have a concomitant diagnosis of epilepsy or have a past history with epileptic seizures .( 16 ) Clinical Manifestations of PNEE PNEE are more often composed of purposeful, asynchronous, apparently consciously integrated motor activity such as thrashing movements of the entire body, opi stotonic posturing of trunk, out of phase limb movements, side to side head movements, forward pelvic thrusting. ( 19,21 ) PNEE patients were more likely to have forceful sustained eye closing at any stage of the seizure and jaw clenching in the tonic phase of convulsive seizures. ( 22,23 ) PNEE is often accompanied by moaning crying (ictal weeping) and stuttering throughout the events. ( 19,24 ) The most common ictal characteristic of PNEE was unresponsiveness without predominant motor manifestations. ( 19 ) P atients with PNEE often describe fluctuating, but more or less continuous, levels of conscious mental activity during their events without the discrete gaps of missing memory that are characteristic of the impaired consciousness during complex partial epil eptic seizures. ( 19 ) More often than in epilepsy PNEE occur in the presence of others and have a more gradual onset (slow increase of symptoms) with abrupt recovery. ( 14,19 ) Pre ictal pseudosleep, in which the seizure arises while the patient seems to be asleep despite electrographic evidence of wakefulness, has been r eported to be specific for PNEE ( 19 ) Autonomic changes can occur with epileptic seizures and PNE E (e.g. coughing, palpitations and pupillary dilatation). In the setting of PNEE,

PAGE 31

31 these autonomic symptoms are likely part of the heightened arousal response attached to panic or other extreme emotional states Etiologies of PNEE PNEE are almost infinitely heterogeneous and are quite different from person to person. Even if the PNEE behavio rs of very different people are morphologically similar, clinical experience reveals that the psychogenic causes may be quite divergent.( 16,19 ) Psychological factors describe the underlying causation of these seizure like behaviors. A common factor is a history of abuse during any time in life prior to PNEE onset. Abuse may have been sexual (most common), physical, or verbal ( 16,19 ) While not all patients who have suffered from abuse will develop PNEE the risk for developing PNEE is clearly increased. Other common examples of such psychological etiology may be personality disorders post traumatic stress disorder, malingering, depression or chronic anxiety, dissociation, somatization disorder, behaviorally oriented concepts of se condary gain and assumption of the sick role (mainly in intellectually impaired persons), personality disorders and organicity. Shaping factors are also import ant in the development of PNEE.( 16,19 ) These factors contribute to shaping the symptoms in the form of seizures like events, as opposed to other movement disorders or other somatic symptoms. A well known shaping factor is living with a relative who has epileptic seizures. Differentiating Epilepsy from PNEE The gold standard diagnostic modality for distinguishing PNEE from epilepsy is inpatient continuous VEEG monitoring. ( 7,14 ) However, there are many limitations to this procedure. First, in order to make a diagnosis, patients must stay in the hospital until all of their typ ical seizure like events are recorded. A typical stay is 3 to 5 days.

PAGE 32

32 VEEG monitoring is a resource that is not universally available, especially in rural areas, and patients with PNEE are misdiagnosed with epilepsy for an average of 7 years.( 19 ) Furthe rmore, once undergoing VEEG, i f one does not have their events before discharge, then a definitive diagnosis cannot be made. Also, VEEG can be quite procedure. In some cases, hospital bills for thousands of dollars may be charged to the patient. As a common seizure provocation technique patients are taken off their seizure medications either immediately or over several days; the speed of removal depends on baseline seizure frequency. Those with epilepsy are at risk for prolonged seizures, status epileptic, need for intubation transfer to an int ensive care unit, and injury. All patients can suffer side effects from rapid withdraw of seizure medication. For all the reasons listed above, a need exist s for the development of alternative diagnostic techniques whereby PNEE can be distinguished from epilepsy. This technique should be one that can be performed safely and in the out patient setting.

PAGE 33

33 CHAPTER 3 MEASURING INTRAHEMIS PHERIC EEG ASYMMETRY IN TLE AND PNEE Asymmetry of EEG activity: In the setting of focal epilepsy, s ingle photon emi ssion tomography, positron emission t omography and magnetic resonance spectroscopy studies have revealed that interictal hypoperfusion, glucose hypometabolism, decreased benzodiazepine binding and metabolic disturbances lateralized to the side of an epi 25 ) These findings are even true when conventional computed tomography or magnetic resonance imaging studies fail to identify a lesion. Studies utilizing spectral power analysis, a technique that measures the amplitude of phy siologic frequency bands, have shown greater entropy of the spectral power in electrodes where interictal discharges appear.( 26 ) This findings provides electrophysiological evidence of brain activity asymmetries in focal epilepsy. Furthermore, p revious research has demonstrated that analyzing hemispheric asymmetries in EEG characteristics may be useful for differentiating focal epilepsy from other controls For example, when comparing patients with focal epilepsy with normal controls or controls with t ension headache, those with focal epilepsy had greater left right asymmetry of total power and alpha power than the control groups.( 27 ) Also, greater asymmetries in sleep spindle intensity (amplitude) ha ve been noted in those with focal epilepsy than those with idiopat hic generalized epilepsy.( 25 ) Furthermore, those with focal epilepsy demonstrated decreased synchrony of brain activity in the area of seizure onset zone than non seizure producing brain regions. These epileptic subjects also have decreased overall brain synchrony when compared to sub jects with chronic facial pain. ( 28 )

PAGE 34

34 For the purpose of this study, we used a novel approach and compared the (left right) interhemispheric asymmetry (IHA) of two quantitatively derived EEG va riables of 1) signal regularity and 2) amplitude variation in patients with TLE and PNEE. The first variable was the pattern matched regularity statistic (PMRS) and the second was the amplitude variation (AV). Both are discussed below. Pattern match reg ularity statistic (PMRS): M otivated by calculations of approximate ent ropy in thermodynamic systems, the PMRS is a probabilistic statistic quantifying signal regularity as shown in Equation 3 1.( 29,30 ) Pr{difference of the next points of xi and x j
PAGE 35

35 The third criterion requires pattern match between x i and x j within a range of (set as 3 in this study). To calculate PMRS, we first d efine a conditional probability, (3 3) Given can be estimated as as in Equation 3 4 (3 4) In Equation 3 5 Finally a PMRS can be estimated. (3 5) As the time series develops into a more regular state, s become larger and PMRS decreases as a result. Amplitude variation: Amplitude variation (AV) is simply the standard deviation of the EEG amplitudes within a detection window This variable has been used in seizure prediction models. ( 29,31 ). However, as far as the investigators in this study are aware, neither the PMRS nor AV have not yet been applied to interictal EEG samples for the purpose of distinguishing those with epilepsy from controls.

PAGE 36

36 CHAPTER 4 STUDY DESIGN AND OUT COME In this stud y, w e tested our hypothesis that inter hemispheric asymmetry (IHA) of the interictal EEG is greater in those with temporal lobe epilepsy (TLE) than those with PNEE We compared IHA of signal regularity using the pattern match r egularity statistic (PMRS) a nd the IHA of amplitude variation (AV). The PMRS calculates the probability that two points will have the same change in slope at the same time, given that two previous points were patterned matched. Further details on the PMRS can be read in the paper by Shiau et al. Cyb ernetics and Systems Analysis 2010 .( 29 ) Methods All EEG data for the project were recorded in the adult Epilepsy Monitoring Unit at the Medical University of South Carolina (MUSC) in Charleston, SC USA Collection and analysis of the E EG data were approved by the MUSC Ins titutional Review Board (IRB), and participants signed informed consent prior to inclusion. EEG samples were collected and analyzed at Optima Neuroscience in Alachua, FL, USA. a diagnosis of TLE or PNEE was confirmed based on a single or multiple typical events having been recorded on VEEG All subjects were adults age 18 years or old. Exclusion criteria included VEEG recordings where a diagnosis was not confirmed because the subject did not have events during the recording. Also, s epileptic and PNEE wer e recorded from the same individual, or if epileptiform ictal activity localized to other brain regions besides the temporal lobes. EEG recordings were obtained using the XLTEK EEG monitoring systems (Oakville, Ontario, Cana da) with a 256 Hz sampling rat e. A 19 electrode scalp

PAGE 37

37 electrode configuration was used according the international 10 20 system (Fig. 1 2 ). A referential montage w as utilized and the referential channel was at a location between Cz and Pz In order to reduce the effects from muscle and movement artifacts, each of the EEG signals were band pass filtered with a low cut filter = 1 Hz and high cut filter = 20 Hz Interictal EEG epochs (at least 10 seconds each) were sampled from recordings obtained from 61 patients (29 with PNEE and 32 with TLE {Left temporal onset in 14, right temporal onset in 10 and independent bilateral temporal onsets in 8}). A total of 610 epochs (10 epochs from each recording) in the relaxed, awake state were sampled from TLE and PNEE patient groups. To reduce c onfounding effects, included interictal EEG epochs were constrained to the following three conditions: 1) no epileptiform discharges; 2) no eye blinking; and 3) presence of a c lear bi posterior alpha rhythm Within each EEG sample epoch, we calculated the PMRS and AV in the F8, T4, F7 and T3 EEG channels utilizing a non overlapping 5.12 second computation window. IHA values were calculated as the difference between right (F8 and T4) and left (F7 and T3) channels, with respect to PMRS and AV values, respe ctively. Within each Both the PMRS and AV variable were found to follow a non normal distribution (Figure 4 3) The Whitne y Mann U test was used to test for significant difference s (p<0.05) in the inter hemispheric differences between the two groups. For a priori power analysis, w e are plann ed a study with 30 TLE subjects and 30 PNEE subjects. The mean and standard deviation of the IHA of the PMRS and AV were not known. The PMRS value (unilateral meansurement ) was known to range from approximately 0.01 to 0.06. The AV value (unilateral measurement ) was known to

PAGE 38

38 range from approximately 1 to 10. For the IHA of the PMRS, a power of 0.995 was calculated based on a standard deviatio n 0.02, i f the true difference in the TLE and PNEE group means was 0.03 A t ype I error probability of 0. 0 5 was used A power of 0.861 was calculated for the IHA of the AV assuming a standard deviation of 2.0 and a true differences in the group means of 2.5, with a Type I error rate of 0.05. Results and Conclusion Both groups included more females than male s (TLE 20/29, PNEE 23/32), and gender ratios were not significantly different (p=0.8035) ( Figure 4 1) The mean age was 36.9 years in the TLE group and 43. 3 y ears in the PNEE group, but the age differences were not sign ificant (p=0.6030) (Figure 4 2) Although length of EEG recording was longer in the TLE group than the PNEE group (84.50 hours versus 70.56 hours), these values were not significantly different. A total of 7 1 outlier values were excluded (60 AV and 11 PMRS). No significant difference was found in AV IHA values (2.9261 vs. 2.6379, p=0. 5065 ) (Figure 4 5) However, TLE samples had significantly higher PMRS asymmetry than PNEE (0.0399 vs 0.0196, p =0.0182 ) (Figure 4 4) We calculated the sensitivity of the PMRS for separating out TLE from PNEE groups, based on empiric mean PMRS asymmetry value for the PNEE (control) group +2 standard devi ations ( 0.0196 + (2*0.01388) on this calculation, 11/32 TLE patients had positive values and the sensitivity of this test was only 34.3%. We found that IHA of the PMRS from the t emporal electrodes is significantly larger in TLE subjects than in NES subjects Our finding is likely due to interruptions in signal

PAGE 39

39 regularity caused by the focal epileptic process in the temporal regions. Patients with focal epilepsy have been shown to have greater asymmetry in spectral power analysis and more delta activity lateralizing to the epileptic focus.( 27,32 ). This intermixed delta activity may be responsible for interruptions in signal regularity. We did not find differences in the IHA of AV. P ossible reasons include that AV IHA may depend on duration of epilepsy, seizure type or severity, or etiology of TLE. However these variables were not collected for this study so their role remains unclear A l imitation of the study is that s ubjects were taking anti epileptic medication. Also, tapering seizure medication is a popular technique for provoking seizures while patients are being monitored in an epilepsy monitoring unit. Therefore, the subjects may have been in the process or tape ring, hold ing, or restarting medications during the recording of the EEG epoch s which were selected for analysis. Furthermore, recording of PNEE patients were compared only recordings of TLE patients. Whether or not AV or PMRS are useful in distinguishin g PNEE from other types of focal epilepsy or even idiopathic generalized epilepsy remains unclear. Finally, we analyzed IHA only in the relaxed, awake state. Different stages of wakefulness and sleep have well described natural fluctuations in frequency, amplitude, and rhythmicity would require separate measurements and analysis for each stage. However, this is an area that may deserve future investigation. In conclusion, our findings suggest that characteristics of the interictal EEG may be useful in di stinguishing patients with TLE from those with PNEE Future studies should focus on more diverse groups of epilepsy patients, additional measures of IHA,

PAGE 40

40 other brain regions (electrodes), and different sleep wake stages in order to improve the potential c linical applicability for separating PNEE from epilepsy. Figure 4 1. Bi = temporal lobe epilepsy and non epileptic seizures=psychogenic non epileptic events. Gender ratios were not significant ly different in the two groups.

PAGE 41

41 Figure 4 2 Bi variate analysis for diagnosis and age (in years). In the tables above epileptic seizures=psychogenic non epileptic events. Age was not significan t ly different in the two groups. Figure 4 3. Distributions of the interhemispheric asymmetry. ( Non norma l) distributions of the interhemispheric asymmetry of the PMRS=pattern matched regularity statistic (left) and the AV=amplitude variation (right).

PAGE 42

42 Figure 4 4. Bi variate analysis of the interhemispheric differences in the PMRS=pattern matched regularity statistic. The interhemispheric PMRS asymmetry was greater in the

PAGE 43

43 Figure 4 5. Bi variate analysis of the interhemispheric differences in the AV=amplitude variation. The interhemispheric AV asymmetry was not significantly different between groups.

PAGE 44

44 LIST OF REFERENCES 1. Greenfield LJ, Geyer JD, Carney PR. Reading EEGs: A practical Approach. Philadelpha: Lippincott Williams & Wilkins; 2010 2. Collura TF. History and Evolution of Electroencephalographic Techniques. Journal of Clinical Neurophysiology 1993; 10: 476 504. 3. Gloor, P.Hans Berger and the discovery of the electroencephalogram. Electroencephalography and Clinical Neurophysiology 1969; S28: 1 36. 4. Chien J. EEG Analysis of brain dynamical behavior with applications in Epilepsy. 2011. Retrieved from: proquest.umi.com/pqdweb?index=0&did=2425198451& SrchM ode=2&sid=1&Fmt=2&VInst=PROD&VType=PQD&RQT=309&VName=PQD& TS=1341245184&clientId=20179 Feb 1, 2012. 5. Schomer DL, Lopes da Silva FH. 6 th ed. Philadelphia: Lippincott, Williams, and Wilkins; 2011. 6. Friedman D, Claa ssen J, Hirsch LJ. Continuous Electroencephalogram Monitoring in the Intensive Care Unit. Anesthesia & Analgesia 2009; 109: 506 23. 7. Cascino GD. Video EEG Monitoring in Adults. Epilepsia. 2002; 43 (S3):80 93. 8. Rosenow F, Lders H. Presurgical evaluation of epilepsy. Brain 2001; 124: 1683 1700. 9. Fisher RS, Boas WE, Blume W, Elger C, Genton P, Lee P, et al. Epileptic seizures and epilepsy: Definitions proposed by the international league against epilepsy (ILAE) and the international bureau fo r epilepsy (IBE). Epilepsia 2005; 46: 470 72. 10. Wyllie E, Cascino GD, Gidal BE, Goodkin HP. Principles and Practice 5 th ed. Philadelphia: Lippincott Williams & Wilkins; 2011. 11. Sander JW. The epidemiology of epilepsy r evisited. Current Opinion in Neurology 2003; 16: 165 70. 12. EEG Monitoring at a Typical Referral Epilepsy Center. Epilepsia 2004: 45 :1150 1153. 13. Diaz Arrastia R, Agostini MA, Madden CJ, and Van Ness PC. Posttraumatic epilepsy: The endophenotypes of a human model of epileptogenesis. Epilepsia 2009; 50(S2): 14 20. 14. Devinsky O, Gazzola D, La France WC. Differentiating between nonepileptic and epileptic seizures. Nature Reviews Neurology 2011; 7: 210 20.

PAGE 45

45 15. Burneo JG, Martin R, Powell T, et al. Teddy bears: an observational finding in patients with non epileptic events. Neurology 2003; 61: 714 15. 16. Bodde NM Brooks JL Baker GA Boon PA, Hendriksen JG, Mulder OG, Aldenkamp AP Psychogenic non epileptic seizures -definition, etiology, treatment and prognostic iss ues: a critical review. Seizure 2 009; 18:543 53. 17. Ko DY, Sahai Srivastava S. Temporal Lobe Epilepsy. Retrived from: emedicine.me dscape.com/article/1184509 overview. Mar 12, 2012 18. Panayiotopoulos CP. The Epilepsies: Seizures, Syndromes and Management. Oxfordshire (UK): Bladon Medical Publishing ; 2005. 19. Bodde NM Brooks JL Baker GA Boon PA Hendriksen JG Aldenkamp AP Psychogenic non epileptic seizures -diagnostic issues: a critical review. Clinical Neurology and Neurosurgery 2009; 111:1 9. 20. Krumholz A, Hopp J. Psychogenic (nonepileptic) seizures. Sem inars in Neurology 2006; 26: 341 50. 21. Geyer JD, Payne TA, Drury I. The value of pelvic thrusting in the diagnosis of seizures and pseudoseizures. Neurology 2000; 54: 227 29. 22. Chung SS, Gerber P, Kirlin KA. Ictal eye closure is a reliable indicator for psychogenic nonepileptic seizures. Neurology 2006; 66:1730 31. 23. Syed TU, Arozullah AM, Suciu GP, et al. Do observer and self reports of ictal eye closure predict psychogenic nonepileptic seizures? Epilepsia 2008; 49 : 898 904. 24. Vossler DG, Haltiner AM, Schepp SK, et al. Ictal stuttering: a sign suggestive of psychogenic nonepileptic seizures. Neurology 2004; 63: 516 19. 25. Clemens B. Mnes A. Sleep spindle asymmetry in epileptic patients. Clinical Neurophysiology 20 00; 111: 2155 59. 26. Inouye T Shinosaki K Sakamoto H Toi S Ukai S Iyama A Katsuda Y Hirano M Abnormality of background EEG determined by the entropy of power spectra in epileptic patients Electroencephalography and Clinical Neurophysiology 1992; 82: 203 07. 27. Drake ME, Padamandan H, Newell SA. lnterictal quantitative EEG in epilepsy. Seizure 1998; 7: 39 42. 28. Warren CP, Hu S, Stead M, Brinkmann BH, Bower MR, Worrell GA. Synchrony in normal and focal epileptic brain: the seizure onset zone is function ally disconnected. Journal of Neurophysiology 2010; 104: 3530 39.

PAGE 46

46 29. Shiau DS, et al. Signal Regularity based Automated Seizure Detection System for Scalp EEG Monitoring. Cybernetics and Systems Analysis 2010; 46: 922 35. 30. Halford JJ, et al. Interictal EE G Dynamics in Patients with Non epileptic Seizures versus those with Temporal Lobe Epilepsy. Abstract. American Epilepsy Society Annual Meeting; San Antonio, TX: 2010. 31. Kuhlmann L Burkitt AN Cook MJ Fuller K Grayden DB Seiderer L Mareels IM Seizure detection using seizure probability estimation: comparison of features used to detect seizures. Annals of Biomedical Engineering 2009; 37: 2129 45. 32. Nuwer MR Frequency analysis and topographic mapping of EEG and evoked potentials in epilepsy Current Opinion in Neurology 2003; 16:165 70.

PAGE 47

47 BIOGRAPHICAL SKETCH Holly Skinner was born in Orlando, FL USA. In 1983 she moved with her family to Tallahassee, FL where she lived until 2001. While in Tallahassee, she graduated from Lincoln High School in 1997, and earned her Bachelor of Science degree in Exercise Physiology from Florida State University in 2001. She then moved to Fort Lauderdale, FL for medical school at Nova Southeastern College of Osteopathic Medicine. Upon completion of medical school in 2005, she moved to Charleston, SC for a one year m edicine Internship, then a four year residency in a dult n eurology After residency, she moved to Gainesville, FL in 2009. While in Gainesville, she completed a one year fellowship in c linical n europhysiology at the University of Florida (UF) Then she worked for the UF Department of Neurology as a c linical l ecturer ( n eurologist ) and participated in the UF Advanced Post graduate Program for Clinical Investi gation Through the program, she was afforded the opportunity to complete a m with a concentration in c linical and t ranslational s cience for which this thesis is written. Under the supervision of her primary mentor, J. Chris Sackellares, she was introduced to Optima Neuroscience Inc. a n eurodiagnostic research co mpany Through mentorship and collaboration with the researchers at Optima, she was able to complete this project. She intends to pursue h er research interests in distinguishing PNEE from epilepsy by way of analy sis of brief EEG epochs