Clinical Neurophysiology
Volume 120, Issue 1 , Pages 3-10, January 2009

Characteristics of generalised epileptiform activity

  • H. Aurlien

      Affiliations

    • Section of Clinical Neurophysiology, Department of Neurology, Haukeland University Hospital, Jonas Liesvei 65, N-5021 Bergen, Norway
    • Corresponding Author InformationCorresponding author. Tel.: +47 55975000; fax: +47 55975164.
    • ,
  • I.O. Gjerde

      Affiliations

    • Department of Neurology, Haukeland University Hospital, Bergen, Norway
    • ,
  • G.E. Eide

      Affiliations

    • Centre for Clinical Research, Haukeland University Hospital, Bergen, Norway
    • Department of Public Health and Primary Health Care, University of Bergen, Bergen, Norway
    • ,
  • J.C. Brøgger

      Affiliations

    • Section of Clinical Neurophysiology, Department of Neurology, Haukeland University Hospital, Jonas Liesvei 65, N-5021 Bergen, Norway
    • Department of Clinical Medicine, University of Bergen, Bergen, Norway
    • ,
  • N.E. Gilhus

      Affiliations

    • Department of Neurology, Haukeland University Hospital, Bergen, Norway
    • Department of Clinical Medicine, University of Bergen, Bergen, Norway

Accepted 15 October 2008.

Article Outline

Abstract 

Objective

To study the age-related occurrence of specific features of generalised epileptiform activity (GEA), their correlation with EEG background activity (BA), and their internal correlation.

Methods

17,723 consecutive routine EEGs from 12,511 patients were annotated and categorised into a database. The first EEG containing GEA from all 325 patients with such activity were selected and categorised for GEA features. The BA was studied in multivariable fractional polynomial regression models including intervening variables. The GEA features were studied in similar models for age-dependency and internal correlation.

Results

High GEA-amplitude and low GEA-frequency correlated with BA slowing. The occurrence of ‘irregular spike/sharp slow wave’ pattern increased with age (p=0.003). Hyperventilation sensitivity was not age-related. There was no correlation between hyperventilation sensitivity and photoparoxysmal response. The age-related probability for specific GEA-types was established.

Conclusions

High GEA-amplitude and low GEA-frequency correlate with BA slowing, indicating cerebral cortical dysfunction. Hyperventilation sensitivity and photoparoxysmal response independently increase the yield of EEG. There is no age-dependency for hyperventilation sensitivity showing that an upper age threshold for hyperventilation provocation is inappropriate.

Significance

The results extend our understanding of GEA and help the electroencephalographer in weighing the various GEA components.

Keywords: Electroencephalography, Generalised epileptiform activity, Alpha rhythm, Background activity, Databases

 

Back to Article Outline

1. Introduction 

The diagnosis of epilepsy is clinical, and good clinical information is essential for the assessment of the significance of epileptiform activity in EEG (Binnie and Stefan, 1999, Sam and So, 2001). The morphology of the epileptiform activity and the response to photic stimulation and activation procedures can, however, aid in determining the type of epilepsy and specific epilepsy syndromes, and thereby also be instrumental in the choice of antiepileptic medication and in evaluating the prognosis.

Generalised epileptiform activity (GEA) can be seen in the EEG in idiopathic generalised epilepsies (IGEs), symptomatic/cryptogenic generalised epilepsies, and in localisation oriented epilepsies with fast generalisation of the epileptiform activity (Koutroumanidis and Smith, 2005). There are no infallible guidelines for EEG identification of these epilepsy types. Nevertheless some rules of thumb are widely accepted: polyspike wave discharges are typically found in epilepsies associated with myoclonus, and photoparoxysmal response (PPR) is most of all associated with juvenile myoclonic epilepsy (JME) (Markand, 2003, Smith, 2005). Generalised spike slow wave or polyspike slow wave discharges with abrupt bilateral synchronous onset and with a frequency of 3Hz or faster is a hallmark of IGE, while similar bursts, but with a frequency lower than 2.5Hz, are indicative for symptomatic generalised epilepsies. Focal or generalised slowing indicating cerebral non-epileptiform pathology is indicative of symptomatic epilepsy (Pillai and Sperling, 2006, Smit et al., 2006, Guerrini, 2006), although quantitative EEG studies have demonstrated some slowing of the background activity also in IGE (Pfurtscheller and da Silva, 1999, Clemens et al., 2000, Clemens, 2004).

EEG-based studies of epileptiform activity are usually limited to clinically well defined epilepsy entities or syndromes. This study, in contrast, focuses on epileptiform activity itself and features linked to this activity from an electroencephalographer’s viewpoint. Including all EEGs with GEA implies a heterogeneous clinical group. Including numerous GEA features available from routine visual EEG examinations makes it possible to establish which of these features correlate with generalised brain pathology, operationally defined as slowing of the background activity. Such knowledge will contribute to a better understanding of the influence of such activity on general brain activity and also help the electroencephalographer to interpret the various components of the epileptiform activity. Examining GEA in a large and unselected population provides information that is inaccessible when studying selected clinical entities only.

The aims of this study were:

To define the GEA features that correlate with pathological background activity.

To study the age-related occurrence of GEA features and GEA-related variables.

To study the correlation between the various GEA features.

Back to Article Outline

2. Methods 

2.1. EEG recordings 

All routine EEGs recorded at Haukeland University Hospital from March 1st 2000 to December 31st 2005 were visually evaluated and described. This included 17,723 EEGs from 12,511 patients. Long-term registrations, and EEGs during general anesthesia, Wada tests, and tilt tests were not included in this study (Wada and Rasmussen, 1960, Low, 1983). Due to the lack of consensus regarding the assessment of rhythmic and periodic EEG patterns encountered in critically ill patients, 40 EEGs from such patients were excluded (Hirsch et al., 2005). Critically ill patients were defined as patients with an EEG referral, hospital admittance, or hospital discharge diagnosis for this stay including coma or acute anoxic brain damage. Suppression-burst complexes, triphasic waves and paroxysmal flattening can appear as epileptiform phenomena, but all 13 EEGs with such activity were excluded due to the controversy about the etiology of this activity (Raegrant et al., 1991, Husain et al., 1999). The first EEG containing GEA from each patient was selected for this study. A total of 366 EEGs marked in routine coding as having GEA were re-evaluated, and 325 EEGs, 181 from females and 144 from males, fulfilled all inclusion criteria (Table 1). These 325 EEGs were scored for the following GEA features: Waveform, bilateral synchronicity, regularity, frequency, amplitude of the sharp component, photoparoxysmal response (PPR), and hyperventilation sensitivity (HVS).

Table 1. GEA-types, GEA features, and GEA-related variables in 325 consecutive GEA patients.
N
Spike/sharp slow wave207
Polyspike slow wave40
Polyspike20
Irregular spike/sharp wave43
Hypsarrhythmia15
Variable valuesN
GenderFemale; male181; 144
Contact-typeInpatient; outpatient86; 239
RBS GEANo; yes268; 57
IctalNo; yes314; 11
FEANo; yes177; 148
CBRDEEaNo; yes233; 92
MedicatedUnknown; no; yes11; 117; 197
Post-ictalUnknown; no; yes19; 281; 25
HVSNot tested; no; yes126; 154; 45
PPRNot tested; no; yes66; 194; 65
MinMaxMeanSDbN
Age (years)0.0289.618.215.1325
GBA-amplitude (μV)540034.433.4321
GBA-frequency (Hz)0.5104.41.6321
AR-amplitude (μV)7.512049.119.8236
AR-frequency (Hz)6149.41.2236
GEA-amplitude (μV)301030175135325
GEA-frequency (Hz)253.20.6063

aCBRDEE, current brain-related disease except epilepsy.

bSD, standard deviation.

The first EEG from all drug-free outpatients with no EEG pathology from the study period were included in the control group (N=3268).

Haukeland University Hospital recruits patients from a population of about 500,000, and is the only provider of EEG within this area. The EEGs were described by one of 6 electroencephalographers. Three of these were certified, experienced electroencephalographers, and three were trainees under supervision.

The age distribution at test time for the patients with GEA is shown in Fig. 1. The mean age was 18.2 years, median 14.8 years. The mean age for the control group was 24.2 years, median 19.6 years. Epilepsy was the referral diagnosis in 318 of the GEA patients (98%). For 300 patients (92%), a clinical diagnosis of epilepsy was made. One hundred and sixty-eight (56%) had an ICD10 diagnosis of generalised epilepsy, 105 (35%) had both diagnoses of generalised and localisation oriented epilepsy, 71 (24%) had only localisation oriented epilepsy diagnosis, whereas 61 (20%) had an unspecified epilepsy diagnosis. For 2 patients (1%), no clinical diagnosis was available. The referral diagnosis was epilepsy in 1885 controls (57%). Totally, 197 GEA patients (61%) received one or more drugs (Table 2), 117 (36%) were drug-free, and this information was missing in 11 (3%). Fifty EEGs with GEA (15%) were recorded after sleep-deprivation with the patient both awake and asleep. No patients received drugs to induce sedation. Standard duration for the EEG registration was 20, and 60min after sleep-deprivation. Hyperventilation provocation was performed in 199 (61%), and photic stimulation in 259 (80%) of the GEA patients. All EEGs were recorded on the digital EEG system Nervus®. The electrodes were placed according to the international 10–20 system (Fp1, Fp2, Fpz, F3, F4, F7, F8, Fz, A1, A2, T3, T4, T5, T6, C1, C2, Cz, P3, P4, Pz, O1, and O2).

Table 2. Drug use in 325 consecutive GEA patients, classified according to the Anatomic Therapeutic Chemical Classification System (WHO Collaborating Centre for Drug Statistics Methodology, 2008).
Drug typeN%
Antiepileptics13842
Respiratory system206
Psycholeptics175
Psychoanaleptics155
Cardiovascular system72
Analgesics62
Antidiabetics41
Antithrombotics41
Antiinflammatory31
Sex hormones or modulators of the genital system31
Systemic hormonal preparations, excl. sex hormones and insulin31
Antiinfectives21
Other258

2.2. EEG database 

The EEG interpretations were structured and categorised in a database using software developed for the purpose (Aurlien et al., 1999). This software automatically collected patient demographic data and administrative test parameters such as name, unique patient identification, date of birth, patient address, referral doctor, test notes, patient notes, and medication from the hospital patient administrative system and the EEG administrative database. For each EEG, the electroencephalographer set one or more ICD10 diagnoses based on the doctor’s EEG referral (Table 3). The system automatically added all ICD10 diagnoses set by clinicians at previous hospital contacts.

Table 3. Referral diagnoses for 325 consecutive GEA patients. Some patients had more than one diagnosis.
DiagnosesN%
Epilepsy31898
Mental retardation41
Encephalitis21
Cerebral palsy21
Myoclonus21
Specific developmental disorders21
Stroke21

2.3. EEG interpretation 

The interpretation was divided into 3 main sections: Alpha rhythm (AR), general background activity (GBA), and EEG events. AR was defined as the dominant posterior rhythm with frequency 8–13Hz that was blocked or attenuated by eye opening (Chatrian et al., 1983). Alpha variant rhythm was defined as with AR characteristics but with frequency outside the 8–13Hz alpha band (Chatrian et al., 1983). AR and alpha variant rhythm were analysed together and is referred to as AR. AR was evaluated for frequency, amplitude, reactivity, and asymmetry. Frequency and amplitude were measured for the occipital leads in a monopolar average montage in an EEG segment where AR was distinctly appearing. They were evaluated as a range with a minimum and maximum value. The maximum value was used for AR-frequency. AR-amplitude was defined as the mean of the upper and lower value.

Background activity was defined as any EEG activity representing the setting in which a given normal or abnormal pattern appeared and from which such a pattern was distinguished (Chatrian et al., 1983). AR was described separately. The remaining continuous general activities were defined as the GBA. Focal, asymmetric, and intermittent activities were not defined as part of the background activity, but described as EEG events. The GBA was evaluated in a monopolar average montage without signs of drowsiness. To assure alertness the patients were instructed to open their eyes during parts of the test. If drowsy, the patients were alerted with an audible signal (knocking), or asked to perform calculations and repeat series of digits. GBA was described as one or more ranges of frequencies and amplitudes. The minimum value was used for GBA-frequency as this value was assumed to be most relevant for EEG pathology (Gloor et al., 1977). GBA-amplitude was defined as the mean of the upper and lower value.

EEG events were defined as any part of the EEG that the electroencephalographer described separately from the frequency and amplitude measurement of the AR and GBA. An example of each type of EEG event was manually marked in the EEG editor with start and stop time. These events were automatically picked up by the description module, and the electroencephalographer then classified them according to the American Standards for Testing and Materials (ASTM, 1994). The ASTM categories were divided hierarchically into 4 main branches: ‘epileptiform pathology’, ‘non-epileptiform pathology’, ‘normal findings/variants’, and ‘extra-cerebral activity’ (Westmoreland and Klass, 1990). Pathological EEGs were defined as having at least one EEG event from one of the first two branches.

2.4. GEA 

The localisation of the EEG events was given by clicking the traces where the activity occurred. The electrodes were divided into 5 regions; frontal, temporal, central, parietal, and occipital. Each event was automatically classified as general or focal. If three or more regions were affected at both sides and with no, mild or moderate asymmetry, the event was classified as general. All other events were classified as focal. All EEGs with one or more general events categorised as ‘epileptiform pathology’ were classified as having GEA, also if there were additional focal epileptiform activity (FEA) in the same EEG.

2.4.1. GEA-types, GEA features, and GEA-related variables 

GEA was categorised into 5 types; ‘Spike/sharp slow wave’, ‘polyspike slow wave’, ‘polyspike’, ‘irregular spike/sharp wave’, and ‘hypsarrhythmia’ (Table 1). ‘Spike/sharp slow wave’ was defined as a pattern of a spike followed by a slow wave or a sharp wave followed by a slow wave. ‘Polyspike slow wave’ was defined as a sequence of multiple spikes and a slow wave. ‘Polyspike’ was defined as a multiple spike complex; several spikes occurring in a sequence without slow waves. ‘Hypsarrhythmia’ was defined as a pattern consisting of high voltage arrhythmic slow waves interspersed with spikes, without consistent synchrony between the two sides of the head or different areas on the same side (Noachtar et al., 1999). Irregular spike/sharp wave’ was defined as paroxysms of spike or sharp waves with or without slow waves and with an irregular frequency of about 3–5Hz (Markand, 2003). If the GEA changed from one category to another during a paroxysm, the GEA-type at the start of the paroxysm was chosen for this study. The GEA and patient were further described by 11 features (Table 1): age (years), gender (female/male), contact-type (outpatient/inpatient), medicated (no/yes), FEA (no/yes), post-ictal (no/yes), ictal (no/yes), current brain-related disease (no/yes), HVS (no/yes), PPR (no/yes), regular bilateral synchronous (RBS) GEA (no/yes), GEA-frequency (Hz), and GEA-amplitude (μV). GEA-frequency was measured only if the GEA appeared rhythmically. The GEA-amplitude was defined as the amplitude of the sharp part of the paroxysm measured in the monopolar montage. GEA-amplitude and frequency were measured during the first part of a paroxysm. The EEG was by us assessed as HVS and PPR if the occurrence of GEA was more than doubled during hyperventilation and flicker provocation, respectively (Waltz et al., 1992). The EEGs were assessed as ‘post-ictal’ if a generalised convulsive seizure had occurred less than 24h before the recording and as ‘ictal’ if a generalised convulsive seizure was observed during the EEG registration. Current brain-related disease was defined as a chronic brain-related ICD10-diagnose set within one month after the EEG or a non-chronic ICD10-diagnose set between one month before and one month after the EEG.

2.4.2. Statistics 

Continuous and binomial dependent variables were analysed using multiple linear and logistic fractional polynomial regression models, respectively (Royston and Sauerbrei, 2005). Age-related changes in the occurrence of GEA-types were analysed using multinomial logistic regression analysis. Pairwise correlations between AR and GBA frequencies and amplitudes, as well as between HVS and PPR, were quantified using Spearman’s correlation coefficient.

The BA variables (AR-amplitude, AR-frequency, GBA-amplitude, and GBA-frequency) were subsequently included as dependent variables and analysed for three groups: All EEGs, EEGs with GEA, and EEGs with RBS GEA. By including the GEA features as independent variables, we tested their individual correlation with the BA variables. GEA features were subsequently included as dependent variables and analysed for two groups: EEGs with GEA and EEGs with RBS GEA. These tests were first performed with only age as independent variable. Then the other GEA features were included as independent variables. As HVS and PPR were tested in only 59% and 77% of GEA patients they were not included as independent variables. GEA-frequency was included as an independent variable only for the RBS GEA group. ‘Medicated’ and ‘post-ictal’ were always first included as independent variables. However, these variables were omitted if not significant to avoid excluding EEGs with missing ‘medicated’ and ‘post-ictal’ values. Index variables were included to test for differences between GEA subtypes. All GEA subtypes were compared to the most common subtype, ‘spike/sharp slow wave’. In some situations ‘irregular sharp waves’, ‘polyspikes’, ‘hypsarrhythmia’, ‘ictal’ and ‘contact-type’ were excluded as covariates due to collinearity. Only significant covariates (p<0.05) were included.

The square root transformation of the AR-amplitude and the natural log (ln) transformation of the GBA-amplitude and GBA-frequency were used to achieve near normality for the error terms. Data were back transformed for graphical display. No transformation was necessary for AR-frequency.

Results were considered significant with p-values <0.05. Stata 9.2 was used for the analyses.

Back to Article Outline

3. Results 

3.1. GEA and background activity 

All EEG background variables were significantly affected in EEGs with GEA compared to control EEGs, as illustrated in Fig. 2 and Table 4 (see also Online Appendix A). High GEA-amplitude and low GEA-frequency correlated with BA slowing (Table 4 and Online Appendix A).

  • View full-size image.
  • Fig. 2. 

    Fitted values for AR-amplitude (a), AR-frequency (b), GBA-amplitude (c), and GBA-frequency (d) by age in 325 GEA patients and in 3268 drug-free outpatient controls. Shaded areas are fitted values ±1.96 SE.

Table 4. Association between GEA-related variables and EEG BA, measured as AR- and GBA-amplitudes and frequencies.
Study groupsIndependent variableEffect variable for EEG background activitySignificance probability (p-value)
GEA patients (N=325) and controls (N=2368)GEA present↑ AR amp< 0.001
↓ AR freq< 0.001
↑ GBA amp< 0.001
↓ GBA freq0.038
Male gender↓ AR freq0.002
↓ GBA freq0.002

GEA patients (N=325)‘Polyspike’↑ AR amp0.037
‘Hypsarrhythmia’↑ GBA amp0.001
↓ GBA freq< 0.001
‘Ictal’↓ AR amp0.040
↑ AR freq0.042
↓ GBA freq0.004
‘Post-ictal’↓ GBA freq0.004
↑ GEA amp↑ AR amp0.003
↑ GBA amp0.006
Male gender↓ GBA freq0.009
CBRDEE↑ GBA amp0.018
↓ GBA freq0.004

Patients with RBS GEA (N=57)↓ GEA freq↓ GBA freq0.005
‘Post-ictal’↓ AR freq0.004
Male gender↓ AR freq0.033

↑, higher; ↓, lower; Amp, amplitude; Freq, frequency; CBRDEE, current brain-related disease except epilepsy.

AR-amplitude, AR-frequency, GBA-amplitude, and GBA-frequency were all pairwise correlated (p<0.001) (Fig. 3). Lower AR-frequency correlated with higher AR-amplitude, higher GBA-amplitude, and lower GBA-frequency.

  • View full-size image.
  • Fig. 3. 

    Pairwise correlation (Spearman’s) between AR-amplitude, AR frequency, GBA-amplitude, and GBA-frequency in 325 consecutive GEA patients (red) and 3268 controls (green). Each background activity variable is plotted against the other ones. Histograms show the number of EEGs with identified AR/GBA-amplitude/frequency in EEGs with GEA and in controls. The black lines in the histograms indicate the normal distribution.

AR could be identified in 236 of the 325 patients with GEA (73%), in 47 of the 57 patients with RBS GEA (82%), and in 2726 of the 3268 controls (83%). GBA could be identified in 321 of the patients with GEA (99%), in all patients with RBS GEA, and in all controls. The patients in which GBA could not be identified had continuous GEA.

3.2. GEA- and GEA-related features 

GEA-amplitude and GEA-frequency were not correlated (p=0.35), nor was the probability of HVS and PPR (p=0.54).

3.2.1. GEA-amplitude 

GEA-amplitude changed with age (p<0.001) (Fig. 4a and Online Appendix A), also if adjusted for other significant covariates (p<0.001). ‘Ictal’ EEG and EEG with RBS GEA both correlated with higher GEA-amplitude (p<0.001). GEA-amplitude changed with age also in the RBS GEA subgroup (p<0.001). In this subgroup EEGs with ‘polyspike slow wave’ correlated with higher GEA-amplitude (p=0.029).

3.2.2. GEA-frequency 

GEA-frequency did not change with age (Fig. 4b) (Online Appendix A). This was true also when adjusted for significant covariates, and also in the RBS GEA subgroup. ‘Polyspike’ correlated with higher GEA-frequency (p<0.001), while ‘ictal’ correlated with lower GEA-frequency (p=0.002).

3.2.3. HVS 

Forty-five of the 199 GEA patients tested had HVS (23%). This probability did not change with age (Online Appendix A), even if adjusted for other GEA features. RBS GEA increased the probability for HVS (odds ratio (OR)=12.7, p<0.001).

Twenty-two out of 37 tested in the RBS GEA subgroup had HVS (59%). This probability was age-independent, also if adjusted for other significant GEA features. HVS probability was higher with the GEA-type ‘spike/sharp slow wave’ than with ‘polyspike slow wave’ (OR=9.0, p=0.005).

Only 23 of 162 tested patients with GEA and without RBS GEA had HVS (14%).

3.2.4. PPR 

Sixty-five of the 259 GEA patients tested with photic stimulation had PPR (25%). The probability for PPR changed with age, but only after adjustments for other significant GEA features and with a maximum at 11 years (p=0.011) (Fig. 5) (Online Appendix A). Higher probability for PPR correlated with ‘polyspike slow wave’ (OR=4.9, p=0.002), ‘polyspike’ (OR=32, p<0.001), female gender (OR=3.8, p=0.001), and no medication (OR=3.0, p=0.003).

  • View full-size image.
  • Fig. 5. 

    Age-related predicted probability for PPR in 259 GEA patients provoked with flicker stimulation. The predicted probabilities are adjusted for significant predictors such as ‘polyspike slow wave’, ‘polyspikes’ ‘gender’, and ‘medicated’. Shaded areas are predicted values ±1.96 SE.

Ten of the 48 individuals in the RBS GEA subgroup had PPR (21%). This changed with age (p=0.049), with maximum probability at 10 years.

3.2.5. RBS GEA 

The probability for RBS GEA changed with age (p=0.021) with maximum probability at 11 years (Fig. 6) (Online Appendix A). This age-related probability was also present when adjusted for other significant covariates (p=0.002). Higher probability for RBS GEA correlated with higher GEA-amplitude (p<0.001).

3.2.6. FEA 

Of the 325 EEGs with GEA 148 (46%) contained also FEA. The probability for FEA did not change with age (Online Appendix A). Lower probability for FEA correlated with ‘polyspike slow wave’ (OR=0.43, p=0.026).

Of the 57 patients with RBS GEA 22 (39%) had FEA in addition to GEA, with no correlation to age or other covariates.

3.2.7. GEA-types 

The occurrence of ‘irregular spike/sharp slow wave’ pattern increased with age (p=0.003) relative to ‘spike/sharp slow wave’ (Fig. 7). ‘Hypsarrhythmia’ decreased with age (p=0.016) and was not seen after age 1 year. The other GEA-types were not age-related relative to the reference-group ‘spike/sharp slow wave’.

3.2.8. Current brain-related disease except epilepsy 

The probability of having chronic brain-related disease except epilepsy increased with age (p=0.013) (Fig. 8), also if adjusted for other significant covariates (p=0.005). Patients with ‘polyspike slow wave’ had a lower probability compared to those with ‘spike slow wave’ (OR=0.14, p=0.008). ‘Medicated’ (OR=1.88, p=0.035) and lower GBA-frequency (OR=1.44, p<0.001) correlated with higher probability of having chronic brain-related disease except epilepsy.

Back to Article Outline

4. Discussion 

All EEG BA parameters were clearly affected in EEGs with GEA as compared to controls: AR and GBA frequencies were lower and amplitudes were higher. Using an EEG-centric approach we have identified GEA-related variables correlating with a change in the EEG BA. These correlations were independent of intervening factors such as age, medication, and ‘current brain-related disease except epilepsy’, as these factors were all included as explanatory variables in the multivariate statistical models used. This is in line with a previous quantitative EEG study demonstrating EEG BA slowing in patients with IGE (Clemens et al., 2000). GEA could induce the BA slowing, be a result of the BA slowing, or there could be a common factor causing both the GBA and the BA slowing. A causal connection is not established, but the first and the last possibility seem most plausible.

GEA-amplitude and lower GEA-frequency correlated with slowing of the EEG BA. This is in accordance with the view that low-frequency GEA indicates symptomatic epilepsy (Markand, 2003, Smith, 2005, Pillai and Sperling, 2006). The amplitude of scalp recorded epileptiform spikes and ictal rhythms depends on multiple factors such as the size and localisation of the cortical area involved, orientation of the electrical field, synchrony, and the amplitude of the cortical signals (Ebersole and Hawes-Ebersole, 2007). The decrease of GEA-amplitude with age could in part be due to less synchrony related to a higher proportion of ‘irregular spike/sharp wave’. However, higher GEA-amplitudes remained correlated with GBA slowing also after adjusting for all these factors. Our findings suggest that higher GEA-amplitude leads to a change in brain function independent of other GEA characteristics.

HVS correlated with RBS GEA and with ‘spike/sharp slow wave’ activity. As many as 59% of the RBS GEA subgroup had HVS, in contrast to only 14% of GEA patients without RBS GEA. Hyperventilation provocation is regarded as more effective in generalised than in focal epilepsies (Dalby, 1969, Blume et al., 1973, Holmes et al., 2004). Our findings show that HVS is a common phenomenon, especially in EEGs with typical IGE features. HVS appearing as 3Hz bilateral synchronous GEA is almost obligate in absence epilepsy in children (Binnie and Stefan, 1999). Due to the close relation between HVS and this age-related epilepsy, an inverse age-correlation would be expected for HVS. However, we found no HVS age-dependency; neither in the total group of patients with GEA, nor in the RBS GEA subgroup. Therefore no upper age-limit for hyperventilation provocation should be applied. Yenjun et al. (2003) found no HVS difference between adolescent and adult onset IGE, in agreement with our findings.

PPR was, in agreement with previous studies, age-dependent with a maximum probability at 16 years of age, and with female predominance (Quirk et al., 1995, Stephani et al., 2004, Fisher et al., 2005, Pillai and Sperling, 2006). PPR was more common in patients with no drug treatment and in the GEA-types ‘polyspike slow wave’ and ‘polyspikes’. This supports the clinical experience of a good drug treatment effect on PPR, and fits with the high proportion of PPR in epilepsies associated with myoclonus, especially in JME (Shiraishi et al., 2001, Guerrini and Genton, 2004, Covanis et al., 2004, Benbadis, 2005, Covanis, 2005, Mendez and Brenner, 2006, Trenite, 2006). The probabilities for HVS and PPR did not correlate, and these provocation procedures should therefore be complementary to increase the yield of the EEG registration. Photic stimulation was of less value in older adults than in children and adolescents.

Adult onset IGE is more common than previously acknowledged, and with EEG features resembling those of adolescent IGE (Yenjun et al., 2003, Marini et al., 2003, Nicolson et al., 2004a). The probability for RBS GEA, the EEG hallmark of IGE, was age-dependent in our study, with a maximum at 12 years. The GEA-amplitude decreased with increasing age, and the GEA-type ‘irregular spike/sharp wave’ increased compared to the ‘spike/sharp slow wave’ which was more prominent at young age. These changes in GEA can account for the observed age-dependent decrease in RBS GEA. In contrast, the GEA-frequency was not age-dependent. This may indicate that the underlying epileptic “generator” is the same at different ages, but with a different expression in the scalp EEG due to changes in the neuronal network.

All patients with GEA in their EEGs were included in this study. The term “generalised” has been a somewhat imprecise expression (Wolf, 2005). However, the ILAE Classification Core Group recently decided to retain the terms “focal” and “generalised”, but with the understanding that the former does not necessarily imply that the epileptogenic region is limited to a small circumscribed area, nor does the latter imply that the entire brain is involved in initialisation of the epileptogenic process (Engel, 2006). GEA often has a frontal amplitude dominance, and does not necessarily show in the most posterior electrodes. This is reflected in our definition of GEA. Patients with GEA could have IGE, symptomatic/cryptogenic generalised epilepsies, localisation oriented epilepsies with fast generalisation of the epileptiform activity, as well as the combination of localisation oriented and generalised epilepsy (Koutroumanidis et al., 1999, Nicolson et al., 2004b, Hirsch, 2005, Jeha et al., 2006). The results from this EEG study should therefore not be directly applied on generalised epilepsy.

Major new findings in this study are the correlation between high GEA-amplitude and slowing of the background activity, the accurate age-related probability for specific GEA-types, the lack of correlation between HVS and PPR, and no HVS age-dependency. HVS and PPR are complimentary in increasing the EEG yield, and no upper age-limit for HVS should be applied. This knowledge extends our understanding of GEA and helps the electroencephalographer in weighing the various GEA components.

Back to Article Outline

Acknowledgements 

We are indebted to Dr. Bjørn Karlsen and Dr. Håvard Skeidsvoll, Section of Clinical Neurophysiology, Department of Neurology, Haukeland University Hospital, Bergen, Norway, and Tom Eichele, Department of Biological and Medical Psychology, University of Bergen, Norway for discussion and comments.

Back to Article Outline

Appendix A. Supplementary data 

Download file

download text

Appendix A. AR and GBA functions

Download file

download text

Appendix B. GEA functions

Download file

download text

Appendix C. HVS and PPR probability functions

Download file

download text

Appendix D. FEA, ‘chronic brain-related disease except epilepsy’, and RBS probability functions

Back to Article Outline

References 

  1. ASTM. Table X2.5 EEG waveform/activity identifiers. In: ASTM Designation: E1467–94. Standard specifications for transferring digital neurophysiological data between independent computer systems, 1994. p. 123.
  2. Aurlien H, Gjerde IO, Gilhus NE, Hovstad OG, Karlsen B, Skeidsvoll H. A new way of building a database of EEG findings. Clin Neurophysiol. 1999;110:986–995
  3. Benbadis SR. Practical management issues for idiopathic generalized epilepsies. Epilepsia. 2005;46:125–132
  4. Binnie CD, Stefan H. Modern electroencephalography: its role in epilepsy management. Clin Neurophysiol. 1999;110:1671–1697
  5. Blume WT, David RB, Gomez MR. Generalized sharp and slow-wave complexes – associated clinical features and long-term follow-up. Brain. 1973;96:289–306
  6. Clemens B. Pathological theta oscillations in idiopathic generalised epilepsy. Clin Neurophysiol. 2004;115:1436–1441
  7. Clemens B, Szigeti G, Barta Z. EEG frequency profiles of idiopathic generalised epilepsy syndromes. Epilepsy Res. 2000;42:105–115
  8. Chatrian GE, Bergamini L, Dondey M, Klass DW, Lennox-Buchtal M, Petersén I. A glossary of terms most commonly used by clinical electroencephalographers. In: The International Federation of Societies for Electroencephalography and Clinical Neurophysiology  editors. Recommendations for the Practice of Clinical Neurophysiology. Amsterdam: Elsevier; 1983;p. 11–26
  9. Covanis A. Photosensitivity in idiopathic generalized epilepsies. Epilepsia. 2005;46:67–72
  10. Covanis A, Stodieck SRG, Wilkins AJ. Treatment of photosensitivity. Epilepsia. 2004;45:40–45
  11. Dalby MA. Epilepsy and 3 per second spike and wave rhythms – a clinical, electroencephalographic and prognostic analysis of 346 patients. Acta Neurol Scand. 1969;S 45:6
  12. Ebersole JS, Hawes-Ebersole S. Clinical application of dipole models in the localization of epileptiform activity. J Clin Neurophysiol. 2007;24:120–129
  13. Engel J. Report of the ILAE classification core group. Epilepsia. 2006;47:1558–1568
  14. Fisher RS, Harding G, Erba G, Barkley GL, Wilkins A. Photic- and pattern-induced seizures: a review for the Epilepsy Foundation of America working group. Epilepsia. 2005;46:1426–1441
  15. Gloor P, Ball G, Schaul N. Brain Lesions That Produce Delta Waves in Eeg. Neurology. 1977;27:326–333
  16. Guerrini R. Epilepsy in children. Lancet. 2006;367:499–524
  17. Guerrini R, Genton T. Epileptic syndromes and visually induced seizures. Epilepsia. 2004;45:14–18
  18. Hirsch E. Childhood epilepsy syndromes with both focal and generalized seizures. Acta Neurol Scand. 2005;112:52–56
  19. Hirsch LJ, Brenner RP, Drislane FW, So E, Kaplan PW, Jordan KG, et al. The ACNS subcommittee on research terminology for continuous EEG monitoring: proposed standardized terminology for rhythmic and periodic EEG patterns encountered in critically ill patients. J Clin Neurophysiol. 2005;22:128–135
  20. Holmes MD, Dewaraja AS, Vanhatalo T. Does hyperventilation elicit epileptic seizures?. Epilepsia. 2004;45:618–620
  21. Husain AM, Mebust KA, Radtke RA. Generalized periodic epileptiform discharges: etiologies, relationship to status epilepticus, and prognosis. J Clin Neurophysiol. 1999;16:51–58
  22. Jeha LE, Morris HH, Burgess RC. Coexistence of focal and idiopathic generalized epilepsy in the same patient population. Seizure-Eur J Epilepsy. 2006;15:28–34
  23. Koutroumanidis M, Hennessy HJ, Elwes RDC, Binnie CD, Polkey CE. Coexistence of temporal lobe and idiopathic generalized epilepsies. Neurology. 1999;53:490–495
  24. Koutroumanidis M, Smith S. Use and abuse of EEG in the diagnosis of idiopathic generalized epilepsies. Epilepsia. 2005;46(Suppl. 9):96–107
  25. Low PA. Quantitative sudomotor axon reflex test in normal and neuropathic subjects. Ann Neurol. 1983;14:573–580
  26. Marini C, King MA, Archer JS, Newton MR, Berkovic SF. Idiopathic generalised epilepsy of adult onset: clinical syndromes and genetics. J Neurol Neurosurg Psychiatry. 2003;74:192–196
  27. Markand ON. Pearls, perils, and pitfalls in the use of the electroencephalogram. Semin Neurol. 2003;23:7–46
  28. Mendez OE, Brenner RP. Increasing the yield of EEG. J Clin Neurophysiol. 2006;23:282–293
  29. Nicolson A, Chadwick DW, Smith DF. A comparison of adult onset and “classical” idiopathic generalised epilepsy. J Neurol Neurosurg Psychiatry. 2004;75:72–74
  30. Nicolson A, Chadwick DW, Smith DF. The coexistence of idiopathic generalized epilepsy and partial epilepsy. Epilepsia. 2004;45:682–685
  31. Noachtar S, Binnie C, Ebersole J, Mauguière F, Sakamoto A, Westmoreland B. A glossary of terms most commonly used by clinical electroencephalographers and proposal for the report form for the EEG findings. In:  Deuschl G,  Eisen A editor. Recommendations for the practice of clinical neurophysiology: guidelines of the International Federation of Clinical Neurophysiology. 2nd revised and enlarged ed.. Amsterdam: Elsevier; 1999;p. 21–41
  32. Pfurtscheller G, da Silva FHL. Event-related EEG/MEG synchronization and desynchronization: basic principles. Clin Neurophysiol. 1999;110:1842–1857
  33. Pillai J, Sperling MR. Interictal EEG and the diagnosis of epilepsy. Epilepsia. 2006;47:14–22
  34. Quirk JA, Fish DR, Smith SJM, Sander JWAS, Shorvon SD, Allen PJ. Incidence of photosensitive epilepsy – a prospective national study. Electroenceph Clin Neurophysiol. 1995;95:260–267
  35. Raegrant AD, Strapple C, Barbour PJ. Episodic low-amplitude events – an under-recognized phenomenon in clinical electroencephalography. J Clin Neurophysiol. 1991;8:203–211
  36. Royston P, Sauerbrei W. Building multivariable regression models with continuous covariates in clinical epidemiology – with an emphasis on fractional polynomials. Method Inform Med. 2005;44:561–571
  37. Sam MC, So EL. Significance of epileptiform discharges in patients without epilepsy in the community. Epilepsia. 2001;42:1273–1278
  38. Shiraishi H, Fujiwara T, Inoue Y, Yagi K. Photosensitivity in relation to epileptic syndromes: a survey from an epilepsy center in Japan. Epilepsia. 2001;42:393–397
  39. Smit CM, Wright MJ, Hansell NK, Geffen GM, Martin NG. Genetic variation of individual alpha frequency (IAF) and alpha power in a large adolescent twin sample. Int J Psychophysiol. 2006;61:235–243
  40. Smith SJM. EEG in the diagnosis, classification, and management of patients with epilepsy. J Neurol Neurosur Psychiatry. 2005;76:2–7
  41. Stephani U, Tauer U, Koeleman T, Pinto D, Neubauer BA, Lindhout D. Genetics of photosensitivity (photoparoxysmal response): a review. Epilepsia. 2004;45:19–23
  42. Trenite DGAK. Photosensitivity, visually sensitive seizures and epilepsies. Epilepsy Res. 2006;70:S269–S279
  43. Wada J, Rasmussen T. Intracarotid injection of sodium amytal for the lateralization of cerebral speech dominance – experimental and clinical observations. J Neurosurg. 1960;17:266–282
  44. Waltz S, Christen HJ, Doose H. The different patterns of the photoparoxysmal response – a genetic-study. Electroenceph clin Neurophysiol. 1992;83:138–145
  45. Westmoreland BF, Klass DW. Unusual EEG patterns. J Clin Neurophysiol. 1990;7:209–228
  46. WHO Collaborating Centre for Drug Statistics Methodology. ATC/DDD Index 2008. http://www.whocc.no/atcddd/ Last accessed November 2008.
  47. Wolf P. Historical aspects of idiopathic generalized epilepsies. Epilepsia. 2005;46:7–9
  48. Yenjun S, Harvey AS, Marini C, Newton MR, King MA, Berkovic SF. EEG in adult-onset idiopathic generalized epilepsy. Epilepsia. 2003;44:252–256

PII: S1388-2457(08)01021-3

doi:10.1016/j.clinph.2008.10.149

Clinical Neurophysiology
Volume 120, Issue 1 , Pages 3-10, January 2009

Article Tools

* Image Usage & Resolution