Clinical Neurophysiology: Clinical EEG


1 Imran I. Ali, 1 Nabeel A. Herial, 1 Noor A. Pirzada, 1 Neeraj Kaplish, and 1 L. John Greenfield ( 1 Neurology, Medical University of Ohio, Toledo, OH )

Rationale: Status epilepticus (SE) from all causes is associated with a 25–30% mortality rate. In patients with hypoxic-ischemic encephalopathy (HIE) and SE, the mortality is even higher. Effect of non-convulsive status epilepticus (NCSE) in this group have not been described previously. We, therefore, wanted to review the EEG features of SE in patients with HIE to correlate them with the clinical findings and also identify patients in which intervention may be associated with a better prognosis.

Methods: The EEG database at Medical University of Ohio from 1999–2004 was reviewed. Patients with HIE and fulfilling the EEG criteria for status epilepticus were identified. EEG were reviewed by two Board Certified Electroencephalographers and charts were reviewed for clinical and EEG information. IRB approval was obtained for this study.

Results: Twelve patients were identified who fulfilled the criteria for for HIE and NCSE. There were six men and six women. Mean age was 60.5 years (range 38–85). Six patients had isolated cardiac arrest, two had purely respiratory arrest while four had simulatneous cardiorespiratory arrest. Six patients were comatose with no clinical evidence of seizures (50%), three patients (25%) had intermittent myoclonus, while two patients (16%) had focal clonic activity noted initially. One additional patient had generalized tonic clonic seizure that was followed by coma and EEG evidence of NCSE.

Irregular bifrontal generalized spike and wave discharges were seen in six patients (50%), generalized periodic epileptiform discharges with evolving pattern were seen in five patients (42%), and focal rhythmic spike and wave activity was seen in one patient (8%). Severe background suppression was noted in all patients. Three patients had intermittent theta activity but the background was unreactive. In nine patients NCSE resolved with treatment but in all twelve patients this did not result in improvement of the background EEG activity. All twelve patients died within 30 days of onset of NCSE.

Conclusions: EEG features of non-convulsive status epilepticus are variable but most patients had generalized periodic epileptiform discharges with evolving frontal discharges or rhythmic frontal spike and wave activity. The EEG pattern did not influence outcome or correlate with specific clinical features.

In our series non-convulsive status epilepticus in patients with hypoxic ischemic encephalopathy was associated with a 100% mortality. Although status epilepticus resolved in 66% of the patients with aggressive treatment, this did not result in improved outcome.

1 Noel Baker, and 1 Paul C. Van Ness ( 1 Neurolgy, University of Texas Southwestern Medical Center, Dallas, TX )

Rationale: Subclinical Rhythmic Epileptiform Discharge of Adults (SREDA) is a non-ictal electrographic variant most frequently characterised as a 5–7 Hz rhythm of abrupt onset and termination and located maximally in the temporoparietoocipital regions. It differs from an ictal pattern in a lack of evolution in frequency, distribution and morphology, and in the subjects' lack of clinical change during the duration of the discharge.(1) SREDA has most often been described in adults in the 4th to the 9th decades, but Najarajan et al have described this pattern in 2 young girls.(2,3) We report its occurrence in a young man whose SREDA had characteristics described as rare in adults.

Methods: In the pediatric division of our institution, approximately 2200–2500 routine EEGs and 150–200 prolonged video-EEG studies are performed each year. Since 2001, only one case of SREDA in a pediatric patient was seen in our institution.

Results: Case Report: The patient initially presented at the age of 14 for staring spells. His neurologic examination and IQ were normal. An intial routine and video EEG were interpreted as showing non-convulsive status, despite his ability to interact during the discharges. He was placed on valproate with no clinical change. A second video EEG 3 years later showed bursts of 5–6 Hz discharges occaisionally slowing to 1–2 Hz which were consistently maximal in the frontal regions and at the vertex, bilaterally synchronous in onset, and persisting in light sleep. The bursts lasted up to 5 minutes in duration and ended abruptly with re-emergence a normal background, and comprised >50% of the record during the 3 days of monitoring. The SREDA persisted unchanged during activation procedures, eye opening and closing, and repetition of syllables such as “lilt.” During the bursts, he repeatedly demonstrated his ability to follow commands, answer questions, play video games, and complete neuropsychologic testing.

Conclusions: SREDA is rare EEG pattern occurring in approximately 1/2500 adult recordings and documented in patients with such non-epileptiform complaints as headaches, dizziness, and depression, as well as seizures. Atypical SREDA variants charcterised as frontally predominant, or occurring as delta frequencies, or having a waxing and waning pattern have also been described, comparable to what we saw in our patient. (2) While our experience at our institution confirms that of Nagarajan et all that SREDA is not seen exclusively in adults, it also suggests that SREDA is more rare in pediatric patients.

  • 1Westmoreland B. and Klass D. A Distinctive Rhythmic EEG Discharge of Adults. Electroencephalography and Clinical Neurophysiology, 1981, 51:186–191.
  • 2Westmoreland B. and Klass D. Unusual variants of subclinical rhythmic electrographic discharge of adults (SREDA). Electroencephalography and Clinical Neurophysiology, 1991, 102:1–4.
  • 3Nagarajan L. et al Subclinical Rhythmic EEG Discharge of Adults: SREDA in Two Children. Pediatric Neurology, 2001; 24:313–316.

1 Ramon E. Bautista, 2 Steven Godwin, and 2 David Caro ( 1 Department of Neurology, University of Florida Health Sciences Center/Jacksonville, Jacksonville, FL ; and 2 Department of Emergency Medicine, University of Florida Health Sciences Center/Jacksonville, Jacksonville, FL )

Rationale: Patients frequently present to the emergency room (ER) with mental status changes but without any immediately apparent cause. The EEG is underutilized in this population, partly due to the time and preparation required to perform it, possibly causing a delay in diagnosis. In this study, we determined if an abbreviated EEG (AEEG) performed immediately upon patient arrival to the ER could provide useful information. We specifically wanted to know if AEEGs can be useful in diagnosing non-convulsive status epilepticus (NCSE).

Methods: We performed a 5-minute AEEG using a preformed electrode placement system on 25 consecutive patients who presented to the ER with mental status changes but without any immediately apparent cause as determined by ER physicians. The AEEGs were performed almost immediately after the patient arrived in the ER without the benefit of a neurologic consultation. The AEEGs were only reviewed after patient discharge and did not impact patient care. We also conducted a retrospective chart review of all enrolled patients after discharge to determine the discharge diagnosis. We then performed a Fisher exact test to determine if the cause of mental status changes could be distinguished by AEEG findings.

Results: Of 25 patients entered into the study, two had a discharge diagnosis of new onset NCSE. These patients had AEEG findings consistent with electrographic seizure activity. Seven patients were later diagnosed to have a metabolic/infectious encephalopathy. The AEEGs in four of these patients showed diffuse slowing and one had interictal epileptiform activity. Eight patients were seen due to post-ictal confusion and seven of them had AEEGs that showed diffuse slowing. Two patient were diagnosed with syncope and both had normal AEEGs. One patient had a left MCA-CVA and the AEEG showed diffuse slowing that was accentuated over the left hemisphere. One patient was diagnosed with dementia and had a normal AEEG. The diagnoses in four patients were uncertain even at the time of discharge. Three of these patients had a normal AEEG while one AEEG was uninterpretable due to movement artifact. Aside from those who were diagnosed to have NCSE, no other patient had an AEEG showing electrographic seizure activity. Using Fisher's exact test, mental status changes due to NCSE could be distinguished by the presence of electrographic seizure activity on AEEG.

Conclusions: Our study suggests that even an abbreviated EEG performed early during the ER visit can diagnose NCSE in patients who present with mental status changes of unknown cause. Future studies with larger sample sizes would help definitively establish the utility of this technique. (Supported by University of Florida Health Sciences Center/Jacksonville Dean's Fund Research Awards for Faculty.)

1 Scott Bearden, 1,2 Basim Uthman, and 2 Stephan Eisenschenk ( 1 Neurology, North Florida/South Georgia Veterans Health System, Gainesville, FL ; and 2 Neurology, University of Florida, Gainesville, FL )

Rationale: Nonconvulsive status epilepticus (NCSE) is most often defined as an epileptic state, without convulsive motor activity, lasting at least 30 minutes with associated ictal EEG activity. A clinically evident, sustained change from baseline behavior or mental status should also occur unless the patient is comatose or in simple partial NCSE. Diagnostic challenges are most evident in NCSE patients presenting with stupor or coma since no or only little behavioral changes are clearly apparent. Identifying EEG patterns that are most probably ictal representations of NCSE could be most valuable in early diagnosis and treatment of NCSE. Delay to diagnosis of adult NCSE may increase morbidity and mortality.

Methods: A thorough review of the literature from 1972 to 2004 identified 72 publications involving NCSE in adults. Clinical and EEG presentations of reported cases were analyzed and supplemented with examples from our clinical practice.

Results: We devised a classification scheme summarizing clinical presentations and ictal EEG patterns of NCSE. Five clinical categories combined with common ictal EEG signatures were identified and referenced in five charts integrating the type of NCSE, common patient symptomatology, ictal EEG, interictal EEG, and prognostic information. Two flow charts (combining clinical presentation and EEG) were developed to aid in classification and provide prognostic information for most adult patients in NCSE whose mental state ranges from mild confusion to stupor and for patients who are comatose. The use of benzodiazepines as a diagnostic aide during EEG was scrutinized.

Conclusions: Classification of prolonged ictal confusion, stupor and coma utilizing clinical and EEG presentations into a simple logical framework is possible and may contribute to early diagnosis and timely treatment of NCSE. High awareness of NCSE as a clinical entity presenting as change in mental status is critical for early diagnosis and intervention.

1 Rachel Berman, 2 Michiro Negishi, 2 R. Todd Constable, 3,4 Edward J. Novotny, 3 Susan Levy, and 4,5 Hal Blumenfeld ( 1 Interdepartmental Neuroscience Program, Yale University, New Haven, CT ; 2 Diagnostic Radiology, Yale University, New Haven, CT ; 3 Pediatrics, Yale University, New Haven, CT ; 4 Neurology, Yale University, New Haven, CT ; and 5 Neurobiology, Yale University, New Haven, CT )

Rationale: Absence seizures occur most commonly in children. They involve brief episodes of staring and non-responsiveness. Electroencephalogram (EEG) recordings of typical childhood absence epilepsy (CAE) show bilateral 3–4 Hz spike-wave discharges. True CAE rarely persists to adulthood. Prior published reports of functional magnetic resonance imaging (fMRI) in adult patients describe variable bilateral frontoparietal fMRI increases, decreases, and thalamic increases during generalized spike-wave, with most studies performed at 1.5T. The goal of the present study is to extend these investigations to typical absence seizures in the pediatric population, which may differ from adult patients, using combined EEG and fMRI at 3T.

Methods: Pediatric patients with typical absence seizures, and no other seizure types, were tested. After application of a 19 channel EEG cap connected to an EEG recorder (Neuroscan NuAmps), fMRI was performed using a 3T MR system (Siemens Trio), with continuous EPI BOLD sequence. Multiple 10 minute imaging runs were repeated up to 6 times. To increase the chance of obtaining absence seizures, patients were asked to hold medications for up to 48 hours, and were sleep deprived the night before scanning. EEG artifact was removed by post-processing using temporal PCA-based gradient noise removal. fMRI data during seizure and baseline intervals were analyzed in a general linear model using SPM2.

Results: During typical childhood absence seizures, we found significant blood oxygenation level dependent (BOLD) signal increases in bilateral thalamus. Moderate increases were found in the cingulate, lateral frontal, and parietal cortex. Decreased BOLD signal was observed in the retrosplenial cortex, and to a lesser degree in the bilateral frontoparietal cortex.

Conclusions: Despite increased challenges of performing fMRI in pediatric patients, high quality data can be obtained in this population. As in adult patients, increases and decreases are seen in frontoparietal and thalamic networks. Further study is needed to determine whether variable involvement of different specific regions of cortex and thalamus during absence seizures is related to variable effects on cognition seen during these seizures. (Supported by Betsy and Jonathan Blattmachr Fund.)

1 Milan P. Borkovic, 1 Nebojsa J. Jovic, and 1 Jasna Jancic ( 1 Clinical for Neurology and Psychiatry for Children and Youth, Medical School, University of Belgrade, Belgrade, Serbia, Yugoslavia; Clinical for Neurology and Psychiatry for Children and Youth, Medical School, University of Belgrade, Belgrade, Serbia, Yugoslavia; and Clinical for Neurology and Psychiatry for Children and Youth, Medical School, University of Belgrade, Belgrade, Serbia, Yugoslavia )

Rationale: Juvenile myoclonic epilepsy (JME) is an idiopathic generalized epileptic syndrome. It is often considered as benign because of its non-lesional nature and favourable clinical long-term prognosis. The characteristic EEG pattern is presented with bilateral symmetric polyspike-wave complexes with anterior accentuation. Focal EEG abnormalities are not exclusive in patients with an unequivocal diagnosis of JME.

Methods: A Group of 39 patients of both sexes (17 male, 22 female), aged from 12 to 26 years (mean 17.8 years) with JME were studied for focal EEG abnormalities recorded during the clinical follow-up. Mean seizure onset was 13.5 years. Diagnosis was made according to the history and EEG records. Antiepileptic treatment was started with valproate.

Results: Initial clinical and EEG assesment was carried-out after the generalized tonic-clonic seizure (mainly induced by sleep-deprivation), in 51% of JME patients. Absences in 21% and myoclonic jerks in 28% of patients were observed as initial epileptic events and reffered for EEG examinations. Complete, stable, long-term seizure control was achieved with valproate use in 87% of patients. In remaining 13% no favorable therapeutic response of AEDs was noted, mainly due to the irregular lifestyle. Repeated EEGs during sleep and when awake showed the interictal polyspike-wave pattern or fast/classical spike-wave discharges. Generalized EEG patterns were prevalent as expected: polyspike-waves or spike-wave discharges non-related to photic stimulation were noted in 66.7% and paroxysmal abnormalities to photic stimulation in 35.9%. Non-specific abnormalities were shown in 10.3%, while, bilateral anterior spikes and sharp waves were recorded in 23.1% of JME adolescents. Localization-related EEG abnormalities (focal slow waves shifting between the hemispheres, spikes, sharp waves and focal onset or asymmetrical generalized discharges) were noted in 23,1% of our JME patients.

Conclusions: In more than one fourth of our JME patients, EEG showed focal abnormalities in addition to the generalized EEG patterns. These features should not be mislead to the erroneous interperetation of EEG characteristic for localization-related epilepsy.

1 Luiz E. Betting, 1 Susana B. Mory, 2 Iscia Lopes-Cendes, 1 Li M. Li, 1 Marilisa M. Guerreiro, 1 Carlos A.M. Guerreiro, and 1 Fernando Cendes ( 1 Department of Neurology, University of Campinas (UNICAMP), Campinas, São Paulo, Brazil ; and 2 Department of Medical Genetics, University of Campinas (UNICAMP), Campinas, São Paulo, Brazil )

Rationale: The objectives of this study were to investigate the EEG profile in a group of patients with clinical diagnosis of idiopathic generalized epilepsy (IGE) followed in a tertiary Hospital and to evaluate the contribution of the records for diagnosis and management of these patients.

Methods: We retrospectively studied clinical and EEG features of 180 consecutive patients with IGE. Eighty patients were diagnosed with juvenile myoclonic epilepsy (JME), 35 had absence epilepsy (AE), 13 had generalized tonic-clonic seizures on awakening (GTCS-A), 28 had generalized tonic-clonic seizures only (TCS) and 24 had adult onset idiopathic generalized epilepsy (AIGE). The EEGs were classified in typical (synchronous generalized spike or polyspike and wave discharges with normal background), atypical (with clear focalities or asymmetries) and normal.

Results: A total of 493 EEG exams were analyzed. The first EEG was normal in 45% of the 180 patients and only 33% had typical abnormalities. AE had a higher proportion of typical exams and needed less sequential exams to register a typical abnormality compared to the other groups. By contrast, the serial EEG profile of TCS and AIGE showed a higher proportion of normal and atypical EEG findings.

Conclusions: These findings support previous recommendations that IGE patients should be treated with appropriate therapy based on clinical history. Waiting for a typical abnormal EEG pattern can generate an unacceptable delay in the correct diagnosis and treatment of these patients. In patients with long term epilepsy the diagnosis may be difficult. Furthermore, serial EEGs can help to elucidate the syndromic diagnosis, especially in patients with TCS and AIGE. (Supported by CAPES, FAPESP.)

1 Farzana Darbari, and 1 Sigmund Jenssen ( 1 Neurology, Drexel Medical College, Philadelphia, PA )

Rationale: Sleep has been shown to influence multiple physiological processes i.e. control of blood pressure, glycemia and protein synthesis, and deprivation of sleep could theoretically lead to sickness. Patients who are admitted to a hospital suffer disturbance of sleep due to external, medical and psychological factors. Disturbed sleep leads to decreased sleep latency. We hypothesized that in-patients have decreased sleep latency as compared to out-patients and that sleep latency decreases with length of stay.

Methods: Otherwise completely normal EEG recordings were reviewed prospectively for sleep onset, which we defined as at 30 seconds of stage two of sleep or briefer if stage two of sleep was seen more than once. Patients with known sleep disorders, deliberately sleep deprived recordings and recordings with focal or generalized abnormalities were excluded. We noted medications, length of stay and if recording had been done in AM or in PM. Comparisons were done using the Mann-Whitney test and Pearson's for correlation.

Results: Thirty out-patient and 24 in-patient EEGs were reviewed. Mean sleep latency for out-patients was 16 minutes and 54 seconds (SD 11 minutes and 17 seconds) and for in-patients 11 minutes and 27 seconds (SD 8 minutes and 35 seconds). (P = 0.0314). Length of hospital stay and sleep latency did not correlate. There was no statistical difference in age, number of patients on sleep enhancing medications or time of day of EEG between the two groups.

Conclusions: In-patients have decreased sleep latency as compared to out-patients. This could be related to disturbed sleep. To see if this is a risk factor for poor recovery needs further investigation.

1 Anita Datta, and 1 David B. Sinclair ( 1 Comprehensive Epilepsy Program, University of Alberta, Edmonton, AB, Canada )

Rationale: The objective of the study was to compare the clinical course and outcome of children with typical and atypical features of Benign Epilepsy of Childhood with Rolandic Spikes (BECRS).

Methods: A retrospective case series design was used in the setting of a tertiary care pediatric hospital. One hundred and twenty six children with BECRS were selected for the study. The patients were divided into two groups based on their clinical presentation. Group A consisted of children with typical features of BECRS. Group B consisted of those with atypical features. Atypical features included age of onset, abnormal neurological exam, developmental delay, and presence of other types of seizures. Patients' charts were reviewed for demographic data, family history, comorbid conditions, atypical clinical features, anticonvulsant drugs, and outcome data.

Results: Atypical features were seen in 47% of patients. Comorbid disorders such as Attention Deficit Hyperactivity Disorder and behavioral problems, although seen in both groups, were slightly more frequent in the atypical group. Sixty-seven percent of children had a family history of migraine and 54% were migraine sufferers. Overall, there was no difference between time to become seizure free between the groups: 62% of the typical group was controlled on medication by two years while 71% of the atypical group was seizure free. Thirty-five percent of the typical BECRS group were incompletely controlled on medication, while 65% of the atypical group required the addition of a second anti-epileptic drug. Resolution of the epilepsy occurred at about the same age in both groups with similar long term outcomes.

Conclusions: Atypical features in BECRS are common. Comorbid disorders are seen in both groups. Personal or family history of migraine appears to be associated with the disorder. Overall, the long term outcome is similar in both groups, although the atypical patients may be initially more difficult-to-control. Resolution of the epilepsy and a good long term outcome was not affected by the presence of atypical clinical features.

1 Ronald G. Davis, and 1 Carl R. Barr ( 1 Pediatric Neurology, Pediatric Neurology, PA, Orlando, FL; and Pediatric Neurology, Pediatric Neurology, PA, Orlando, FL )

Rationale: We report on a large series (>25) of pediatric patients who presented to our neurology practice with various neurologic referable complaints, excluding clinical seizure activity, who were discovered to have EEG abnormalities. Patient ages varied from 3–18years. Symptoms included headache with vertigo, vertigo, visual field alterations including Alice-in -Wonderland changes and atypical speech/language dyspraxia. Prior to presentation, there had been no history of seizure or seizure like activity noted in any of the patients by history and record review.

Methods: Routine surface EEG was completed on these patients as part of the workup. Findings included focal sharp wave discharges, focal spike and wave discharges, rhythmic slowing and generalized bursts of spike and polyspike and wave discharges of the non-absence type. A significant percentage also activated with sleep. Neuroimaging was normal in all cases. Therapeutic intervention varied and included the use of antiepileptic medications.

Results: A positive response to symptoms (>50% improvement in symptoms) was noted in >50% of patients. Of the initial nonresponders, a second antiepileptic medication was tried. Of those who did not respond to 2 medication trials, discontinuation of medication was recommended. No significant adverse events were reported for any patients who were treated.

Conclusions: In conclusion, EEG abnormalities can be seen in pediatric patients with atypical neurologic complaints without clinical seizures.How these electrographic abnormalities contribute to symptoms is unclear. Additionally, guidelines for medication intervention will need to be established, though further investigation will be required. Given these finding and significantly positive clinical response in this limited number of patients, EEG should be considered frequently in the diagnostic evaluation of similar pediatric patients.

1 Eliane Kobayashi, 1 Andrew P. Bagshaw, 1 Jean Gotman, and 1 François Dubeau ( 1 EEG Department, Montreal Neurological Institute and Hospital, McGill University, Montreal, Canada )

Rationale: Cerebral cavernous angiomas (CCAs) are frequently associated with focal intractable epilepsy, and their epileptogenicity is usually attributed to the rim of hemosiderin deposits. Co-recording of EEG and fMRI is a new method to investigate the effects of epileptic discharges in the focus and at distance.

Methods: We used EEG-fMRI to identify BOLD responses to interictal spikes in patients with CCAs. Spikes were marked according to spatial distribution, with each spike type determining one EEG-fMRI study. Combined maps of t-statistics were created to assess positive (activation) and negative (deactivation) changes in BOLD signal related to the spikes. Responses were analysed in the lesional and perilesional area, and at distance. The raw fMRI signal in the lesion and its boundaries was evaluated for signal loss due to the susceptibility artefact associated with the lesion and the hemosiderin.

Results: Eight patients, three with multiple CCAs, were scanned, all with temporal lobe epilepsy and spikes. One patient had bilateral spikes, analyzed separately, and three studies with less than two spikes were excluded. Therefore, six studies were analyzed, all showing BOLD responses: four had only activation, one had only deactivation and one had both responses. No response was found in the lesion itself or immediate periphery, and the raw fMRI signal showed loss in lesional and perilesional areas in all patients. Responses involved the perilesional area in 2/6 studies, both concordant with the spike topography, one as activation and the other one as deactivation. In all patients, areas of BOLD responses were identified at distance from the lesion. In the four studies where the spikes were ipsilateral to the temporal CCA, there was activation in that temporal lobe. In the two studies where the spikes were contralateral to the temporal CCA, the response was concordant with the spike in one study, and unrelated to either the spike or the lesion in the other one.

Conclusions: Despite susceptibility artifacts caused by iron deposits, we demonstrated EEG-fMRI responses involving the perilesional areas in 33% of studies, supporting their role in epileptogenicity of CCAs. The frequently distant responses are in accordance with previous studies in lesional and non-lesional temporal lobe epilepsy. These responses may be the result of distant effect of spikes, which, in CCAs, could be enhanced by the possible presence of non-visible microvascular lesions. (Supported by The Canadian Institutes of Health Research (CIHR), Preston Robb fellowship from the MNI and the Savoy Foundation for Epilepsy.)

1 Maria D.C. Gomes, 1 Eliana Garzon, 1 Elza M.T. Yacubian, and 1,2 Americo C. Sakamoto ( 1 Neurology and Neurosurgery, Federal University of São Paulo, São Paulo, São Paulo, Brazil ; and 2 Neurology, Psychiatry and Psychology, University of São Paulo, Ribeirão Preto, São Paulo, Brazil )

Rationale: Hypsarritmia is a well-known EEG pattern that usually appears associated to a severe form of infantile epilepsy. Different patterns of hypsarritmia are recognizable on visual analysis, making it difficult to correlate to outcome. In this study we propose the use of scales for objective analysis of EEG background and hypsarritmia, searching for predictors of EEG outcome in West Syndrome.

Methods: EEGs of children with diagnosis of West Syndrome and infantile spasms were analyzed. Patients were included in the study between October 2002 to October 2004. EEG background parameters encompassed presence of background gradient, amount of delta activity, sleep spindles and assimetry; hypsarritmia parameters included amplitude and percentage of hypsarritmia in the EEG record. Patients were divided in controlled (group 1) and uncontrolled (group 2) and their scores compared. Clinical variables including age of onset of spasms, age at the time of the examination, previous use of antiepileptic drugs, etiology and MRI data were additionally analyzed.

Results: 24 EEGs were assessed and scored. Global scores varied from 7 to 18, background scores from 2 to 12, and hypsarritmia scores from 4 to 9, respectively (groups 1 and 2). Age of onset of spasms were classified as early onset, classical and late onset, and the global EEG scores were 11.8 (sd = 2.28), 12.61 (sd = 3.09) and 13, respectively. Differences were not statiscally significant. Patients'age at the time of the evaluation were also divided in up to 4 months, 4 to 9 months, 9 to 12 months and over 12 months. Global EEG scores regarding patients' age were, respectively 14, 12.50 (sd = 3.12); 11.67(sd = 2.50) and 12.78 (sd = 3.15). There was no correlation between EEG scores and duration of epilepsy, etiology, MRI findings. Lower scores were found in controlled patients, while higher scores in uncontrolled patients. Overall, only background scores were correlated with outcome.

Conclusions: In patients with West syndrome EEG scores were not affected by clinical and etiological data. Global EEG scores do not have prognostic value, but background scores were predictive of EEG outcome (Supported by FAPESP.)

1 Jeffrey W. Britton, 1 Gena R. Ghearing, 1 Eduardo E. Benarroch, and 1 Gregory D. Cascino ( 1 Department of Neurology, Mayo Clinic Rochester, Rochester, MN )

Rationale: The importance of the insular cortex and adjacent temporal lobe in cardiovascular regulation has been demonstrated. Some studies have suggested that cortical cardiac autonomic influences may be lateralized, with sympathetic representation lateralized to the right insula and parasympathetic to the left. Ictal bradycardia, although rare, offers an opportunity to study the functional anatomy of cardiac rate control.

Methods: A query was performed on all electronic reports of prolonged EEG monitoring studies performed at Mayo Clinic Rochester from January 1990 to December 2004. Ictal bradycardia was defined as an R-R interval of greater than 2.0 seconds or activation of cardiac pacing during a recorded seizure. All clinical and electrophysiological data in identified patients were reviewed to determine localization and lateralization of EEG activity at seizure onset and at the onset of bradycardia in these patients. A literature review of previously reported cases was also conducted.

Results: Thirteen patients were identified out of 6,168 patients who underwent video-EEG monitoring at our institution during the study period. Ictal bradycardia was present in 29 of 60 recorded seizures in these patients. Seizure onset was localized to the temporal lobe in all patients. EEG activity at seizure onset did not consistently lateralize to either hemisphere. Seven patients had ictal bradycardia with right temporal seizure onset, five with left temporal seizures and one with bitemporal seizure onset. Seizure activity was bilateral at bradycardia onset in 9 of 13 patients. One patient had seizure activity confined to the right hemisphere, and 3 had only left hemisphere activity at bradycardia onset.

One hundred six cases of ictal bradycardia were found in the literature, with 80 of these patients having simultaneous EEG and ECG monitoring. Of these, seizure onset was localized to the temporal lobe in 35 patients, to the frontotemporal region in 15, the frontal lobe in 13 and the occipital lobe in one. Information regarding the EEG lateralization at seizure onset was available in 55 cases, but only 21 of these cases had information on EEG lateralization at the onset of bradycardia. While 34 of 55 patients had seizure onset lateralized to the left hemisphere at seizure onset, 11 of 21 had bilateral activity at bradycardia onset.

Conclusions: Both our series and a review of the ictal bradycardia literature support the association of ictal bradycardia with temporal lobe onset seizures. However, ictal bradycardia does not appear to be a consistent lateralizing finding. The distribution of seizure activity at the onset of bradycardia was bilateral in our series, suggesting that bradycardia is more likely to occur in the setting of bilateral hemispheric seizure activity. These data do not support the presence of a lateralized cortical zone associated with parasympathetic cardiovascular activity in humans. (Supported by NINDS P01 NS32352-P2.)

1 Preeti Gupta, 2 Robert Rothermel, 1 Darren R. Fuerst, and 1 Aashit K. Shah ( 1 Neurology, Wayne State University School of Medicine, Detroit Medical Cetner, Detroit, MI ; and 2 Psychiatry/Psychology, Detroit Medical Center, Wayne State University School of Medicine, Detroit, MI )

Rationale: The Wada test is a neurological test used for epileptic patients who are scheduled for epilepsy surgery to evaluate memory function and language dominance by injecting a short acting anesthetic, sodium amobarbital, in each internal carotid artery separately, anesthetizing the part of the brain perfused by that circulation. We have observed shivering in a portion of patients undergoing the Wada test. The relationship of shivers in the context of this test has not been delineated. Therefore, in this study, we attempt to attain a further understanding regarding this clinical phenomenon.

Methods: We studied a total of 14 patients (9.5–50.25 years) prospectively who have undergone the Wada test since 2003. Multiple variables including age, race, sex, site of injection, dose of amobarbital, duration of shivering, duration of slowing on the EEG (as a surrogate for continued effects of amobarbital on the brain), and hemispheric dominance were analyzed.

Results: Of the fourteen patients, six (43%) experienced shivering during the Wada test. The duration of shivering was similar to the duration of amobarbital's effect as monitored by the presence of slowing on EEG. No significant difference in the two groups was noted in terms of age, race, sex, dose of amobarbital, and hemispheric dominance. Of these six patients, four patients shivered after injection to the left ICA and two patients shivered after injection to the right ICA.

Conclusions: There is no difference among patient characteristics in those who shivered vs. those who did not, suggesting an anatomical cause for this clinical observation. The anterior cerebral artery and the anterior communicating artery, branches of the ICA, perfuse the anterior hypothalamus. Thus, sodium amobarbital injected into the ICA likely causes a temporary functional lesion of the anterior hypothalamic-preoptic area (heat center). This area normally inhibits the primary motor center for shivering. This “lesion” likely causes a release of the primary motor center for shivering from the inhibitory effects from the “heat center” resulting in the clinical manifestation of shivering.

1 Yun Jung Her, 1 Joon Soo Lee, 2 Jong Doo Lee, and 1 Heung Dong Kim ( 1 Pediatrics, Institute of Handicapped Children, Yonsei University, College of Medicine, Seoul, Seodaemun-ku, Korea ; and 2 Diagnostic Radiology, Yonsei University, College of Medicine, Seoul, Seodaemun-ku, Korea )

Rationale: Changes in metabolism due to the differences of cerebral blood flow during ictal and interictal period can be quite various when observed by 2-deoxy-2-[18F]fluro-D-glucose(FDG)-PET. So, we performed EEG and PET concurrently to evaluate the effect of epileptiform and non-epileptiform discharge of EEG on glucose metabolism. We tried to reduce the rate of false lateralization of PET reading and evaluate the relationship between EEG and PET result, thus deciding the appropriateness of concurrent study of PET and PET-EEG along with the importance of PET-EEG in diagnosing and localizing a lesion.

Methods: 73 children with pediatric epilepsy who received PET and PET-EEG simultaneously were included in our study. We studied their clinical aspects of seizure, the severity of focal slow wave during interictal period with the frequency and localization of polyspike or sharp wave to compare with the result of PET. Also, we evaluated the relationship between pathologic results and the concurrency rate between PET and PET-EEG in patients who received operation.

Results: 73 patients received PET and PET-EEG concurrently. Their sex ratio was 1.8:1. When grouping the patients according to the site of the lesion in EEG, 53 children proved to have a local lesion showing the ratio of 2.7:1 to children without it. Among 53 children, 27 had focal seizure while 26 had secondary encephalopathy. The frequency of epileptiform discharge was 18% in no/rare, 10% in occasional, and 72% in frequent group while showing 0%, 43% and 68% rate of concurrency with PET each with high correlation (p = 0.001, r = 0.491). Severity of non-epileptiform discharge was classified as no, mild, moderate, and severe and were 21%, 34%, 23%, and 22% of the whole group along with 13%, 52%, 65%, 69% of concurrence rate (p = 0.001, r = 0.365). According to EEG study, 29% of focal lesion was located in frontal, 14% in centrotemporal, and 8% in occipital area. 22% showed lateralization and 27% without it. The rate of concurrency between epileptiform discharge and PET according to the location of focal lesion was 67%, 80%, 67%, 75%, and 5% (p = 0.001). The same rate was 57%, 30%, 50%, 88%, 25% (p = 0.002) between non-epileptiform discharge and PET showing certain difference. 10 patients went through operation. 6 showed concurrent result in PET and PET-EEG and 4 out of them had concurrency with 24 hr video EEG and brain MRI as well. They had no seizure event after the surgery and proved to have cortical microdysgenesis or dysplasia in pathologic evaluation.

Conclusions: Epileptiform discharge and non-epileptiform discharge in EEG showed certain association with hypometabolim in PET study. In conclusion, we recommend PET-EEG to reduce false lateralization and to localize lesion in cases of high frequency or severity.

1 Mark D. Holmes, 2 Micah Brown, and 2,3 Don M. Tucker ( 1 Neurology, University of Washington, Seattle, WA ; 2 Electrical Geodesics, Inc., Eugene, OR ; and 3 Psychology, University of Oregon, Eugene, OR )

Rationale: This study was undertaken to examine the spatiotemporal dynamics of interictal epileptiform discharges (IEDs) in temporal lobe epilepsy by applying methods that improve both spatial localization and temporal precision in mapping the rapidly evolving discharges.

Methods: Eight patients with temporal lobe epilepsy underwent dense array 256 channel scalp EEG recordings. Localization of IEDs at 10 msec intervals in relation to a standard MRI model was accomplished with dipole fits and with two distributed linear inverse methods of source analysis, using probabilistic gray matter voxels as source constraints. The results were compared to conventional EEG waveform interpretations and the limited source analysis possible with conventional EEG.

Results: The three methods of source localization yielded convergent results. The initial onset of the averaged IED localized to one basal-mesial temporal lobe. Adjacent temporal lobe regions become involved within 10–20 msec. In several patients, there was rapid spread to the opposite temporal lobe and, in some cases, to extratemporal regions before returning to the original temporal lobe. Both temporal lobes often were sources for IEDs. Each IED showed unique propagation patterns. Initial IED temporal involvement was concordant with the site of surgery in most subjects. Restriction of mapping to a conventional EEG montage resulted in fewer IED detections and frequently misleading results from both visual inspection and source localization.

Conclusions: Discerning the rapid propagation of IEDs in temporal lobe epilepsy can be achieved by applying source analysis to dense array recordings. Dynamic brain mapping improves discharge detection, reduces margin of error in source analysis calculations, enhances reliability of anatomic localization of discharges with msec resolution, and yields insight into cortical network involved in epileptic discharges.

1 Yue-Loong Hsin, 2 Tomor Harnod, and 1 Shin-Zong Lin ( 1 Neurology, Tzu Chi General Hospital, Hualien, Taiwan ; and 2 Neurosurgery, Tzu Chi General Hospital, Hualien, Taiwan )

Rationale: Because of the excellent therapeutic effect and high safety of deep brain stimulation (DBS) in certain neurologic disorders, its clinical application has been extended to the treatment of epilepsy. Therefore, the procedure of stimulation electrode implantation provides an opportunity for studying neurophysiology of deep brain structures. We investigated the ictal EEG activity of subthalamic nucleus (STN) in treatment of patient with refractory epilepsy by deep brain stimulation (DBS).

Methods: Seven patients with refractory epilepsy underwent implantation of DBS leads into bilateral STN. In addition to the extension of STN leads as intracranial EEG electrodes, subdural strips were implanted to the corresponding cortical regions of epileptogenesis. To obtain ictal and interictal epileptiform discharges, prolonged video-EEG recording was performed before delivering electric impulses into the STN. In a patient, the intracranial EEG was co-registered with electromyography of muscles involved in the myoclonic seizures.

Results: Over 50 epileptic seizures including partial seizures, atypical absence seizures, myoclonic seizures, tonic seizures, tonic-clonic seizures, and unclassified seizures were recorded. Ictal EEG activities in the STN had fast alpha activity, rhythmic spiking, and quasi/periodic spike/polyspike-and-slow-wave complexes that were not different from cortical epileptic patterns. Where there were bilaterally synchronous and rhythmic spikes and spike/polyspike-and-slow-wave discharges in generalized seizures would be faithfully reflected at bilateral STN simultaneously with opposite polarity. Additionally, myoclonic bursting was present at the STN.

Conclusions: The STN generates epileptic EEG activity. The STN participates in the network interactions between cortical and subcortical structures. Diverse seizure activities of the STN reflect the pathogenesis of epilepsy that should influence the outcome of STN DBS.

1 Mohammed M.S. Jan ( 1 Neurosciences, King Faisal Specialist Hospital & Research Center, Jeddah, Saudi Arabia )

Rationale: Electroencephalography (EEG) is an important tool in investigating children with neurological disorders, particularly epilepsy. The objectives were to examine the relationship between clinical indications and EEG results, and assess the predictability of a normal result.

Methods: 438 consecutive pediatric EEGs were included prospectively. One certified electroencephalographer (EEGer) reviewed EEG requisitions and recorded his prediction of a normal result. EEGs were reviewed separately and the relationship between the clinical indications and EEG abnormalities was recorded.

Results: Children's mean age was 5 years (SD 4.2). Pediatric neurologists ordered 32% of EEGs. The first EEG was studied in 65% of cases. Overall, 55% of the EEGs were abnormal. Repeat EEGs were 2 times more likely to be abnormal (95% CI 1.3–3, p = 0.001). Established epilepsy, using antiepileptic drugs, and sleep record, highly correlated with an abnormal result (p < 0.0001). The EEGer predicted 26% of the EEGs to be normal. A normal EEG was correctly predicted in 97% of non-epileptic paroxysmal events, however, normalization of EEG was correctly predicted in only 54% of children with seizures. EEGs of 15 (3.4%) children with epilepsy revealed unexpected findings that completely changed their management.

Conclusions: A normal EEG is highly predictable in non-epileptic paroxysmal events. EEGs of children with epilepsy are not predictable and may yield unexpected results.

1 Howard L. Kim, 2 Pauline Filipek, 2 Joseph H. Donnelly, 2 Anne E. Tournay, and 3 Teri M. Book ( 1 Neurology, University of California at Irvine College of Medicine, Irvine, CA ; 2 Pediatrics, University of California at Irvine College of Medicine, Irvine, CA ; and 3 For OC Kids, Orange, CA )

Rationale: In a tertiary care center, we performed video-EEG monitoring on 32 autistic children, and found that 19 had interictal epileptiform abnormalities (IEAs). All recorded seizures in this group were non-epileptic. Therefore, we question whether the IEAs in these children represent an intrinsic part of the autistic phenotype

Methods: Video-EEG monitoring studies of autistic children (n = 32) were compared with those of non-autistic children with other cerebral disorders (n = 20). This comparison group included children with developmental delay, learning disability, and attention deficit disorder, who like to autistic group, underwent continuous EEG or video-EEG monitoring to exclude epileptic seizures. None of the patients from either group proved to have epilepsy. Specific parameters that were examined included the occurrence of IEAs and the types, i.e., focal, multifocal, and generalized IEAs.

Results: The autistic group consisted of 27 boys and 5 girls with a median age of 5 years (2–13 years). The comparison group consisted of 12 boys and 7 girls with a median age of 7.5 years (1–18 years). Nineteen of the 32 (59%) autistic children and 12 of the 20 (60%) children in the comparison group had IEAs, respectively. Of these children with IEAs, 11 of 19 (58%) in the autistic group and only 3 of 12 (25%) in the comparison group had generalized IEAs. Multifocal IEAs also tended to occur more often in the autistic group than in the comparison group (32% vs. 8%). In contrast, focal IEAs tended to predominate in the comparison group (75%) compared to the autistic group (42%), although a minority of patients in each group had co-existence of focal/multifocal and generalized IEAs.

Conclusions: IEAs may occur in non-epileptic children with a variety of cerebral disorders. However, the abnormalities in children with autism tend to differ from those of a similar cohort without autism. Specifically, generalized IEAs may constitute an intrinsic part of the autistic phenotype.

1 Oh-Young Kwon, 1 Yeongsoo Kim, 1 Ki-Jong Park, 1 Nack-Cheon Choi, and 1 Byeong Hoon Lim ( 1 Department of Neurology, Gyeongsang National University College of Medicine, Jinju, Gyeongnam, Korea )

Rationale: True anterior temporal electrodes (TATEs, T1 and T2) are as effective as mini-sphenoidal electrodes, and significantly superior to anterior temporal electrodes (ATEs, F7 and F8) to detect epileptiform discharges. In addition, non-invasiveness makes TATEs more useful in clinical application. Low Resolution Brain Electromagnetic Tomography (LORETA) is a mathematical model for the reconstruction of current source from digitalized scalp electroencephalography (EEG). Using LORETA, we investigated the current source distribution patterns of interictal spikes on TATEs and those on ATEs respectively, and tried to find the difference between them.

Methods: In the each digital scalp EEG of epileptic patients showing the interictal spikes of maximum amplitude on T1, T2, F7 or F8, we selected the most typical spike. The EEGs of ten patients were enrolled for the spike on each electrode respectively. All the 40 patients were epileptics and had complex partial seizures. The EEG data were collected from 21 channels using an average reference montage. In each typical spike, one time frame showing the highest amplitude of global power field was selected and LORETA image was acquired using that time frame. The average LORETA images for each electrode were obtained respectively. The Statistical non-Parametric Map (SnPM) between spikes on T1 and those on F7, and the SnPM between the spikes on T2 and those on F8 were acquired.

Results: The current source density in the average LORETA image of T1 spikes were maximum in Brodmann areas (BAs) 20 and 21. That of T2 spikes were maximum in BAs 13 and 21. The current sources of T1 and T2 spikes were mainly located in the anterior temporal area, and rare in the frontal lobe. The current source density in average LORETA image of F7 spikes were maximum in BAs 21, 13, 45 and 46. That of F8 spikes were maximum in BAs 13, 21 and 11. The current sources of F7 and F8 spikes were mainly located in the anterior temporal area and it was also observed in the insular cortex and the inferior frontal area. The difference of current source density in the SnPM between T1 and F7 spikes was maximal in BA 22 and minimal in BA 45. That difference between T2 and F8 spikes was maximal in BA 29 and minimal in BA 6.

Conclusions: Both of TATEs and ATEs may mainly reflect the current source of the anterior temporal area. The current source of the inferior frontal gyrus may be more easily detected by F7 than T1 and that of the superior temporal gyrus by T1 than F7. The current source of the superior frontal and precentral gyrus may be more easily detected by F8 than T2 and that of the posterior cingulate gyrus by T2 than F8. Thus TATEs may have less influences of the current source of the frontal lobe and reflect the current source of the temporal area more specifically than ATEs.

1 Pål G. Larsson, and 1 John Wilson ( 1 Department of Neurodiagnostics, National Centre for Epilepsy, Sandvika, Norway )

Rationale: The last ten months we carried out a total of 440 recordings of which more than 200 had a manifold increase of epileptic activity during sleep. The majority of these patients had a seizure disorder, but some had only cognitive and/or behavioural disorders. The purpose of this study was to characterize the epileptic activity and relate it to the clinical diagnosis.

Methods: All recordings, including at least one night, with epileptic activity during sleep were included. Since many patients had recordings before and after start of drug treatment, only the first recording was included, in total 114. Referrals, and other available information were used for diagnosis. The patients were grouped as having epilepsy, attention deficit (ADHD), autism spectrum disorder (ASD) or Landau-Kleffner syndrom (LKS). Spike index was calculated as the fraction of time when there was less than 3 s between spikes in 10-minutes epochs.

Results: The mean age of the patients was 11 years. There was only one patient below the age of five and three above 17. The mean spike wave index for the whole group was 45.6% during slow wave sleep, 15.8% during REM-sleep and 9.4% awake. The mean number of detected spikes was 15872 during sleep. In the patient group with seizures the mean increase in epileptic activity from awake to sleep was 81.4 times. In patients without seizures the mean increase was 51 times (n = 16). The difference is not significant (p = 0.33). ADHD, ASD and LKS in patients without seizures had a mean increase of 38.1 (n = 5), 9.5 (n = 2) and 22.9 (n = 4) times respectively. The mean inter spike interval during slow sleep was 1.1 s in the seizure group and 2.4 in the non-seizure group (p = 0.0034).

Conclusions: There was a significant longer inter spike interval in the group without seizures compared to the epilepsy group. This may indicate that higher spike frequency increases the change of clinical seizures. The increase in epileptic activity during sleep was highest in the ADHD group without seizures, but the difference did not reach significance compared to the epilepsy group. The suprisingly high increase in epileptic activity during sleep in the ADHD group warrants further studies. All groups showed large spread that partly reflects the age differences and partly may be due to heterogeneity in the groups.

1 Tobias Loddenkemper, and 1 Richard C. Burgess ( 1 Department of Epilepsy and Sleep Disorders, The Cleveland Clinic Foundation, Cleveland, OH )

Rationale: Left hemispheric predominance of interictal epileptiform discharges (IEDs) has been highlighted by a previous study from our center. Lateralization becomes slightly more prominent in adults than in children. However, no data on lateralization of benign focal epileptiform discharges of childhood (BFED) are available. The objective of our study was to determine the distribution, frequency and type of epileptiform abnormalities as a function of age and to compare it with BFEDs and slowing.

Methods: The current study retrospectively reviewed all routine (25,935) and multi-hour (5449) EEGs during a consecutive 10 year period. 7704 patients under the age of 18 years were included. All EEGs were read according to a systematic EEG classification system (Lüders et al., 1993). EEGs were reviewed for unilateral epileptiform discharges, unilateral BFED and slowing. The lateralization of the abnormality was counted only once for each patient by including just the first abnormal EEG. Additionally subsequent EEGs of all infants under the age of 3 years with unilateral IED were reviewed for change in lateralization of IED over time.

Results: EEG recordings with unilateral regional IED were seen in 612 (8%) pediatric patients; left-sided IED were seen in 260 patients (55%), right-sided in 213 (45%) and both right and left-sided in 139. There was no difference between infants (40 left, 33 right) and older children (220 left, 180 right).

BFEDs were seen in 136 (49 left, 50 right, 37 bilateral) out of 7704 pediatric patients (1.8%). No BFEDs were seen under 3 years of age. Continuous unilateral regional slow was seen in 193 patients, with 97 (20 infants) on the left and 96 (18 infants) on the right. Intermittent slow (less than 80% of the recording) was seen in 199 patients (32 infants) on the left (61.6%) and in 123 patients (9 infants) on the right.

In 118 infants with unilateral or bilateral regional IED, the lateralization of IED in subsequent EEGs was reviewed. 32 patients with uni- or bilateral IEDs had subsequent EEGs, with localization changed in 9. Two patients converted from left to right (one of these two underwent epilepsy surgery on the left), and none converted from right to left. Five patients with initially left sided and two with initially right sided IED showed bilateral IEDs subsequently. Two patients with initially bilateral AEDs had subsequently right-sided IED in one and left-sided IED in the other.

Conclusions: Regional epileptiform discharges were seen in approximately 8% of patients. Interictal epileptiform discharges and intermittent slow were more frequently seen in the left hemisphere. Age adjusted analysis of the data showed no difference between infants and older children. Continuous slow and BFEDs were equally distributed between both hemispheres. Longitudinal follow up of consecutive infant EEGs did not reveal a trend towards IED conversion from one side to the other.

1 Fumisuke Matsuo ( 1 Neurology, University of Utah School of Medicine, Salt Lake, UT )

Rationale: Advanced digital technology significantly improved accuracy of EEG feature analysis, but detection of interictal focal epileptiform transients (IFET) with clinical relevance continues to depend on human visual analysis. Particularly problematic has been the difficult-to-define spatiotemporal IFET profile that experienced interpreters find no difficulty in recognizing and applying to polygraphic EEG analysis. EEG geometry was examined in IFET polygraphic derivations from randomly chosen clinical cases. The primary objective was to delineate the time-series characteristics of well-formed IFET. The secondary objective was to demonstrte polygraphic variations that indicate subtle variation in IFET source location within an area of unit interelectrode distance.

Methods: During 2004 a total of 53 standard EEG cases revealed a varying number of IFET. A representative IFET chosen from each case was reformatted in common average reference derivations. One derivation with the largest IFET deflection was chosen, approximating its source to a 10–20 scalp electrode. The 53 representative IFET were reviewed for waveform criteria culled from the literature, and ranked in a descending order by evaluating the degree of fit, when polygraphically superimposed. Well-formed IFET were then reformatted in serial bipolar derivations for detailed examination of the IFET peak and phase relationship.

Results: The first 5 ranked IFET referred to common average reference are summarized in Figure. inline image

A well-formed IFET consists of 3 major peaks with different waveform. The third (blunt) peak varied most, and attenuated in lesser-ranked IFET. When displayed in a chain of 5 serial bipolar derivations, phase relationship surrounding the IFET peak differed among cases. This finding was confirmed by examining multiple IFET within each case

Conclusions: Digital technology enables the viewer to reformat EEG data off-line and improve EEG feature extraction. The notion of well-formed IFET is elementary, but robust, because it allows to supplement voltage criteria with additional clinically relevant geomatric paramaters, and can improve automated IFET screening and source modeling. Demonstration of phase relationship surrounding the IFET peak specific to each case confirmed the feasibility of demonstrating subtle variation in location of an equivalent current dipole relative to head-surface electrodes. This method is easy to apply in clinical settings and can expand the scope of EEG analysis in space domain.

1 Daniela Minecan, 1 Janel Schneider, and 1 Lawrence Hudson ( 1 Neurology, University of Michigan Medical Center, Ann Arbor, MI )

Rationale: Nonconvulsive status epilepticus(NCSE) is a heterogenous entity, often associated with subtle clinical signs. There are few reports on the electroencephalographic(EEG) patterns in such patients of pediatric age range. NCSE can be further subdivided into 2 groups: complex partial and generalized.

Electroencephalograms(EEGs) are known to have a fairly high degree of sensitivity and specificity for identifying NCSE.Some of the patterns previously described in association with NCSE in the pediatric patients include: generalized synchronous discharges, alpha coma, burst suppression pattern.

We describe an unusual pattern that was observed in three of our patients that presented with mental status(MS) changes and were diagnosed with NCSE.

Methods: Three patients(age 3–8 years) who presented with MS changes and decreased responsiveness were evaluated at the University of Michigan Hospital. Clinical features(age at presentation, previous history of any kind of seizures, duration of symptoms) and EEGs were assessed. Other diagnostic tests were obtained in each of the patients, according to the clinical presentation.

The EEGs were performed with a number of standard bipolar and referential montages, according to the international 10–20 system of electrode placement.

All three patients were loaded with antiepileptic drugs, and two of them required antibiotic treatment as well.

Results: In two of the patients, the background activity was characterized by very high amplitude(>250uV) delta activity(1–2Hz) and which was diffusely distributed. This was at times rhythmic, but without clear evolution. In one patient, there were multifocal independent sharp waves noted, with no specific hemisphere predominance. In the other patient, independent right/left frontal sharp waves were present.

The third patient EEG was notable for a very high amplitude delta activity(1–2Hz) over the right hemisphere and loss of faster frequencies within the same region. Bursts of high amplitude rhythmic 2.5 Hz spike and wave activity, up to 15 seconds in duration, were present over the left frontal region and less often over the left temporal region. No distinct clinical events were noted, other than the MS changes.

All EEG patterns improved in paralel with the clinical symptoms, with gradual return of the MS to the previous baseline status.

Conclusions: The very high amplitude rhythmic delta activity, whether focal or generalized, is an unusual and not previously described EEG pattern in NCSE. All our patients were younger than 10 years at the time of onset, which raises the question of an age dependent underlying pathophysiologic mechanism responsible.

More EEG studies of patients in NCSE will need to be observed in order to confirm how prevalent this EEG pattern might be and to better characterize its clinical associations(e.g.age, etiology).

This may also contribute to the better and early identification of this category of patients and rapid institution of treatment.

1 Hian-Tat Ong, 2 Karen J.L. Lim, 1 Poh-Chan Low, 2 Stacey K.H. Tay, 1 Shang-Chee Chong, and 2 Poh-Sim Low ( 1 Children's Medical Institute, National University Hospital, Singapore, Singapore ; and 2 Paediatrics, National University of Singapore, Singapore, Singapore )

Rationale: Hyperventilation (HV) is an important activation procedure during routine pediatric electroencephalograph (EEG) recording. Nevertheless, there is little scientific evidence to allow standardization of the HV procedure, in particular the duration of HV required, to be deemed adequate. We attempt to determine the adequacy of HV in pediatric subjects during EEG recording, by assessing the degree of EEG build-up of slow activity and the corresponding non-invasive end-tidal carbon dioxide (EtCO2) measurement.

Methods: Following informed consent by the parents or adult caregivers, we assessed a total of 117 pediatric subjects aged 5 to 16 years who were scheduled for routine EEG studies during a 3-year period from April 2002 to March 2005. The inclusion criteria were the subject's ability to perform adequate HV and a normal waking EEG background. All subjects performed HV for 4 minutes. An electronic metronome was used to synchronize the over-breathing for all subjects in order to achieve an optimal respiratory rate of 30 per minute. A nasal cannula was attached to the subject and the EtCO2 values were recorded using a Nellcor NPB-75 capnometer. The EEG build-up of slow activity was categorized according to the severity of slowing within a 10-second epoch of the EEG recording, as previously reported by Yamatani et al. The four categories consisted of: an increase in amplitude but no intermixed slow waves (+), mild or partial transient slowing (++), moderate slowing (+++), and marked slowing with disappearance of background activity (++++).

Results: The mean EtCO2 values and their standard deviations after starting HV were (30.7 ± 5.0) mm Hg at 1 minute, (27.7 ± 5.0) mm Hg at 2 minutes, (26.2 ± 4.8) mm Hg at 3 minutes, and (26.1 ± 4.9) mm Hg at 4 minutes. The number of subjects showing the 4 different degrees of EEG slowing at various time periods after commencing HV is shown in the table.

Conclusions: Our study showed that when performing hyperventilation, there was a correlation between the EtCO2 values achieved and the degree of EEG slowing present, for the duration of HV till 3 minutes. However, the EtCO2 values and the degree EEG slowing did not show a significant difference between HV for 3 minutes versus HV for 4 minutes. Thus, the results of our study suggest that good hyperventilation effort for 3 minutes is sufficient and this could be used for standardization of the duration of HV required in routine pediatric EEG recording. (Supported by National Healthcare Group (Singapore) cluster research fund: Grant Number NHG-PRP-01139.)

Table 1. Number of subjects showing different degree of EEG slowing during various intervals of HV
 Degree of EEG slowing
HV at 1 minute564912 0
HV at 2 minutes335724 3
HV at 3 minutes22572810
HV at 4 minutes31512510

1 Maryann Pelkey, 1 Sanjeev V. Kothare, 1 Geoffrey Harrison, and 1 Divya S. Khurana ( 1 Section of Neurology, St. Christopher's Hospital for Children, Philadelphia, PA )

Rationale: Attainment of sleep during an EEG is important as interictal epileptiform discharges are often enhanced during drowsiness, stage I and stage II sleep and transition from drowsiness to arousal. Sleep deprivation is often used as a technique to obtain sleep during a routine EEG recording while some EEG laboratories routinely sedate young children to obtain a sleep recording. Both techniques have their drawbacks. Hyperventilation during an EEG is often followed by drowsiness; we therefore hypothesized that hyperventilation early in the course of an EEG would encourage sleep production. Conversely, photic stimulation early in an EEG delays sleep due to its arousing effect.

Methods: This was a prospective study of children over the age of 2 years who were scheduled for a non sleep deprived EEG between 9–11 am in the Neurophysiology Laboratory of a tertiary children's hospital over a 3 month period. The first 32 children received photic stimulation in the beginning of the EEG and hyperventilation at the end; the next 32 children were hyperventilated early in the course of the EEG and photic stimulation was performed at the end of the EEG. Both groups were assessed for presence or absence of sleep during their EEG recording.

Results: One of the 32 (3.1%) children in the first group who received photic stimulation in the beginning of the recording transitioned to sleep. Twelve of 32 children (37.5%) who hyperventilated at the beginning of the recording attained stage II sleep. This was statistically significant (p = 0.001).

Conclusions: Hyperventilation performed early during an EEG increases the likelihood of attaining sleep, while photic stimulation early in the recording decreases the chances of inducing sleep. Therefore, in order to optimize attainment of sleep, pediatric neurophysiology labs should routinely begin the EEG with hyperventilation and end with photic stimulation

1 Mark Quigg, 2 Boris Kovatchev, and 2 Daniel Cox ( 1 Neurology, University of Virginia, Charlottesville, VA ; and 2 Psychiatric Medicine, University of Virginia, Charlottesville, VA )

Rationale: Driving performance is a skill in which impaired vigilance due to diabetic hypoglycemia (HG) has both personal and public health consequences. Studies with the use of a driving simulator show that driving impairment occurs coincident with impaired vigilance measured by EEG that accompanies onset of hypoglycemic symptoms. We evaluate whether recent driving history or awareness of HG during driving are associated with changes in EEG-vigilance measured while driving during euglycemia (EG) or during HG.

Methods: Simulated driving was monitored in two 30 minute trials on 2 consecutive days. During EG driving, blood glucose (BG) was maintained through an insulin clamp to levels ∼100 mg/dl. For HG driving, BG was gradually decreased from an initial target of ∼70 mg/dl to ∼50 mg/dl. Self-treatment of hypoglycemia was scored as “present “ or “absent” depending on whether the patient drank a “sugar soda” (a non-sugar, diet soda) to correct perceived HG symptoms during simulated driving. Recent driving mishaps were scored according to the patients' self-report of driving accidents or moving citations within the past two years. Multi-channel digital EEG was acquired was acquired during simulated driving trials. Relative EEG spectral power data was averaged into five-minute epochs to correspond with BG sampling rate. The difference between relative alpha and theta powers at the central vertex (ATDiff) was used to quantify vigilance. Outcomes were modeled using binary logistic regression with self-treament (+/−) and driving mishaps (+/−) as dependent variables, and BG and ATDiff during EG and HG epochs as independent variables.∼

Results: Thirty-two subjects completed both EG and HG trials. 60% self-treated HG during HG driving, and 68% admitted to recent driving mishaps. Driving history and self-treatment were not significantly associated (Chi square = 2.54, p = 0.28). Subjects overall showed increasing delta power with decreasing BG. In the case of driving mishaps, increasing ATDiff during both EG (B = 4.13, p = 0.008) and HG (B = 3.41, p = 0.02) driving trials predicted driving mishaps. In the case of self-treatment of HG, decreasing ATDiff during HG driving did not significantly predict self-treatment. However, values of ATDiff increasingly diverged with worsening BG, with self-treating subjects experiencing decreasing ATDiff.

Conclusions: Task-acquired EEG shows worse vigilance in diabetics who successfully recognize and treat HG. EEG vigilance is impaired in diabetics who have no self-reported history of driving mishaps. These findings suggest that mechanisms independent of EEG vigilance may allow compensation for hypoglycemia-associated performance.

1 Rajesh RamachandranNair, 1 Jacob M. Muthalaly, 1 Manoj Parameswaran, and 1 Vellani T. Ajithkumar ( 1 Neurology, Medicine, Pediatrics, Medical College Hospital, Kozhikode, India )

Rationale: Both acute disseminated encephalomyelitis (ADEM) and Herpes Simplex encephalitis (HSE) can present with alteration in sensorium and seizures. Access to immediate MRI brain may not be available in developing world. Early diagnosis and treatment determines outcome in HSE. Steroids, used to treat ADEM, can be harmful in HSE. Aim of our study was to compare the early electroencephalographic (EEG) findings in ADEM and HSE to aid dignostic differentiation.

Methods: Patients presenting with seizures and altered sensorium, who were subsequently diagnosed either as ADEM or HSE based on MRI brain and virological studies (CSF-PCR and rising titer of CSF HSV antibody), were included in the study. EEGs performed within 72 hours after the onset of neurological symptoms were analysed for background activity, interictal epileptiform activity (including periodic lateralized epileptiform discharges-PLEDs) and ictal discharges. Incidence of abnormal EEG patterns in ADEM and HSE were compared using X2/Fisher's test.

Results: EEGs were performed within 72 hours in 21 patients (age 2–56 years, mean 18.6 yrs) with ADEM and 32 patients (age 4–72 years; mean 23.7 yrs) with HSE. All 53 EEGs were abnormal due to background slowing or epileptic discharges.

Table 2. Early EEG findings in ADEM and HSE
VariableADEMHSEp value
Total No of patients during study period3937-
No of patients who satisfied clinical inclusion criteria2334-
No of patients who satisfied clinical and EEG criteria2132-
Background slowing1932-
Only Focal 714-
Only Fronto-temporal 3140.02 
Extra-Frontotemporal 4 00.02 
Epileptiform discharges 9240.007
Generalised 28-
Focal/Multifocal 716-
Only Frontotemporal 3160.007
Extra-Frontotemporal 4 00.02 
PLEDs 2100.05 
BiPLEDs 040.12 
Seizures 1 70.08 

Background slowing was seen (generalised & focal) in both ADEM and HSE. But frontotemporal slowing occurred more frequently in HSE. Slowing confined to extra-frontotemporal area was seen only in ADEM. Same pattern was seen with interictal epileptiform discharges also. PLEDs were seen in ADEM also (1-frontoparietal and 1-frontocentrotemporal). In HSE, all the PLEDs were frontotemporal.There was a trend towards higher incidence of electrographic seizures in HSE.

Conclusions: Focal background slowing and epileptiform discharges (including PLEDs) confined to frontotemporal areas suggested the possibility of HSE. Presence of focal slowing and interictal epileptiform discharges confined to extra-frontotemporal areas suggested the possibility of ADEM as these features were not seen in HSE. Based on early EEG features, ADEM and HSE with similar clinical presentation can be differentiated in selected cases. This could be important in developing countries where facility for immediate MRI brain is not readily available.

1 Eishi Asano, 1 Robert Rothermel, 1 Michael Behen, 1 Otto Muzik, 1 Diane C. Chugani, 1 Carol Pawlak, 2 James Janisse, 2 Joel Ager, and 1 Harry T. Chugani ( 1 Pediatrics, Neurology, Psychiatry, Radiology, Electroneurodiagnostics, Children's Hospital of Michigan, Wayne State University, Detroit, MI ; and 2 Center for Health Effectiveness Research, Wayne State University, Detroit, MI )

Rationale: Prediction of cognitive function has been difficult in children with epilepsy, since cognitive dysfunction in epilepsy involves multiple causes. We determined whether occipital alpha peak frequency during quiet wakefulness can independently predict cognitive function in children with focal epilepsy.

Methods: A total of 33 children (age: 6–17 years; 20 boys) with focal epilepsy were included in the study. None of the subjects had visual, hearing, or motor deficits. Children with idiopathic childhood epilepsy with occipital paroxysms or central-temporal spikes were also excluded. Occipital alpha peak frequency during quiet wakefulness with eye closure and interictal spike frequency during wakefulness were quantitatively measured by scalp EEG. Interictal spike frequency was categorized into four groups (0, 0–1, 1–10, and >10 per minute; Asano et al, 2004). A hierarchical multiple regression analysis was performed to determine whether age-adjusted occipital alpha peak frequency can predict the full-scale IQ independently from spike frequency category and the number of antiepileptic medications taken by the patient.

Results: An analysis of the zero-order correlations revealed that each of the three predictors (age-adjusted occipital alpha peak frequency, spike frequency category and the number of anti-epileptic medications) is about equally correlated to full-scale IQ and each of these predictors accounted for 20%−23% of the variance (p < 0.01 for all three predictors). In a hierarchical multiple regression analysis, the categorical measure of spike frequency and the number of medications used were entered in the first step and accounted for 31% of the variance in the IQ measure (F(2, 30) = 6.77, p = 0.004). In a second step of the analysis the age-adjusted occipital alpha peak frequency was entered and accounted for an additional 13% of the variance (F(1, 29) = 6.49, p = 0.016).

Conclusions: Occipital alpha peak frequency during quiet wakefulness is an important factor independently predicting the full-scale IQ in children with focal epilepsy even after controlling for both interictal spike frequency and the number of anti-epileptic medications. Quantitative measurement of occipital alpha peak frequency and interictal spike frequency data extracted from a single video-EEG session combined with the number of antiepileptic drugs may provide important information for prognosis of overall cognitive function in children with focal epilepsy. (Supported by NIH: K23NS047550.)

1 Michelle J. Shapiro, 2 D. Barry Sinclair, and 3 S. Nizam Ahmed ( 1 Department of Neurology, University of Saskatchewan, Saskatoon, SK, Canada ; 2 Department of Pediatric Neurology, University of Alberta, Edmonton, AB, Canada ; and 3 Department of Neurology, University of Alberta, Edmonton, AB, Canada )

Rationale: Intermittent photic stimulation (IPS) is used routinely as an activation procedure for electroencephalographs (EEGs) in the pediatric population. Photosensitivity is relatively more prevalent in children than in adults. The purpose of this study was to determine the proportion of pediatric epilpetiform EEGs in which photic stimulation altered the overall impression of the EEG, thus adding additional information to the baseline.

Methods: One hundred consecutive EEGs performed at the University of Alberta Hospital on patients 18 years or younger, that had unequivocal epileptiform abnormalities and activation by IPS were reviewed. Demographics including patient age, sex, clinical diagnosis, and reason for the EEG were collected. A determination was made as to whether IPS altered the overall impression of the EEG.

Results: There were 62 males and 38 females. Forty-two patients were noted to have photic driving. Fifty patients had evidence of epileptic activity during photic stimulation, which was determined not to be linked to the stimulus. In 3 patients, there was epileptic activity during IPS with possibile association to the stimulus. In 3 patients IPS provided information not available otherwise. Out of these 3 patients, one had photosenstive Juvenile Myoclonic Epilepsy, the second had occipital spikes that generalized during IPS, and the third patient had evidence of occipital dominant spikes and polyspikes during IPS, a finding not seen on the baseline EEG. The finding of photosensitivity did not alter the management plan for any of these patients.

Conclusions: This study demonstrates a prevalence of 3% photosensitivity amongst pediatric patients with known epileptic abnormalities on their EEG. These numbers are lower than those described by Wolf and Goosses (1), who reported a prevalence of 10%, in their review of of more than 1000 patients. The discrepancy could be related to the difference in sample size or merely to genetic heterogenity in the different samples. Interestingly all our photosensitive patients were females supporting the previously published literature on higher prevalence of photosensitivity in females. Considering the prevalence of photosensitivity in the epileptic population and the benign nature of the procedure, we support this activation procedure in routine practice.


1. Wolf P, Goosses R. Relation of photosensitivity to epileptic syndromes. J Neruol Neruosurg Psychiat 1986;49: 13861391.

1 Ulrich Specht, 2 Matthias Hoppe, 1 Rupprecht Thorbecke, 2 Alois Ebner, and 1 Bernd Pohlmann-Eden ( 1 Rehabilitation Unit, Bethel Epilepsy Center, Bielefeld, Germany ; and 2 EEG-Video-Department, Bethel Epilepsy Center, Bielefeld, Germany )

Rationale: The uncertainty of increased risks of seizure provocation due to conditions of the workplace is considered to contribute to the high rates of unemployment in patients with epilepsy. In Germany, the diagnosis of epilepsy frequently leads to restrictions concerning workplaces with computer monitor displays, even in patients without photosensitivity. Current knowledge about photosensitive patients does not give evidence for an increased risk of seizures in such a setting, e.g. because of the high refresh frequencies (≥75 Hz) of modern PC monitor displays; however studies with EEG registration in patients working at a PC monitor display have not been undertaken.

Methods: Patients 16 years and older with a documented photoparoxysmal response (PPR), i.e. epileptiform EEG discharges (ED) provoked by intermittent photic stimulation (IPS), were included. They underwent a standardized EEG procedure with simultaneous video recording. It comprised three evaluation periods in randomized order, each lasting 10 minutes: 1) baseline period (resting condition), 2) PC-test period (reading and editing a text at a personal computer using a word processing program), 3) control period (reading and editing a similar text printed on a white paper). For the PC-test period, a commercially available personal computer with a 17 inches cathode ray tube (CRT) monitor display (resolution 800x600 dpi, refresh rate of 85 Hz) was used. The number and duration of ED during each of the evaluation periods were assessed by two blinded examiners. In addition, standard IPS was performed in order to document present PPR.

Results: Sixteen patients were evaluated. During the procedure, four did not show PPR on IPS and thus were excluded. In 8 of the remaining 12 patients, ED occurred at least during one of the three evaluation periods. There was no statistical significant increase in number and duration of ED in the PC-test period compared to the control and baseline periods. ED tended to occur more frequently in the baseline period. No clinical seizure was recorded during the study.

Conclusions: In accordance with the current literature using IPS, this study does not give evidence for an increased risk of seizure precipitation in photosensitive patients when using a word processing program displayed on a commercially available 17′′ CRT PC monitor display. Thus, restrictions concerning workplaces with computer monitor displays such as office work are not warranted. (Supported by UCB Pharma GMBH, Kerpen, Germany.)

1 Ian A. Stein, and 1 Geetha Chari ( 1 Neurology, SUNY-Health Science Centers of Brooklyn, Brooklyn, NY )

Rationale: To characterize EEG abnormalities found on EEGs of pediatric psychiatric inpatients.

Methods: A retrospective analysis of all EEGs ordered by the Kings County Hospital Center inpatient pediatric psychiatric unit between July 2000 and December 2004 was performed. EEGs were performed using the guidelines of the ACNS. Clinical information was restricted to the history provided on the EEG request forms. The EEGs were interpreted by a neurologist board certified in clinical neurophysiology. Reports were written according to ACNS guidelines. The results of this study were obtained by analyzing information from these reports

Results: Between January 2000 and December 2004 a total of 411 EEGs were performed on patient admitted to the pediatric psychiatry service at Kings County Hospital. 24 patients had repeat studies. 387 EEGs were included into the analysis. There were 282 males and 105 females. The age of patients ranged between 5 and 17 years, with an average age of 11.5 years. 134 patients (35%) had abnormal EEGs 93 patients had a non-epileptiform abnormality (e.g. background slowing, focal slowing) alone, while 26 patients had only epileptiform activity. 20 patients had both epileptiform and non-epileptiform abnormalities. Of the 46 (11.9%) patients with epileptiform activity, 26 patients had focal discharges, and 18 patients had generalized discharges. 2 patients had both focal and generalized activity.

Conclusions: There is a greater incidence of EEG abnormalities in our study when compared with previously published data on normal children. While the rates of abnormal EEG findings in our population were similar to those seen in the adult psychiatric patients, epileptiform activity was more frequent in pediatric patients. Further studies are required to determine the clinical significance of these findings. (Supported by SUNY-Health Sciences Center of Brooklyn, Department of Neurology.)

1 James X. Tao, 1 Amit Ray, 1 Baldwin Maria, 1 Sona Shah, 1 Susan Hawes-Ebersole, and 1 John S. Ebersole ( 1 Department of Neurology, The University of Chicago, Chicago, IL )

Rationale: To determine the cerebral EEG substrates of scalp EEG seizure patterns, and thus to assess the value of scalp ictal recording in the lateralization and localization of seizure onset zones in patients with temporal lobe epilepsy.

Methods: We recorded simultaneously 26 channels of scalp EEG with sub-temporal supplementary electrodes and 46–98 channels of intracranial EEG in presurgical candidates with temporal lobe epilepsy. Subdural electrodes covered the anterior two-thirds of the temporal lobe in all patients. We studied the influence of intracranial EEG source area and synchrony at seizure onset on corresponding scalp EEG correlates. Eighty-six simultaneous intracranial- and scalp-recorded seizures from 23 patients were evaluated.

Results: Thirty-four intracranial ictal discharges (40%) from 9 patients (39%) had sufficient source area (namely > 10 cm2) and synchrony at seizure onset to produce a nearly simultaneous focal scalp EEG ictal pattern. Forty-one intracranial ictal discharges (48%) from 10 patients (43%) gradually achieved the necessary source area and synchrony over several seconds to generate a scalp EEG ictal pattern. These scalp rhythms were lateralized, but less localizable. Eleven intracranial ictal discharges (13%) from 4 patients (17%) recruited the necessary source area, but lacked sufficient synchrony to generate clearly localized or lateralized scalp EEG ictal patterns.

Conclusions: Sufficient source area and synchrony are mandatory cerebral EEG substrates for generating scalp-recordable ictal EEG patterns. When the necessary source area and synchrony are achieved at or near seizure onset, the scalp EEG pattern can localize the cerebral seizure onset zone. However, if achieving necessary source area and synchrony requires several seconds, during which time more distant propagation may occur, the resultant scalp ictal pattern may only lateralize, but not localize, the seizure onset zone. (Supported by Brain Research Fundation.)

1 Mark D. Holmes, 2,3 Don M. Tucker, 2 Micah Brown, 1,4 Russell P. Saneto, and 5 Jeffrey G. Ojemann ( 1 Neurology, University of Washington, Seattle, WA ; 2 Electrical Geodesics, Inc., Eugene, OR ; 3 Psychology, University of Oregon, Eugene, OR ; 4 Division Pediatric Neurology, University of Washington, Seattle, WA ; and 5 Neurological Surgery, University of Washington, Seattle, WA )

Rationale: We describe a case report of a subject with intractable extratemporal epilepsy who underwent pre-operative longterm EEG-video monitoring (LTM) with 128 channel scalp recordings, following completion of standard (sparse array) LTM. She underwent subdural, intracranial grid and strip LTM recordings, after standard methods failed to provide adequate seizure localization. We compare the results of seizure onsets predicted on the basis of dense array scalp EEG to that obtained from the invasive recordings.

Methods: A 13 year old girl, a potential surgical candidate, presented with medically refractory daily complex partial seizures. Clinical exam and MRI studies were unrevealing. Standard EEGs demonstrated abundant interictal discharges over left posterior quadrant, particularly over left parietal regions (P3). Standard scalp LTM recorded seizures that were poorly localized over the left posterior quadrant. One of her habitual clinical seizures was captured during 48 hours of dense array LTM. Using a linear inverse method of EEG source analysis (LAURA), seizure onset and propagation patterns were displayed on a standard MRI model, prior to invasive studies. Invasive LTM studies included subdural grid and strips placed over the left, and subdural strips placed over the right, posterior quadrants.

Results: Analysis of dense array data predicted that seizure onset occurred over the left inferior posterior temporal-occipital regions, followed by rapid spread to right lateral posterior temporal cortex, left lateral posterior temporal cortex, and subsequently to the left superior parietal region. These predictions, including both seizure onset and propagation patterns, were confirmed on analysis of invasive ictal EEG recordings. Surgical resection was performed, based on the invasive EEG data.

Conclusions: Dense array EEG, used in conjunction with source analysis and a realistic MRI model, has the potential to localize seizure onsets and propagation patterns, when standard noninvasive methods fail. This is the first report demonstrating that dense array LTM studies are possible, and that predictions based on these studies are confirmed on the basis of intracranial EEG recordings. Future research will determine if the use of these new techniques will reduce the need for invasive LTM in the surgical evaluation of subjects with difficult epilepsy.