ORIGINAL RESEARCH
Infraslow EEG Changes in Infantile Spasms Kenneth A. Myers,* Luis E. Bello-Espinosa,*†‡ Xing-Chang Wei,‡§ and Morris H. Scantlebury*†‡
Purpose: Infantile spasms (IS) are a devastating epileptic encephalopathy syndrome of infancy. Analysis of infraslow EEG activity (ISA) has shown potential in the presurgical evaluation of patients with epilepsy and in differentiating between focal and generalized epilepsy syndromes. Infraslow EEG activity analysis may provide insights into the pathophysiology of some difficult-to-treat epilepsy syndromes, such as IS. To our knowledge, there are no published reports describing ISA in patients with IS. The purpose of this study was to describe ictal patterns of ISA in patients with IS and to correlate with clinical data. Methods: EEG recordings of all cases of IS in the past 10 years at the Alberta Children’s Hospital were reviewed. Inclusion criteria were a technically adequate video EEG recording that captured at least one spasm. For each patient, the first 10 confirmed spasms were examined. Spasms were evaluated for changes in ISA, which were either generalized, lateralized, or absent ISA (g-ISA, l-ISA, or n-ISA, respectively). Results were correlated with treatments, clinical course, and information pertinent to likely etiology of the IS. Results: A total of 77% of spasms were associated with ISA; 57% with g-ISA, 20% l-ISA, and 21% n-ISA. All patients with exclusively g-ISA showed at least a partial response to initial therapy, while this was the case in 66.7% of those with at least some l-ISA and 50% of those with exclusively n-ISA. Other seizure types occurred in 60% of patients with exclusively g-ISA versus 83% with some l-ISA and all patients with exclusively n-ISA. Conclusions: Ictal ISA was observed in the majority of IS. Trends were observed suggesting that the presence of exclusive g-ISA changes may be a positive prognostic factor in IS. Key Words: Infantile spasms, Infraslow oscillations, Epileptic encephalopathy, thalamus. (J Clin Neurophysiol 2014;31: 600–605)
I
nfantile spasms (IS) are a rare, but often devastating epileptic encephalopathy of infancy with an incidence of 1 to 2 per 10,000 live births (Brna et al., 2001; Sidenvall and Eeg-Olofsson, 1995). Infantile spasms usually present before the first year of life and, even in patients with focal brain abnormalities, usually involve bilateral flexion, extension, or mixed flexion–extension spasms. The EEG patterns in IS are well characterized with conventional frequency range (1–70 Hz) EEG. Eleven different ictal subtypes have been described, involving combinations of slow waves, sharp and slow waves, attenuation, fast activity, and rhythmic slow waves with generalized attenuation after the seizure being the most common
From the *Section of Neurology, Department of Paediatrics, Faculty of Medicine, University of Calgary, Calgary, Canada; †Department of Clinical Neurosciences, Faculty of Medicine, University of Calgary, Calgary, Canada; ‡Alberta Children’s Hospital Research Institute for Child and Maternal Health, Calgary, Canada; and §Department of Radiology, Faculty of Medicine, University of Calgary, Calgary, Canada. Address correspondence and reprint requests to Morris H. Scantlebury, MD, 278 HMRB, 3300 Hospital Drive, Calgary, AB T2N 4N1, Canada; e-mail: morris. [email protected]. Copyright Ó 2014 by the American Clinical Neurophysiology Society
ISSN: 0736-0258/14/3106-0600
600
(Kellaway et al., 1979). Hypsarrhythmia is the classic interictal EEG pattern, defined as a high amplitude multifocal spike wave discharges with a disorganized background. Although hypsarrhythmia sometimes improves with arousal or immediately after a cluster of spasms (Hrachovy et al., 1984), in many patients, it occupies most of the EEG recording. Some have proposed that hypsarrhythmia be considered a form of status epilepticus accounting in large part for the abrupt development regression seen in these patients. There is now emerging evidence that an extended evaluation of the EEG beyond the conventional frequencies between 1 and 70 Hz may have significantly impact the treatment of patients with IS. For example, frequencies above 70 Hz, known as high frequency oscillations, have been studied in IS and have shown potential utility in making treatment decisions including presurgical planning (Iwatani et al., 2012; Nariai et al., 2011). Although these studies support a role for the careful evaluation of high frequency oscillations in patients with IS, analysis of frequencies below the conventional frequency range may also be of benefit. Infraslow activity (ISA) EEG analysis has shown potential in differentiating absence and focal onset seizures (Rodin et al., 2008) as well as preepilepsy surgical planning (Modur et al., 2012). Despite these potential applications, there have been no published reports, to our knowledge, involving ISA analysis in early childhood epilepsy in general and IS specifically. This is an important gap in the literature because recent studies have shown that ISA has the potential to be useful in seizure characterization and localization and may eventually help in guiding epilepsy management by helping in the choice of medications and in the identification of epileptogenic cortical regions in preparation epilepsy surgeries (Constantino and Rodin, 2012; Modur et al., 2012; Rampp and Stefan, 2012; Rodin and Modur, 2008; Rodin et al., 2008, 2009). This study is a single institution, retrospective analysis of ISA during infantile spasms and correlates the findings with the patients’ clinical characteristics.
METHODS The research protocol was approved by the University of Calgary’s Conjoint Health Research Ethics Board. Two attending epileptologists (L.E.B. and M.H.S.) performed a search of the pediatric neurology clinic’s patient records and generated a list of all patients in the past 10 years with a diagnosis of infantile spasms. This list, with no additional clinical information, was provided to a pediatric neurology resident (K.A.M.) who reviewed the EEG recordings masked to the patient demographics. The results of the analysis were independently analyzed by L.E.B. and M.H.S. Only patients younger than 2 years of age at the time of the recording were included to minimize the confounding effects of brain maturation on ISA patterns. Minimum requirements for inclusion were a technically adequate video EEG recording that captured at least one infantile spasm, confirmed on the official EEG
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
report. Events suspicious for spasms were identified by technician or epileptologist notes on the recording. All EEG analysis was conducted using the XLTek NeuroWorks system (Natus Medical Inc, San Carlos, California). These were first examined using a bipolar montage and conventional frequency range (1–70 Hz). Events were included if there was clinical evidence of a spasm on the video recording, correlated with classic EEG findings. These events were considered confirmed spasms and analyzed for ISA changes. Once an event was confirmed as a spasm, ISA was analyzed with an ipsilateral ear referential montage using a band pass filter with a frequency range of 0.01 to 0.1 Hz (Rodin et al., 2008). Spasms were analyzed at a time base of 15 mm/s and sensitivity that varied from 10 to 50 mV/mm. For each patient, the first 10 confirmed spasms were analyzed. The recording was examined under these new settings for ISA 30 seconds before and after the spasm. Although the study was focused on ictal changes, this relatively wide time window was thought appropriate because waveforms in the infraslow range have quite long periods (10–100 seconds). Changes were initially recorded as present or absent. If present, the changes were classified as generalized (symmetric) or lateralized, with laterality defined as the amplitude in one hemisphere being more than 50% the contralateral side (Figs. 1 and 2). Patients were then grouped as those having: 1. Exclusively generalized or absent ISA (G-ISA), 2. At least one event with lateralized ISA (L-ISA), or 3. No ISA change with any spasms (N-ISA). We also recorded if the spasm occurred in the context of a cluster (cluster defined as 2 or more spasms occurring in series with an interval of .30 seconds between spasms). Once the EEG analysis was completed for all patients, pediatric neurology clinic notes and
Infraslow EEG in Infantile Spasms
brain imaging reports were reviewed. Imaging results, treatments, clinical course over the 2 years after spasm onset, and information pertinent to likely etiology of the infantile spasms were recorded. Additionally, a pediatric neuroradiologist (X-C.W.), masked to the EEG results, reviewed each patient’s brain MRI and report specifically on any thalamic pathology.
Statistical Analysis Chi-squared analysis was used to compare proportions between groups. Kruskal–Wallis one-way analysis of variance on ranks was used for multiple group comparisons where appropriate.
RESULTS Patient Characteristics The EEG recordings of 136 patients were reviewed, and 13 patients met criteria for inclusion. The reasons for exclusion were (1) spasms were not captured during the recoding and (2) video was not available to analyze the spasms captured for ISA. When known, the mean age of spasm onset was 7.0 6 4.3 (mean 6 SD) months. There were no differences in the age of onset of spasms between the groups (P ¼ 0.391; Kruskal–Wallis one way analysis of variance on ranks). IS was due to a symptomatic etiology in 12/13 (92.3%) patients. The etiologies were diverse and are listed in Table 1. Vigabatrin was used as an initial therapy for spasms in 11/13 patients (84.6%), whereas the remaining 2 patients received adrenocorticotropic hormone. The mean age at EEG for analysis was 8.8 6 5.3 months.
FIG. 1. Generalized ISA change. Patient 5: 8 months old with cortical dysplasia. EEG recording settings, 20 mV/mm and 10 mm/s. Arrow denotes clinical spasm. Copyright Ó 2014 by the American Clinical Neurophysiology Society
601
K. A. Myers et al.
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
FIG. 2. Lateralized ISA change. A, Individual spasm. Patient 9: 4 months old with hypoxic ischemic encephalopathy. EEG recording settings, 30 mV/mm and 10 mm/s. B, Cluster of spasms. Patient 3: 5 months old with mitochondrial disorder. EEG recording settings, 50 mV/mm and 5 mm/s. Arrows denote clinical spasms.
Spasms Characteristics There were a total of 101 spasms analyzed, of which 55/101 (54%) occurred in the context of a cluster. Of the total number of spasms, 78 (77%) were associated with ISA, and 58/101 (57%) of spasms were associated with generalized ISA (g-ISA), 20/101 602
(20%) with lateralized (l-ISA), and 23/101 (23%) with no ISA (n-ISA). After completion of the study, a less formal analysis was done to investigate whether lateralization in the infraslow range correlated with similar asymmetry in the conventional frequency Copyright Ó 2014 by the American Clinical Neurophysiology Society
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
TABLE 1.
Clinical Characteristics and EEG Results Etiology
Age of Onset
Age at EEG (months)
Number of Spasms Analyzed
MCD Meningitis Mitochondrial IVH MCD Genetic/IVH Meningitis/IVH TS HIE Genetic/Metabolic MCD MCD Genetic/Metabolic
14 5 3 Unclear 6 12 9 12 2.5 Unclear 5 2 5
14 5 5 22 8 12 9 12 4 3 5 10 5
9 1 8 7 3 3 10 10 10 10 10 10 10
Patient 1 2 3 4 5 6 7 8 9 10 11 12 13
Patient 1 2 3 4 5 6 7 8 9 10 11 12 13
Infraslow EEG in Infantile Spasms
Spasms With l-ISA (%) 2 0 8 4 0 0 4 0 1 1 0 0 0
(22) (0) (100) (57) (0) (0) (40) (0) (10) (10) (0) (0) (0)
Spasms With n-ISA (%) 0 0 0 1 2 0 0 0 0 0 0 10 10
(0) (0) (0) (14) (67) (0) (0) (0) (0) (0) (0) (100) (100)
Overall ISA Pattern L-ISA G-ISA L-ISA L-ISA G-ISA G-ISA L-ISA G-ISA L-ISA L-ISA G-ISA N-ISA N-ISA
Total Number Spasms ISA (%) 9 1 8 6 1 3 10 10 10 10 10 0 0
Response to Initial Treatment (Vigabatrin/ACTH) Partial (Vigabatrin) Complete (Vigabatrin) Complete (Vigabatrin) No Response (Vigabatrin) Partial (Vigabatrin) Partial (ACTH) Partial (ACTH) Complete (Vigabatrin) Complete (Vigabatrin) No Response (Vigabatrin) Partial (Vigabatrin) No Response (Vigabatrin) Partial (Vigabatrin)
(100) (100) (100) (86) (33) (100) (100) (100) (100) (100) (100) (0) (0)
Spasms With g-ISA (%) 7 1 0 2 1 3 6 10 9 9 10 0 0
(78) (100) (0) (29) (100) (100) (60) (100) (90) (90) (100) (0) (0)
Development of other Seizure Types (1/2) 1 (Tonic, Absence) 1 (Myoclonic, Focal dyscognitive) 1 (GTC, Myoclonic) 1 (Focal clonic, Focal dyscognitive) 2 1 (Myoclonic, Focal clonic) 1 (Focal dyscognitive) 1 (Focal dyscognitive) 2 1 (Myoclonic, Tonic, Focal clonic) 2 1 (Focal) 1 (Myoclonic)
ACTH, adrenocorticotropic hormone; G-ISA, patients with exclusively g-ISA 6 n-ISA with spasms; HIE, hypoxic ischemic encephalopathy; IVH, intraventricular hemorrhage; L-ISA, patients with at least one spasm with l-ISA; MCD, malformation of cortical development; N-ISA, patients with exclusively n-ISA with spasms; TS, tuberous sclerosis.
range. Although asymmetry in the conventional range was observed in some cases, there was no clear relation to the patterns seen with ISA.
Patient-Specific Results
A mean of 7.8 6 3.3 spasms were analyzed per patient with no significant differences in the number of spasms analyzed between the groups. Only 2/13 patients (15%) did not show clear ictal ISA with any of their captured spasms (N-ISA). Of the remaining patients with clear ictal ISA, 6/13 (46%) had evidence of lateralization with at least one of their spasms (L-ISA). The remaining 5 patients (38%) were classified as having exclusively g-ISA 6 n-ISA with their spasms (G-ISA). All of the G-ISA patients showed at least a partial response to initial therapy, while this was the case in 4/6 L-ISA patients (66.7%) and 1/2 N-ISA patients (50%). Interestingly, two symptomatic patients received adrenocorticotropic hormone as an initial therapy, one with G-ISA and the other with L-ISA, with both showing a partial response to treatment. There were no significant differences between the groups regarding response to treatment. Regarding long-term outcome, 3/5 (60%) patients with G-ISA developed other seizure types compared with 5/6 (83%) with L-ISA and 2/2 (100%) with N-ISA. These differences were not statistically significant. Copyright Ó 2014 by the American Clinical Neurophysiology Society
MRI Analysis of Thalamic Pathology
The thalamic MRI findings are summarized in Table 2. In summary, thalamic damage was seen in 8/13 (62%) patients and was clearly asymmetric in only 2 patients. Both (100%) of these patients showed L-ISA while this lateralized pattern was seen in 2/6 (33%) patients with symmetric thalamic pathology and 2/5 (40%) patients with no apparent thalamic abnormality.
DISCUSSION
To our knowledge, this is the first study describing the ictal ISA patterns in patients with IS. The major finding is that ictal ISA occurs in the majority of patients with IS. ISA may be generalized or lateralized, with many patients showing a mix of both patterns and some having no apparent change whatsoever. Ictal ISA was recorded in more than 80% of spasms in this study. Taking into consideration the ictal baseline shifts seen before the onset of focal onset seizures (Rodin et al., 2008), we propose that ISA recorded at the time of an infantile spasm is pathologic and related to thalamic dysfunction. There are several lines of evidence supporting a role for thalamic dysfunction in the generation of ictal ISA in patients with IS. First, the thalamus is thought to play a critical role in the generation of ISA (Hughes et al., 2011). One recent study showed that damage to the thalami and basal ganglia in patients with 603
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
K. A. Myers et al.
TABLE 2. Patient
MRI Analysis of Thalami Pattern of ISA Expression
1
L-ISA
2
G-ISA
3
L-ISA
4
L-ISA
5
G-ISA
6
G-ISA
7
L-ISA
8
G-ISA
9
L-ISA
10
L-ISA
11
G-ISA
12
N-ISA
13
N-ISA
General MRI Findings Bilateral ventriculomegaly and colpocephaly; agenesis of corpus callosum Diffuse cerebral atrophy; delayed myelination Diffuse cerebral atrophy Extensive white matter abnormalities, ventriculomegaly and multiple areas of porencephaly Restricted diffusion, mainly in midbrain and globi pallidi Exvacuo hydrocephalus and right frontoparietal porencephalic cyst; enlarged posterior fossa Exvacuo hydrocephalus; multifocal encephalomalacia Bilateral cortical and subcortical tubers; subependymal nodules Bilateral areas of restricted diffusion; left occipitotemporal infarct Normal at 11 weeks; later, generalized cerebral atrophy and T2 hyperintensity in globi pallidi, diencephalon, and dentate nucleus Ventriculomegaly; corpus callosum dysgenesis; pachygyria; extensive white matter abnormalities; multicystic encephalomalacia; hypoplastic cerebellar vermis Aqueductal stenosis with hydrocephalus; thin corpus callosum; absent septum pellucidum; multiple subependymal heterotopias Bilateral cytotoxic edema (developed after vigabatrin)
Thalamic Abnormality (1/2)
Thalamic Abnormality (Symmetric or Asymmetric)
Type of Thalamic Abnormality
2
2
2
1
Symmetric
1
Symmetric
1
Asymmetric (R . L)
Initially restricted diffusion; later, atrophy and gliosis Restricted diffusion (resolved on subsequent study) Gliosis and volume loss
1
Symmetric
Restricted diffusion
1
Symmetric
Mild volume loss
1
Asymmetric (L . R)
2
2
1
Symmetric
2
2
Restricted diffusion consistent with infarct/ischemia; subtle hemorrhage 2
2
2
2
2
2
2
1
Symmetric
Subtle restricted diffusion; later, gliosis and volume loss 2
Diffuse cytotoxic edema (developed after vigabatrin)
G-ISA, patient had exclusively generalized or absent infraslow activity with spasms; L-ISA, patient had signs of lateralization of infraslow activity with at least some spasms; N-ISA, patient did not change in infraslow activity with any spasms.
hypoxic ischemic encephalopathy was predictive of progression to IS (Gano et al., 2013). The tendency for spasms to occur around the sleep–wake transition, and for hypsarrhythmia to become more prominent during sleep, are both suggestive of thalamocortical dysfunction. In patients with Lennox–Gastaut syndrome, a highly related catastrophic epilepsy, there is a higher recorded frequency of cyclical alternating pattern of non-REM sleep, and both the generalized spike and wave pattern and seizures occur more frequently during the active phase of the cyclical alternating pattern (Eisensehr et al., 2001). This is an important observation as cyclical alternating 604
pattern is characterized by slow oscillations of ,0.1 Hz, occurring with a periodicity of 20 to 35 seconds in stages 2 and 3 of natural non-REM sleep (Terzano and Parrino, 2000) and is thought to be linked to thalamocortical volleys given its widespread EEG expression and due to its ability to modulate diffuse ascending and descending pathways (Eisensehr et al., 2001; Terzano et al., 1991). Given the probable relationship between thalamic dysfunction and ISA expression, asymmetric damage to the thalami of patients with IS is a possible explanation for the l-ISA observed in some patients in this study. Patients with exclusively g-ISA can be Copyright Ó 2014 by the American Clinical Neurophysiology Society
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
speculated to have more symmetric thalamic involvement. Of note, two patients had n-ISA correlating with spasms, which may indicate relative preservation of “normal” thalamic function. Based on this theory, the degree and symmetry of thalamic involvement could be inferred by an analysis of the ISA activity, which could have both prognostic and therapeutic significance. Our MRI analysis of thalamic abnormalities was somewhat supportive of this theory; however, autopsy analysis or analysis of other surrogate electrophysiological markers for thalamic function, such as sleep spindles (Ho et al., 2014), would likely be the best approach to assess thalamic pathology and any associated asymmetry thereof. Notably, a recent article reported that symmetrical thalamic hyperperfusion on ictal SPECT (single-photon emission computerized tomography) during infantile spasms was a positive prognostic factor, when compared with asymmetric hyperperfusion (Haginoya et al., 2013). The functional mechanisms underlying ISA are currently poorly understood. However, electrophysiological recordings conducted in in vitro thalamic preparations suggest that ISA is likely generated through the activation of G-coupled inward rectifying potassium channels (Hughes et al., 2011), which are the downstream effectors of G-coupled receptors, including the GABAB receptor (Luscher et al., 1997). Notably, diffuse overexpression of G-coupled inward rectifying potassium 2 channels has been described in the brains of the Ts65Dn mouse model of Down syndrome, a model that forms the basis of an important animal model of IS induced by exposing the Ts65Dn mice to the GABAB agonist gammabutyrolactone (Cortez et al., 2009). Together, these results suggest that ictal ISA recorded in association with the spasms may in part be due to abnormal G-coupled inward rectifying potassium channel expression in the thalamus among other regions of the brain. Although, not statistically significant, the results of this study suggest that the presence of exclusively g-ISA may have a positive prognostic factor in IS. However, significant caveats are the small sample size, which included a high number of individuals with symptomatic IS. This latter factor precluded an analysis of the patterns observed in cryptogenic versus symptomatic cases. A second factor was the variability in treatment of IS, and the reliance on clinical notes to determine the extent of response to initial treatment. Furthermore, our decision to analyze only the first 10 spasms was somewhat arbitrary, and the possibility exists that substantially more patients might show signs of lateralized ISA if more spasms were analyzed. An optimal study would be conducted prospectively, including a standard minimum number of events, allowing for a consistent sampling of ISA patterns for each patient. In conclusion, this study demonstrates the exciting potential and need for further study, of ISA in childhood epilepsy in general and IS in particular. Overall, the results of our study demonstrate that
Copyright Ó 2014 by the American Clinical Neurophysiology Society
Infraslow EEG in Infantile Spasms
ISA changes frequently occur concurrently with IS and may provide prognostic information for these individuals. Much more work is required to better clarify the potential utility of this tool in the pediatric population, which will include defining normal patterns at different ages and clarification of the underlying physiology. REFERENCES Brna PM, Gordon KE, Dooley JM, Wood EP. The epidemiology of infantile spasms. Can J Neurol Sci 2001;28:309–312. Constantino T, Rodin E. Peri-ictal and interictal, intracranial infraslow activity. J Clin Neurophysiol 2012;29:298–308. Cortez MA, Shen L, Wu Y, et al. Infantile spasms and Down syndrome: a new animal model. Pediatr Res 2009;65:499–503. Eisensehr I, Parrino L, Noachtar S, et al. Sleep in Lennox-Gastaut syndrome: the role of the cyclic alternating pattern (CAP) in the gate control of clinical seizures and generalized polyspikes. Epilepsy Res 2001;46:241–250. Gano D, Sargent MA, Miller SP, et al. MRI findings in infants with infantile spasms after neonatal hypoxic-ischemic encephalopathy. Pediatr Neurol 2013;49:401–405. Haginoya K, Uematsu M, Munakata M, et al. The usefulness of subtraction ictal SPECT and ictal near-infrared spectroscopic topography in patients with West syndrome. Brain Dev 2013;35:887–893. Ho AW, Myers KA, Bello-Espinosa LE, Scantlebury MH. Infra-slow EEG activity and sleep spindle expression: Potential window into thalamic function in infantile spasms (Conference Abstract). Paper presented at: 13th International Child Neurology Congress; May 4–9, 2014; Iguazu Falls, Brazil. Hrachovy RA, Frost JD Jr, Kellaway P. Hypsarrhythmia: variations on the theme. Epilepsia 1984;25:317–325. Hughes SW, Lorincz ML, Parri HR, Crunelli V. Infraslow (,0.1 Hz) oscillations in thalamic relay nuclei basic mechanisms and significance to health and disease states. Prog Brain Res 2011;193:145–162. Iwatani Y, Kagitani-Shimono K, Tominaga K, et al. Ictal high-frequency oscillations on scalp EEG recordings in symptomatic West syndrome. Epilepsy Res 2012;102:60–70. Kellaway P, Hrachovy RA, Frost JD Jr, Zion T. Precise characterization and quantification of infantile spasms. Ann Neurol 1979;6:214–218. Luscher C, Jan LY, Stoffel M, et al. G protein-coupled inwardly rectifying K1 channels (GIRKs) mediate postsynaptic but not presynaptic transmitter actions in hippocampal neurons. Neuron 1997;19:687–695. Modur PN, Vitaz TW, Zhang S. Seizure localization using broadband EEG: comparison of conventional frequency activity, high-frequency oscillations, and infraslow activity. J Clin Neurophysiol 2012;29:309–319. Nariai H, Nagasawa T, Juhasz C, et al. Statistical mapping of ictal high-frequency oscillations in epileptic spasms. Epilepsia 2011;52:63–74. Rampp S, Stefan H. Ictal onset baseline shifts and infraslow activity. J Clin Neurophysiol 2012;29:291–297. Rodin E, Modur P. Ictal intracranial infraslow EEG activity. Clin Neurophysiol 2008;119:2188–2200. Rodin E, Constantino T, van Orman C, House P. EEG infraslow activity in absence and partial seizures. Clin EEG Neurosci 2008;39:12–19. Rodin E, Constantino T, Rampp S, Modur P. Seizure onset determination. J Clin Neurophysiol 2009;26:1–12. Sidenvall R, Eeg-Olofsson O. Epidemiology of infantile spasms in Sweden. Epilepsia 1995;36:572–574. Terzano MG, Parrino L. Origin and significance of the cyclic alternating pattern (CAP): review article. Sleep Med Rev 2000;4:101–123. Terzano MG, Parrino L, Spaggiari MC, et al. Discriminatory effect of cyclic alternating pattern in focal lesional and benign rolandic interictal spikes during sleep. Epilepsia 1991;32:616–628.
605
Infraslow EEG Changes in Infantile Spasms Kenneth A. Myers,* Luis E. Bello-Espinosa,*†‡ Xing-Chang Wei,‡§ and Morris H. Scantlebury*†‡
Purpose: Infantile spasms (IS) are a devastating epileptic encephalopathy syndrome of infancy. Analysis of infraslow EEG activity (ISA) has shown potential in the presurgical evaluation of patients with epilepsy and in differentiating between focal and generalized epilepsy syndromes. Infraslow EEG activity analysis may provide insights into the pathophysiology of some difficult-to-treat epilepsy syndromes, such as IS. To our knowledge, there are no published reports describing ISA in patients with IS. The purpose of this study was to describe ictal patterns of ISA in patients with IS and to correlate with clinical data. Methods: EEG recordings of all cases of IS in the past 10 years at the Alberta Children’s Hospital were reviewed. Inclusion criteria were a technically adequate video EEG recording that captured at least one spasm. For each patient, the first 10 confirmed spasms were examined. Spasms were evaluated for changes in ISA, which were either generalized, lateralized, or absent ISA (g-ISA, l-ISA, or n-ISA, respectively). Results were correlated with treatments, clinical course, and information pertinent to likely etiology of the IS. Results: A total of 77% of spasms were associated with ISA; 57% with g-ISA, 20% l-ISA, and 21% n-ISA. All patients with exclusively g-ISA showed at least a partial response to initial therapy, while this was the case in 66.7% of those with at least some l-ISA and 50% of those with exclusively n-ISA. Other seizure types occurred in 60% of patients with exclusively g-ISA versus 83% with some l-ISA and all patients with exclusively n-ISA. Conclusions: Ictal ISA was observed in the majority of IS. Trends were observed suggesting that the presence of exclusive g-ISA changes may be a positive prognostic factor in IS. Key Words: Infantile spasms, Infraslow oscillations, Epileptic encephalopathy, thalamus. (J Clin Neurophysiol 2014;31: 600–605)
I
nfantile spasms (IS) are a rare, but often devastating epileptic encephalopathy of infancy with an incidence of 1 to 2 per 10,000 live births (Brna et al., 2001; Sidenvall and Eeg-Olofsson, 1995). Infantile spasms usually present before the first year of life and, even in patients with focal brain abnormalities, usually involve bilateral flexion, extension, or mixed flexion–extension spasms. The EEG patterns in IS are well characterized with conventional frequency range (1–70 Hz) EEG. Eleven different ictal subtypes have been described, involving combinations of slow waves, sharp and slow waves, attenuation, fast activity, and rhythmic slow waves with generalized attenuation after the seizure being the most common
From the *Section of Neurology, Department of Paediatrics, Faculty of Medicine, University of Calgary, Calgary, Canada; †Department of Clinical Neurosciences, Faculty of Medicine, University of Calgary, Calgary, Canada; ‡Alberta Children’s Hospital Research Institute for Child and Maternal Health, Calgary, Canada; and §Department of Radiology, Faculty of Medicine, University of Calgary, Calgary, Canada. Address correspondence and reprint requests to Morris H. Scantlebury, MD, 278 HMRB, 3300 Hospital Drive, Calgary, AB T2N 4N1, Canada; e-mail: morris. [email protected]. Copyright Ó 2014 by the American Clinical Neurophysiology Society
ISSN: 0736-0258/14/3106-0600
600
(Kellaway et al., 1979). Hypsarrhythmia is the classic interictal EEG pattern, defined as a high amplitude multifocal spike wave discharges with a disorganized background. Although hypsarrhythmia sometimes improves with arousal or immediately after a cluster of spasms (Hrachovy et al., 1984), in many patients, it occupies most of the EEG recording. Some have proposed that hypsarrhythmia be considered a form of status epilepticus accounting in large part for the abrupt development regression seen in these patients. There is now emerging evidence that an extended evaluation of the EEG beyond the conventional frequencies between 1 and 70 Hz may have significantly impact the treatment of patients with IS. For example, frequencies above 70 Hz, known as high frequency oscillations, have been studied in IS and have shown potential utility in making treatment decisions including presurgical planning (Iwatani et al., 2012; Nariai et al., 2011). Although these studies support a role for the careful evaluation of high frequency oscillations in patients with IS, analysis of frequencies below the conventional frequency range may also be of benefit. Infraslow activity (ISA) EEG analysis has shown potential in differentiating absence and focal onset seizures (Rodin et al., 2008) as well as preepilepsy surgical planning (Modur et al., 2012). Despite these potential applications, there have been no published reports, to our knowledge, involving ISA analysis in early childhood epilepsy in general and IS specifically. This is an important gap in the literature because recent studies have shown that ISA has the potential to be useful in seizure characterization and localization and may eventually help in guiding epilepsy management by helping in the choice of medications and in the identification of epileptogenic cortical regions in preparation epilepsy surgeries (Constantino and Rodin, 2012; Modur et al., 2012; Rampp and Stefan, 2012; Rodin and Modur, 2008; Rodin et al., 2008, 2009). This study is a single institution, retrospective analysis of ISA during infantile spasms and correlates the findings with the patients’ clinical characteristics.
METHODS The research protocol was approved by the University of Calgary’s Conjoint Health Research Ethics Board. Two attending epileptologists (L.E.B. and M.H.S.) performed a search of the pediatric neurology clinic’s patient records and generated a list of all patients in the past 10 years with a diagnosis of infantile spasms. This list, with no additional clinical information, was provided to a pediatric neurology resident (K.A.M.) who reviewed the EEG recordings masked to the patient demographics. The results of the analysis were independently analyzed by L.E.B. and M.H.S. Only patients younger than 2 years of age at the time of the recording were included to minimize the confounding effects of brain maturation on ISA patterns. Minimum requirements for inclusion were a technically adequate video EEG recording that captured at least one infantile spasm, confirmed on the official EEG
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
report. Events suspicious for spasms were identified by technician or epileptologist notes on the recording. All EEG analysis was conducted using the XLTek NeuroWorks system (Natus Medical Inc, San Carlos, California). These were first examined using a bipolar montage and conventional frequency range (1–70 Hz). Events were included if there was clinical evidence of a spasm on the video recording, correlated with classic EEG findings. These events were considered confirmed spasms and analyzed for ISA changes. Once an event was confirmed as a spasm, ISA was analyzed with an ipsilateral ear referential montage using a band pass filter with a frequency range of 0.01 to 0.1 Hz (Rodin et al., 2008). Spasms were analyzed at a time base of 15 mm/s and sensitivity that varied from 10 to 50 mV/mm. For each patient, the first 10 confirmed spasms were analyzed. The recording was examined under these new settings for ISA 30 seconds before and after the spasm. Although the study was focused on ictal changes, this relatively wide time window was thought appropriate because waveforms in the infraslow range have quite long periods (10–100 seconds). Changes were initially recorded as present or absent. If present, the changes were classified as generalized (symmetric) or lateralized, with laterality defined as the amplitude in one hemisphere being more than 50% the contralateral side (Figs. 1 and 2). Patients were then grouped as those having: 1. Exclusively generalized or absent ISA (G-ISA), 2. At least one event with lateralized ISA (L-ISA), or 3. No ISA change with any spasms (N-ISA). We also recorded if the spasm occurred in the context of a cluster (cluster defined as 2 or more spasms occurring in series with an interval of .30 seconds between spasms). Once the EEG analysis was completed for all patients, pediatric neurology clinic notes and
Infraslow EEG in Infantile Spasms
brain imaging reports were reviewed. Imaging results, treatments, clinical course over the 2 years after spasm onset, and information pertinent to likely etiology of the infantile spasms were recorded. Additionally, a pediatric neuroradiologist (X-C.W.), masked to the EEG results, reviewed each patient’s brain MRI and report specifically on any thalamic pathology.
Statistical Analysis Chi-squared analysis was used to compare proportions between groups. Kruskal–Wallis one-way analysis of variance on ranks was used for multiple group comparisons where appropriate.
RESULTS Patient Characteristics The EEG recordings of 136 patients were reviewed, and 13 patients met criteria for inclusion. The reasons for exclusion were (1) spasms were not captured during the recoding and (2) video was not available to analyze the spasms captured for ISA. When known, the mean age of spasm onset was 7.0 6 4.3 (mean 6 SD) months. There were no differences in the age of onset of spasms between the groups (P ¼ 0.391; Kruskal–Wallis one way analysis of variance on ranks). IS was due to a symptomatic etiology in 12/13 (92.3%) patients. The etiologies were diverse and are listed in Table 1. Vigabatrin was used as an initial therapy for spasms in 11/13 patients (84.6%), whereas the remaining 2 patients received adrenocorticotropic hormone. The mean age at EEG for analysis was 8.8 6 5.3 months.
FIG. 1. Generalized ISA change. Patient 5: 8 months old with cortical dysplasia. EEG recording settings, 20 mV/mm and 10 mm/s. Arrow denotes clinical spasm. Copyright Ó 2014 by the American Clinical Neurophysiology Society
601
K. A. Myers et al.
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
FIG. 2. Lateralized ISA change. A, Individual spasm. Patient 9: 4 months old with hypoxic ischemic encephalopathy. EEG recording settings, 30 mV/mm and 10 mm/s. B, Cluster of spasms. Patient 3: 5 months old with mitochondrial disorder. EEG recording settings, 50 mV/mm and 5 mm/s. Arrows denote clinical spasms.
Spasms Characteristics There were a total of 101 spasms analyzed, of which 55/101 (54%) occurred in the context of a cluster. Of the total number of spasms, 78 (77%) were associated with ISA, and 58/101 (57%) of spasms were associated with generalized ISA (g-ISA), 20/101 602
(20%) with lateralized (l-ISA), and 23/101 (23%) with no ISA (n-ISA). After completion of the study, a less formal analysis was done to investigate whether lateralization in the infraslow range correlated with similar asymmetry in the conventional frequency Copyright Ó 2014 by the American Clinical Neurophysiology Society
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
TABLE 1.
Clinical Characteristics and EEG Results Etiology
Age of Onset
Age at EEG (months)
Number of Spasms Analyzed
MCD Meningitis Mitochondrial IVH MCD Genetic/IVH Meningitis/IVH TS HIE Genetic/Metabolic MCD MCD Genetic/Metabolic
14 5 3 Unclear 6 12 9 12 2.5 Unclear 5 2 5
14 5 5 22 8 12 9 12 4 3 5 10 5
9 1 8 7 3 3 10 10 10 10 10 10 10
Patient 1 2 3 4 5 6 7 8 9 10 11 12 13
Patient 1 2 3 4 5 6 7 8 9 10 11 12 13
Infraslow EEG in Infantile Spasms
Spasms With l-ISA (%) 2 0 8 4 0 0 4 0 1 1 0 0 0
(22) (0) (100) (57) (0) (0) (40) (0) (10) (10) (0) (0) (0)
Spasms With n-ISA (%) 0 0 0 1 2 0 0 0 0 0 0 10 10
(0) (0) (0) (14) (67) (0) (0) (0) (0) (0) (0) (100) (100)
Overall ISA Pattern L-ISA G-ISA L-ISA L-ISA G-ISA G-ISA L-ISA G-ISA L-ISA L-ISA G-ISA N-ISA N-ISA
Total Number Spasms ISA (%) 9 1 8 6 1 3 10 10 10 10 10 0 0
Response to Initial Treatment (Vigabatrin/ACTH) Partial (Vigabatrin) Complete (Vigabatrin) Complete (Vigabatrin) No Response (Vigabatrin) Partial (Vigabatrin) Partial (ACTH) Partial (ACTH) Complete (Vigabatrin) Complete (Vigabatrin) No Response (Vigabatrin) Partial (Vigabatrin) No Response (Vigabatrin) Partial (Vigabatrin)
(100) (100) (100) (86) (33) (100) (100) (100) (100) (100) (100) (0) (0)
Spasms With g-ISA (%) 7 1 0 2 1 3 6 10 9 9 10 0 0
(78) (100) (0) (29) (100) (100) (60) (100) (90) (90) (100) (0) (0)
Development of other Seizure Types (1/2) 1 (Tonic, Absence) 1 (Myoclonic, Focal dyscognitive) 1 (GTC, Myoclonic) 1 (Focal clonic, Focal dyscognitive) 2 1 (Myoclonic, Focal clonic) 1 (Focal dyscognitive) 1 (Focal dyscognitive) 2 1 (Myoclonic, Tonic, Focal clonic) 2 1 (Focal) 1 (Myoclonic)
ACTH, adrenocorticotropic hormone; G-ISA, patients with exclusively g-ISA 6 n-ISA with spasms; HIE, hypoxic ischemic encephalopathy; IVH, intraventricular hemorrhage; L-ISA, patients with at least one spasm with l-ISA; MCD, malformation of cortical development; N-ISA, patients with exclusively n-ISA with spasms; TS, tuberous sclerosis.
range. Although asymmetry in the conventional range was observed in some cases, there was no clear relation to the patterns seen with ISA.
Patient-Specific Results
A mean of 7.8 6 3.3 spasms were analyzed per patient with no significant differences in the number of spasms analyzed between the groups. Only 2/13 patients (15%) did not show clear ictal ISA with any of their captured spasms (N-ISA). Of the remaining patients with clear ictal ISA, 6/13 (46%) had evidence of lateralization with at least one of their spasms (L-ISA). The remaining 5 patients (38%) were classified as having exclusively g-ISA 6 n-ISA with their spasms (G-ISA). All of the G-ISA patients showed at least a partial response to initial therapy, while this was the case in 4/6 L-ISA patients (66.7%) and 1/2 N-ISA patients (50%). Interestingly, two symptomatic patients received adrenocorticotropic hormone as an initial therapy, one with G-ISA and the other with L-ISA, with both showing a partial response to treatment. There were no significant differences between the groups regarding response to treatment. Regarding long-term outcome, 3/5 (60%) patients with G-ISA developed other seizure types compared with 5/6 (83%) with L-ISA and 2/2 (100%) with N-ISA. These differences were not statistically significant. Copyright Ó 2014 by the American Clinical Neurophysiology Society
MRI Analysis of Thalamic Pathology
The thalamic MRI findings are summarized in Table 2. In summary, thalamic damage was seen in 8/13 (62%) patients and was clearly asymmetric in only 2 patients. Both (100%) of these patients showed L-ISA while this lateralized pattern was seen in 2/6 (33%) patients with symmetric thalamic pathology and 2/5 (40%) patients with no apparent thalamic abnormality.
DISCUSSION
To our knowledge, this is the first study describing the ictal ISA patterns in patients with IS. The major finding is that ictal ISA occurs in the majority of patients with IS. ISA may be generalized or lateralized, with many patients showing a mix of both patterns and some having no apparent change whatsoever. Ictal ISA was recorded in more than 80% of spasms in this study. Taking into consideration the ictal baseline shifts seen before the onset of focal onset seizures (Rodin et al., 2008), we propose that ISA recorded at the time of an infantile spasm is pathologic and related to thalamic dysfunction. There are several lines of evidence supporting a role for thalamic dysfunction in the generation of ictal ISA in patients with IS. First, the thalamus is thought to play a critical role in the generation of ISA (Hughes et al., 2011). One recent study showed that damage to the thalami and basal ganglia in patients with 603
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
K. A. Myers et al.
TABLE 2. Patient
MRI Analysis of Thalami Pattern of ISA Expression
1
L-ISA
2
G-ISA
3
L-ISA
4
L-ISA
5
G-ISA
6
G-ISA
7
L-ISA
8
G-ISA
9
L-ISA
10
L-ISA
11
G-ISA
12
N-ISA
13
N-ISA
General MRI Findings Bilateral ventriculomegaly and colpocephaly; agenesis of corpus callosum Diffuse cerebral atrophy; delayed myelination Diffuse cerebral atrophy Extensive white matter abnormalities, ventriculomegaly and multiple areas of porencephaly Restricted diffusion, mainly in midbrain and globi pallidi Exvacuo hydrocephalus and right frontoparietal porencephalic cyst; enlarged posterior fossa Exvacuo hydrocephalus; multifocal encephalomalacia Bilateral cortical and subcortical tubers; subependymal nodules Bilateral areas of restricted diffusion; left occipitotemporal infarct Normal at 11 weeks; later, generalized cerebral atrophy and T2 hyperintensity in globi pallidi, diencephalon, and dentate nucleus Ventriculomegaly; corpus callosum dysgenesis; pachygyria; extensive white matter abnormalities; multicystic encephalomalacia; hypoplastic cerebellar vermis Aqueductal stenosis with hydrocephalus; thin corpus callosum; absent septum pellucidum; multiple subependymal heterotopias Bilateral cytotoxic edema (developed after vigabatrin)
Thalamic Abnormality (1/2)
Thalamic Abnormality (Symmetric or Asymmetric)
Type of Thalamic Abnormality
2
2
2
1
Symmetric
1
Symmetric
1
Asymmetric (R . L)
Initially restricted diffusion; later, atrophy and gliosis Restricted diffusion (resolved on subsequent study) Gliosis and volume loss
1
Symmetric
Restricted diffusion
1
Symmetric
Mild volume loss
1
Asymmetric (L . R)
2
2
1
Symmetric
2
2
Restricted diffusion consistent with infarct/ischemia; subtle hemorrhage 2
2
2
2
2
2
2
1
Symmetric
Subtle restricted diffusion; later, gliosis and volume loss 2
Diffuse cytotoxic edema (developed after vigabatrin)
G-ISA, patient had exclusively generalized or absent infraslow activity with spasms; L-ISA, patient had signs of lateralization of infraslow activity with at least some spasms; N-ISA, patient did not change in infraslow activity with any spasms.
hypoxic ischemic encephalopathy was predictive of progression to IS (Gano et al., 2013). The tendency for spasms to occur around the sleep–wake transition, and for hypsarrhythmia to become more prominent during sleep, are both suggestive of thalamocortical dysfunction. In patients with Lennox–Gastaut syndrome, a highly related catastrophic epilepsy, there is a higher recorded frequency of cyclical alternating pattern of non-REM sleep, and both the generalized spike and wave pattern and seizures occur more frequently during the active phase of the cyclical alternating pattern (Eisensehr et al., 2001). This is an important observation as cyclical alternating 604
pattern is characterized by slow oscillations of ,0.1 Hz, occurring with a periodicity of 20 to 35 seconds in stages 2 and 3 of natural non-REM sleep (Terzano and Parrino, 2000) and is thought to be linked to thalamocortical volleys given its widespread EEG expression and due to its ability to modulate diffuse ascending and descending pathways (Eisensehr et al., 2001; Terzano et al., 1991). Given the probable relationship between thalamic dysfunction and ISA expression, asymmetric damage to the thalami of patients with IS is a possible explanation for the l-ISA observed in some patients in this study. Patients with exclusively g-ISA can be Copyright Ó 2014 by the American Clinical Neurophysiology Society
Journal of Clinical Neurophysiology Volume 31, Number 6, December 2014
speculated to have more symmetric thalamic involvement. Of note, two patients had n-ISA correlating with spasms, which may indicate relative preservation of “normal” thalamic function. Based on this theory, the degree and symmetry of thalamic involvement could be inferred by an analysis of the ISA activity, which could have both prognostic and therapeutic significance. Our MRI analysis of thalamic abnormalities was somewhat supportive of this theory; however, autopsy analysis or analysis of other surrogate electrophysiological markers for thalamic function, such as sleep spindles (Ho et al., 2014), would likely be the best approach to assess thalamic pathology and any associated asymmetry thereof. Notably, a recent article reported that symmetrical thalamic hyperperfusion on ictal SPECT (single-photon emission computerized tomography) during infantile spasms was a positive prognostic factor, when compared with asymmetric hyperperfusion (Haginoya et al., 2013). The functional mechanisms underlying ISA are currently poorly understood. However, electrophysiological recordings conducted in in vitro thalamic preparations suggest that ISA is likely generated through the activation of G-coupled inward rectifying potassium channels (Hughes et al., 2011), which are the downstream effectors of G-coupled receptors, including the GABAB receptor (Luscher et al., 1997). Notably, diffuse overexpression of G-coupled inward rectifying potassium 2 channels has been described in the brains of the Ts65Dn mouse model of Down syndrome, a model that forms the basis of an important animal model of IS induced by exposing the Ts65Dn mice to the GABAB agonist gammabutyrolactone (Cortez et al., 2009). Together, these results suggest that ictal ISA recorded in association with the spasms may in part be due to abnormal G-coupled inward rectifying potassium channel expression in the thalamus among other regions of the brain. Although, not statistically significant, the results of this study suggest that the presence of exclusively g-ISA may have a positive prognostic factor in IS. However, significant caveats are the small sample size, which included a high number of individuals with symptomatic IS. This latter factor precluded an analysis of the patterns observed in cryptogenic versus symptomatic cases. A second factor was the variability in treatment of IS, and the reliance on clinical notes to determine the extent of response to initial treatment. Furthermore, our decision to analyze only the first 10 spasms was somewhat arbitrary, and the possibility exists that substantially more patients might show signs of lateralized ISA if more spasms were analyzed. An optimal study would be conducted prospectively, including a standard minimum number of events, allowing for a consistent sampling of ISA patterns for each patient. In conclusion, this study demonstrates the exciting potential and need for further study, of ISA in childhood epilepsy in general and IS in particular. Overall, the results of our study demonstrate that
Copyright Ó 2014 by the American Clinical Neurophysiology Society
Infraslow EEG in Infantile Spasms
ISA changes frequently occur concurrently with IS and may provide prognostic information for these individuals. Much more work is required to better clarify the potential utility of this tool in the pediatric population, which will include defining normal patterns at different ages and clarification of the underlying physiology. REFERENCES Brna PM, Gordon KE, Dooley JM, Wood EP. The epidemiology of infantile spasms. Can J Neurol Sci 2001;28:309–312. Constantino T, Rodin E. Peri-ictal and interictal, intracranial infraslow activity. J Clin Neurophysiol 2012;29:298–308. Cortez MA, Shen L, Wu Y, et al. Infantile spasms and Down syndrome: a new animal model. Pediatr Res 2009;65:499–503. Eisensehr I, Parrino L, Noachtar S, et al. Sleep in Lennox-Gastaut syndrome: the role of the cyclic alternating pattern (CAP) in the gate control of clinical seizures and generalized polyspikes. Epilepsy Res 2001;46:241–250. Gano D, Sargent MA, Miller SP, et al. MRI findings in infants with infantile spasms after neonatal hypoxic-ischemic encephalopathy. Pediatr Neurol 2013;49:401–405. Haginoya K, Uematsu M, Munakata M, et al. The usefulness of subtraction ictal SPECT and ictal near-infrared spectroscopic topography in patients with West syndrome. Brain Dev 2013;35:887–893. Ho AW, Myers KA, Bello-Espinosa LE, Scantlebury MH. Infra-slow EEG activity and sleep spindle expression: Potential window into thalamic function in infantile spasms (Conference Abstract). Paper presented at: 13th International Child Neurology Congress; May 4–9, 2014; Iguazu Falls, Brazil. Hrachovy RA, Frost JD Jr, Kellaway P. Hypsarrhythmia: variations on the theme. Epilepsia 1984;25:317–325. Hughes SW, Lorincz ML, Parri HR, Crunelli V. Infraslow (,0.1 Hz) oscillations in thalamic relay nuclei basic mechanisms and significance to health and disease states. Prog Brain Res 2011;193:145–162. Iwatani Y, Kagitani-Shimono K, Tominaga K, et al. Ictal high-frequency oscillations on scalp EEG recordings in symptomatic West syndrome. Epilepsy Res 2012;102:60–70. Kellaway P, Hrachovy RA, Frost JD Jr, Zion T. Precise characterization and quantification of infantile spasms. Ann Neurol 1979;6:214–218. Luscher C, Jan LY, Stoffel M, et al. G protein-coupled inwardly rectifying K1 channels (GIRKs) mediate postsynaptic but not presynaptic transmitter actions in hippocampal neurons. Neuron 1997;19:687–695. Modur PN, Vitaz TW, Zhang S. Seizure localization using broadband EEG: comparison of conventional frequency activity, high-frequency oscillations, and infraslow activity. J Clin Neurophysiol 2012;29:309–319. Nariai H, Nagasawa T, Juhasz C, et al. Statistical mapping of ictal high-frequency oscillations in epileptic spasms. Epilepsia 2011;52:63–74. Rampp S, Stefan H. Ictal onset baseline shifts and infraslow activity. J Clin Neurophysiol 2012;29:291–297. Rodin E, Modur P. Ictal intracranial infraslow EEG activity. Clin Neurophysiol 2008;119:2188–2200. Rodin E, Constantino T, van Orman C, House P. EEG infraslow activity in absence and partial seizures. Clin EEG Neurosci 2008;39:12–19. Rodin E, Constantino T, Rampp S, Modur P. Seizure onset determination. J Clin Neurophysiol 2009;26:1–12. Sidenvall R, Eeg-Olofsson O. Epidemiology of infantile spasms in Sweden. Epilepsia 1995;36:572–574. Terzano MG, Parrino L. Origin and significance of the cyclic alternating pattern (CAP): review article. Sleep Med Rev 2000;4:101–123. Terzano MG, Parrino L, Spaggiari MC, et al. Discriminatory effect of cyclic alternating pattern in focal lesional and benign rolandic interictal spikes during sleep. Epilepsia 1991;32:616–628.
605
