Electroencephalography and clinical Neurophysiology, 82 (1992) 408-414 © 1992 Elsevier Scientific Publishers Ireland, Ltd. 0013-4649/92/$05.00
Sequential E E G m a p p i n g may differentiate "epileptic" from "non-epileptic" rolandic spikes W. Van der Meij a, A.C. Van Huffelen a, G.H. Wieneke a and J. Willemse b aDepartment of Clinical Neurophysiology and b Department of Child Neurology, University Hospital Utrecht, Utrecht (The Netherlands) (Accepted for publication: 10 December 1991)
Sequential topographic mapping was performed to differentiate "epileptic" from "non-epileptic" rolandic spikes. Twenty-four children without any indication of organic brain lesion were divided into a group with epilepsy and a group without epilepsy. The group with epilepsy was subdivided into "classical BECT" (benign focal epilepsy of childhood with centro-temporal spikes) and "non-classical BECT." Sequential mapping of the rolandic spikes revealed two different topographic patterns: a pattern of stationary potential fields and a pattern of non-stationary potential fields. The topographic pattern of stationary potential fields was morphologically represented by a single spike-and-wave complex whereas that of non-stationary potential fields was morphologically represented by a "double" spike-and-wave complex. Among the non-stationary topographic patterns represented by a "double" spike, one specific sequence of changes of potential fields was found. This sequence started with a dipolar field, with the negative pole in the frontal region and the positive pole in the centro-temporal region, morphologically represented by the small first spike of the "double" spike-and-wave complex. This dipolar field, changed to a unipolar or dipolar field, with a negative potential field in the centro-temporal region and, sometimes, a simultaneous positive potential field in the frontal region, morphologically represented by the prominent rolandic spike. This characteristic pattern was found to be significantly related to classical BECT. Key wards: Rolandic spike; Double spike; Sequential topographic analysis
The rolandic spike is a characteristic finding in the interictal EEG of children with benign focal epilepsy of childhood with centro-temporal spikes (BECT; Nayrac and Beaussart 1958; Lombroso 1967). The typical seizures in this idiopathic form of focal epilepsy have an onset of sensory a n d / o r motor manifestations in the oropharyngeal region (Loiseau and Beaussart 1973). The location of the accompanying epileptiform EEG abnormality in the centro-temporal region or, more precisely, in the inferior central region (Faure and Loiseau 1960; Lombroso 1967; Beaussart 1972) corresponds with the cortical area in which this oropharyngeal region is represented (Penfield and Jasper 1954). However, the rolandic spike is not exclusively associated with this specific ictal symptomatology. Other seizure types, such as unilateral seizures or generalized seizures, are also encountered in BECT. The cornerstones for the diagnosis BECT are epileptic seizures, the absence of any neurological deficit and the presence of rolandic spikes in the EEG, but the semiology of the seizures is not decisive (Lerman 1985). Thus, an epileptic syndrome with unilateral seizures without in-
Correspondence to: W. Van der Meij, Department of Clinical Neurophysiology, University Hospital Utrecht, P.O. Box 85500, 3508 GA Utrecht (The Netherlands).
volvement of the oropharyngeal region or with generalized seizures will be diagnosed as BECT when other clinical and EEG criteria are fulfilled. Moreover, the rolandic spike is not restricted to BECT, or even to epilepsy (Lerman and Kivity-Ephraim 1981). The rolandic spike also occurs in the EEG of children with a neurological deficit with or without epilepsy (Gibbs and Gibbs 1964; Trojaborg 1966; Roger et al. 1972), in the EEG of non-epileptic children with various complaints including behavioural problems (Green 1961; Fois et al. 1968), and even in normal children (Eeg-Olofsson et al. 1971; Cavazzuti et al. 1980). Several studies have been devoted to the topography of the rolandic spikes in BECT and describe a characteristic location and distribution of the potential fields of the interictal spikes, i.e., negativity in the centrotemporal region and simultaneous positivity in the frontal region (Gibbs and Gibbs 1952; Faure and Loiseau 1959; Blume 1982; Gregory and Wong 1984; Liiders et al. 1987; Wong 1989). A few studies report ictal EEG findings (Dalla Bernardina and Tassinari 1975; Gutierrez et al. 1990; Roger et al. 1990). To our knowledge, no study compared the characteristics of "epileptic" with "non-epileptic" rolandic spikes. The question is raised whether there are specific EEG features that can distinguish between them. To answer
SEQUENTIAL EEG MAPPING OF ROLANDIC SPIKES
this question, we sequentially mapped rolandic spikes and examined their topographic distribution.
Children were selected for this study on the basis of EEG and clinical criteria. Of 3123 EEGs performed in children aged 0-16 years in whom an EEG was recorded during the period 1988-1990 on request of pediatricians, child neurologists or child psychiatrists, the patients who showed rolandic spikes were eligible for this study. Rolandic spikes were defined as spikes or sharp waves localized in the centro-temporal region, in an otherwise normal EEG. Clinical criteria were set t o exclude children with organic brain lesion. An extensive history was taken, using a questionnaire. The symptomatology of the seizures was questioned conscientiously. All children underwent standardized neurological examination (Touwen 1979) and their brains were examined by CT scan a n d / o r MRI. Children with behavioural or learning problems were subjected to a neuropsychological examination (the WISC-revised, a Continuous Performance-Vigilance task, the Benton visual retention test, the Trailmaking task (A + B), the Stroop test and a motor performance task consisting of tapping, pegboard, aiming trace tracking and steadiness) with special attention being paid to indications of organic brain disorders. Children with neurological deficit, CT scan a n d / o r MRI abnormalities or neuropsychological indications of organic brain lesion were excluded. Twenty-four children, mean age 9.4 years (range: 6-13 years) fulfilled all criteria and were selected for the study. On the basis of clinical data, the children were placed in one of two groups, one with epilepsy, the other without epilepsy. The epilepsy group was subdivided into a "classical BECT" group and a "nonclassical BECT" group. The terminology "classical BECT" was applied to those children with characteristic focal seizures with ictal symptoms starting in or restricted to the oropharyngeal region (n = 8). Children with unilateral seizures without involvement of the oropharyngeal region (n = 2) or generalized seizures (n = 5) were classified as having "non-classical BECT." The group without epilepsy (n = 9) included 4 children with behavioural problems, 3 children with learning disorders, 1 child suspected of having Bell's palsy and 1 child who exhibited vasovagal collapse.
Fig. 1. Electrode positions. The filled circles indicate the electrode positions of the international 10-20 system, the open circles the interposed electrodes. The interposed positions are designated according to the neighbouring electrode positions of the 10-20 system, i.e., F5C5 indicates the interposed electrode midway between electrode F5 and C5.
electrodes were interposed in the centres of the quadrangles or triangles formed by the electrodes of the 10-20 system (Buchsbaum et al. 1985) and at extrapolated positions on the horizontal line drawn at the level of the nasion and inion (Chatrian et al. 1985), midway between two positions of the 10-20 system (Fig. 1). Conventional 10 mm Ag-AgC1 electrodes were used and fixed to the scalp by means of collodion 3%. Conductive jelly was applied. The electrode impedance was always less than 2 k ~ . When in the standard EEG rolandic spikes were detected, a common reference montage was chosen for further analysis with the preauricular electrode contralateral to the centro-temporal spike focus as a reference. A Nicolet Pathfinder 2 system was used for the 32-channel EEG recording. The bandwidth was 0.5-70 Hz ( - 3 dB). The sensitivity of the recorded EEG was adapted to the amplitude of the spikes. The EEG signal was sampled with a frequency of 250 Hz and the recording time was 12 min: 3 rain in the eyes open condition, 3 min in the eyes closed condition and 6 min in that of these two conditions that gave a maximum number of spikes and a minimum number of artifacts.
Data processing and analysis Methods
Data acquisition A 32-channel interictal EEG was recorded. In addition to the 21 electrodes of the 10-20 system, 11
The EEG signal was displayed on the monitor of the Pathfinder 2 system. Spikes with an amplitude of at least 1.5 times the mean amplitude of the background activity were selected visually. The channel that revealed the greatest amplitude of the spike was determined. The selected spikes were aligned at their maxi-
410 mum negativity in this channel and averaged. The number of spikes contributing to the averaged spikeand-wave complex ranged from 11 to 50 spikes (median: 42). A baseline correction was applied by using the mean amplitude of an averaged EEG segment of 204 msec ending 52 msec before the spike maximum, as the baseline amplitude. The amplitudes of the different components of the averaged spike-and-wave complex were measured in relation to this baseline. All 32 channels were used to study the topography. For each averaged spike the ascending phase, the negative maximum and the descending phase, from 64 msec before to 56 msec after the negative maximum of the spike, were analysed by means of sequential topographic mapping with steps of 4 msec. In these topographic maps, the amplitudes are expressed in a rainbow colour scale with 45 grades. A voltage scale that was symmetrical around zero was applied. The negative extreme in this scale is represented by the colour white, the positive extreme by blue, and zero by yellow. When in a patient more than one EEG was recorded during the study period that contained rolandic spikes, the first EEG showing these spikes was used for analysis. When rolandic spikes occurred independently over both hemispheres, the side that corresponded with the unilateral seizure symptomatology was chosen for analysis. When the seizure symptomatology showed no lateralization, the side at which the spikes occurred most frequently was chosen for analysis. In the 6 children with bilateral spikes, the morphological and topographical features were identical for both loci and therefore the choice of analysing only one of these foci did not influence the results. Thus, 24 averaged spikes of 24 children were available for analysis.
Topographic analysis of the rolandic spike Distribution of the potential fields, at the maximum of the averaged spike. The negative maximum of the averaged spike was found at one of the interposed or 10-20 electrodes in the centro-temporal region (Table I, rolandic spike). In 15 patients simultaneously with this negative maximum a positive potential field occurred over several electrodes in the frontal region (Table I, rolandic spike). No differences between the two groups were found with respect to the location of the spike maxima and their potential field distribution. The spike maxima were localized at an electrode of the 10-20 system in 9/24 patients (37.5%) and at an interposed electrode in 15/24 patients (62.5%). In these 15 patients the electrode position was determined at which the negative maximum would have been localized if only the electrodes of the 10-20 system had been used. The mean ratio of the ampli-
w. VAN DER MEIJ ET AL. tude of the spike maximum at the interposed electrode to the amplitude at this 10-20 electrode position was 1.2 (S.D.: 0.2).
Spatio-temporal analysis of the ascending and descending phase of the averaged spike. Sequential topographic analysis of the ascending and descending phases of the averaged spikes indicated that there were 2 different topographic patterns: (1) A stationary pattern (n = 13): the locations of the negative and, when present, the positive potential fields were stationary during the ascending and descending phases of the spike and similar to the locations at the maximum amplitude of the spike (Fig. 2.1). Thus, only the amplitude of the potential fields changed, not the topography. (2) A non-stationary pattern in which the locations of the negative and positive potential fields changed over time (n = 11). A sequence was observed starting with a dipolar field with a maximum negativity anteriorly and a maximum positivity posteriorly. This changed into a unipolar field or to a dipolar field that had the opposite polarity compared to the initial dipolar field (Fig. 2.2). The negative extremes of the first dipolar field were located (Table I, first spike) at the frontal electrode (F3/F4) in 4 patients, at the fronto-central electrode (F2C2) in 3 patients, at the fronto-¢entrotemporal electrode (F5C5/F6C6) in 2 patients, in 1 patient at the central electrode (C3) and in 1 patient at the centro-parietal electrode (CIP1). The positive extremes of the first dipolar field were located (Table I, first spike) at the centro-parieto-temporal electrode (C5P5/C6P6) in 6 patients, in 1 patient at electrodes T3 and T5, in 2 patients at the temporo-ocipital electrode ( T 5 0 1 / T 6 0 2 ) and in 2 patients at the occipital electrode (O1/O2). The locations of the extremes of the unipolar or dipolar field following the first dipolar field were similar to those found at the maximum amplitude of the stationary pattern (Table I, rolandic spike). Within this non-stationary pattern, a pattern of a similar sequence of topography of potential fields was noted in 6 patients (Table I, (*) and Fig. 2.2): a dipolar field with the negative extreme at the frontal (F3/F4) or fronto-central (F1C1/F2C2) electrode and the positive extreme at the centro-parieto-temporal (C5P5/ C6P6) electrode (Fig. 2.2, first topographic map) changed into a unipolar field with negativity at the fronto-centro-temporal (F6C6) electrode or the central (C3/C4) electrode (Fig. 2.2, second map). In 3 patients the unipolar field subsequently changed into a dipolar field with the negative extreme in the centro-temporal region and a positive field in the frontal region (Fig. 2.2, third map). The 6 patients in whom this specific non-stationary topographic pattern was found, all belonged to the epilepsy group (all except one to the classical BECT group). This specific sequence of
SEQUENTIAL EEG MAPPING OF ROLANDIC SPIKES
Fig. 2. 1: topographic analysis by means of sequential mapping of the averaged rolandic spike-and-wave complex of patient 22. In the ascending and descending phases of the spike the development and decline of a stationary dipolar field with a maximum at moment 1024 msec was observed. The orientation of this dipolar field was constant. The negative extreme is represented by the colour white, the positive extreme by blue, and zero by yellow. 2: sequential topographic analysis of the "double" spike in patient 2 showing a specific non-stationary pattern: a dipolar field with (at moment 996) the negative extreme located at electrode F4 and the positive extreme at electrode C6P6, morphologically represented by the small first spike of the "double" spike in Fig. 3.2. This dipolar field changed via an intermediate phase with a unipolar negative potential field in the centro-temporal region to a dipolar field with (at moment 1024) the negative extreme at C6P6 and the positive extreme in the frontal region, represented by the prominent rolandic spike in Fig. 3.2. The negative extreme is represented by the colour white, the positive extreme by blue, and zero by yellow.
changing locations of potential fields was not observed in the group of patients without epilepsy. In the remaining 5/11 patients in whom a non-stationary pattern was found, either the negative extreme was not located frontally (patients 19, 21) or the positive extreme was not located centro-parieto-temporally (patients 10, 18, 19, 20, 21). Statistical analysis revealed no significant differences in the occurrence of the stationary and non-stationary topographic patterns between the groups with
epilepsy and without epilepsy. However, when tested with the Fisher exact test the occurrence of the specific non-stationary topographic pattern differed significantly between the classical BECT group and the group without epilepsy (P = 0.025) but not between the classical BECT and the non-classical BECT groups.
Morphology of the rolandic spike-and-wavecomplex As with the sequential topographic analysis 2 morphological patterns were found:
W. V A N D E R M E I J E T AL. 1024
Location of the negative and positive extremes of the rolandic spike-and-wave complex in 24 patients: patients 1-15: group with epilepsy (classical BECT: 1-8; non-classical BECT: 8-15); patients 16-24: group without epilepsy. Location of the negative and positive extremes of the first spike (if present) of the "double" spike-and-wave complex. Patient
First spike (if present)
Frontal positivity 1
Classical BECT 1 2 3 4 5 6 7 8
* * * *
F4 F4 F4 F2C2
C6P6 C6P6 C6P6 C6P6
F6C6 F6C6 C4 C4 C3 C6P6 C5P5 P4
+ + + + + +
Non-classical BECT 9 10 11 12 * 13 14 15
F5C5 F5C5 T4 C4 C5P5 C5P5 C5P5
+ + + +
F5C6 F6C6 T4 T3 CA C5P5 C5P5 C5P5 C2P2
+ + + + + -
No epilepsy 16 17 18 19 20 21 22 23 24
F6C6 C1P1 F2C2 C3
T602 T3-T5 02 T501
* Patients with the specific non-stationary topographic pattern. Loc a t i o n of the negative extreme of the prominent rolandic spike. A
' 1024 I I
F~co . . . . . .
"-----3 i v
/ ' - , .
,, i '
6o~vT 2 0 0
rn s e ¢
Fig. 3. 1: the averaged rolandic spike-and-wave complex of patient 22. Positivity in the frontal region occurred simultaneously with the negative extreme at C5P5. T h i s morphological pattern represents the topographic pattern of Fig. 2.1. 2: the averaged rolandic spike-and-
wave complex of patient 2 demonstrating a "double" spike-and-wave complex. A small amplitude spike reached its negative m a x i m u m earlier than the prominent rolandic spike. The maximum negativity of the small first spike was located at electrode F4, the simultaneously occurring m a x i m u m positivity was located at electrode C6P6. The maximum negativity of the prominent rolandic spike was found at electrode F6C6, the simultaneous maximum positivity in the frontal region. This morphological pattern represents the topographic pattern in Fig. 2.2.
simultaneous positive extreme, when present, was always located in the frontal region.
(1) The topographic pattern of stationary potential fields was morphologically represented by the single spike-and-wave complex (Fig. 3.1; n = 13). (2) The topographic pattern of non-stationary potential fields was morphologically represented by a complex, consisting of a "double" spike followed by a slow wave (Fig. 3.2; n = 11). The first dipolar field in the topographic sequence of non-stationary potential fields was represented by the small first spike and the subsequent unipolar or dipolar field by the prominent rolandic spike. The topographic pattern of non-stationary potential fields, morphologically represented by the "double" spike-and-wave complex, included the 6 patients with the specific topographic pattern of changing potential field locations. This "double" spike configuration consisted of a small amplitude spike of which the negative maximum was smaller and earlier (varying
from 16 to 36 msec in the 11 patients) than the negative maximum of the prominent rolandic spike. The first spike reached its negative maximum in the ascending phase of the rolandic spike. Sometimes the first spike was noticed to be superimposed on the ascending phase of the rolandic spike and thus observed in the same channel (n = 2; Fig. 3.2, electrode position F6C6), but in most patients (n = 9) the first spike and the rolandic spike were most dearly identifiable in different channels (Fig. 3.2, electrode positions C6P6 and F4).
Discussion In recent years several studies have been devoted to the clinical and EEG aspects of BFEC with centrotemporal spikes (BECT) (for review see" Lerman 1985; Liiders et al. 1987).
SEQUENTIAL EEG MAPPING OF ROLANDIC SPIKES
The distribution of potential fields of rolandic spikes in BECT is dipolar with a negative pole in the centralmidtemporal region and a positive pole in the frontal region (Gibbs and Gibbs 1952; Faure and Loiseau 1959; Blume 1982; Liiders et al. 1987). This distribution of potential fields is considered to be almost pathognomonic for BECT (Liiders et al. 1987). Gregory and Wong (1984) and Wong (1989) assumed that a single generator orientated tangentially to the surface of the scalp and located in the lower rolandic region was the source of this dipolar field. In 7 out of the 12 centro-temporal spike-and-wave complexes studied by Gregory and Wong (1984) in patients with BECT, a complex onset was described with a low voltage "onset dipole" field consisting of positivity in the midposterior-temporal region and negativity in the ipsilateral frontal region, reversing to a high voltage "main dipole" with negativity in the temporal region and positivity in the frontal region. To explain this pattern of spike onset the authors postulated that there is a single generator that reverses its polarity within one spike. However, the above-mentioned studies concern interictal EEG findings. Little has been published about ictal records. Dalla Bernardina and Tassinari (1975) reported an ictal EEG record in a 10-year-old boy with BECT. At the onset of the seizure, in which myoclonus of the right face occurred, rhythmic spikes were found in the left temporal and centro-parietal regions. Roger et al. (1990) recently reported an ictal EEG record coinciding with motor signs in the right face. Rhythmic spikes were found which started in the left central region. Gutierrez et al. (1990) recorded a subclinical EEG seizure in a child diagnosed as having BECT the very morning after the occurrence of a nocturnal focal seizure with manifestations in the oropharyngeal region. During sleep an EEG discharge was observed which consisted of spikes with a potential field distribution that was reversed in comparison with that of the interictal spikes, i.e., the "ictal" spikes were negative in the frontal region with simultaneous positivity at the central and midtemporal electrodes. The present study differs from previous reports with regard to both the clinical criteria and neurophysiological methods used. Only children without any indication of neurological deficit, based on neurological, neuroradiological and neuropsychological examination, were selected for analysis, thereby excluding children with a brain lesion. By dividing the patients into two groups, one with and one without epilepsy, we were able to examine, unlike other authors (Gregory and Wong 1984; Gutierrez et al. 1990), whether specific spike characteristics were associated with epilepsy. The neurophysiological methods used in this study provided both high spatial and temporal resolution. The extension of the 21 electrodes of the 10-20 system
to a 32-electrode system, with electrodes at interposed positions, provided optimal coverage of the centrotemporal region. The increase in spatial resolution thus obtained has two advantages: the localization and the amplitude measurements of the spike maxima are more accurate. The spike maxima frequently appeared to be located at an interposed electrode rather than at an electrode of the 10-20 system, and the amplitudes were on average 20% larger than those found with the 10-20 system. This facilitated the discrimination of a specific non-stationary topographic pattern. The precision of time analysis of an analogue EEG signal recorded on paper and the assessment of the complexity of spatial data are restricted when done by eye. Conversion of data to a digital signal, sequential analysis of the signal at defined intervals of short duration, and transformation of the data into a topographic map improve the accuracy of time and space measurements. The sample frequency of 250 Hz, used in the present study, which gives a temporal resolution of 4 msec, enabled an accurate analysis of the rolandic spike-and-wave complex. Topographic analysis of the maximum negativity of the rolandic spike did not reveal a specific location of the maximum negativity or a potential field distribution specific for one of the groups of subjects. In the group with epilepsy as well as the group without epilepsy, a dipolar potential field with negativity in the centrotemporal region and positivity in the frontal region was found in some patients and a unipolar negative potential field in the centro-temporal region in others. Sequential mapping of the rolandic spike, however, revealed two different topographic patterns, which could be described as a pattern of stationary potential fields and a pattern of non-stationary potential fields. Morphologically, the stationary pattern appeared to be represented by a single spike-and-wave complex, the non-stationary topographic pattern by a "double" spike-and-wave complex. The pattern of stationary potential fields, represented by a single spike-and-wave complex, was equally divided over the two patient groups as was the pattern of non-stationary potential fields, represented by a "double" spike-and-wave complex. However, within the pattern of non-stationary potential fields, represented by a "double" spike-andwave complex, a specific topographic pattern was found which only occurred in patients with epilepsy, especially in patients with classical BECT. In this pattern a sequence starting with a dipolar field with one pole in the frontal region and the other pole in the centrotemporal region was noted, morphologically represented by the first spike of the "double" spike-and-wave complex, changing into a potential field distribution with negativity in the centro-temporal region and, when present, simultaneous positivity in the frontal region, represented by the prominent rolandic spike. Gregory
and Wong (1984) reported an onset (type I) with a similar first dipolar field in their study of rolandic spikes in BECT. In the present study, this phenomenon appeared to be significantly related to classical BECT. Interestingly, the "ictal" spikes reported by Gutierfez et al. (1990) showed, as far as visual assessment of the 18-channel EEG allows, a distribution of potential fields similar to the specific topographic pattern of changing potential fields represented by the first spike of the "double" spike-and-wave complex. The results suggest that the prominent rolandic spike does not represent a specific clinical entity as it is found in the EEG of children with or without epilepsy. On the contrary, the occurrence of a first spike before the prominent rolandic spike with a specific pattern of changing potential fields represents an EEG feature that is significantly related to classical BECT. The first spike of this specific "double" spike-and-wave complex conforms with the "ictal" spikes of the study of Guttierez et al. (1990). One may hypothesize that this first spike represents a true "epileptic" spike and triggers the non-specific prominent rolandic spike. A study is in progress to calculate the localization of the sources of scalp recorded potential fields by means of dipole modelling. This might enable us to localize the source of the dipolar field responsible for the first spike of the "double" spike, different from the localization of the dipole source responsible for the second, prominent, rolandic spike and correlating with the characteristic focal seizure symptomatology in BECT. This research was supported by CLEO (TNO National Epilepsy Research Committee). We would like to thank Mrs. G. Konings-Van der Steen, Mrs. M. Musbach and R.P. Schoobaar for their technical assistance.
Beaussart, M. Benign epilepsy of children with rolandic (centro-temporal) paroxysmal loci. Epilepsia, 1972, 13: 795-811. Blume, W.T. Atlas of Pediatric Electro-encephalographpy. Raven Press, New York, 1982: 144-145. Buchsbaum, M.S., Hazlett, E., Sicotte, N., Ball, R. and Johnson, S. Geometric and scaling issues in topographic electroencephalography. In: F.H. Duffy (Ed.), Topographic Mapping of Brain Electrical Activity. Butterworth, London, 1986: 325-337. Cavazzuti, G.B., Capella, L. and Nalin, A. Longitudinal study of eleptiform EEG patterns in normal children. Epilepsia, 1980, 21: 43-55. Chatrian, G.E., Lettich, E. and Nelson, P.L Ten percent electrode system for topographic studies of spontaneous and evoked EEG activities. Am. J. EEG Technol., 1985, 25: 83-92. Dalla Bernardina, B. and Tassinari, C.A. EEG of a nocturnal seizure
W. VAN DER ME1J ET AL. in a patient with benign epilepsy of childhood with rolandic spikes. Epilepsia, 1975, 16: 497-501. Eeg-Olofsson, O., Peters6n, L. and Selleden, V. The development of the electroencephalogram in normal children from the age of 1 through 15 years. Paroxymal activity. Neurop/idiatrie, 1971, 4: 375-404. Faure, J. and Loiseau, P. Une corr61ation clinique particuli~re des pointes-ondes sans signification focale. Rev. Neurol., 1960, 102: 399-406. Fois, A., Borgheresi, S. and Luti, E. Clinical correlates of focal epileptic discharges in children without seizures. A study of 110 cases. Helv. Paediat. Acta, 1968, 3: 257-265. Gibbs, F.A. and Gibbs, E.L. Jacksonian seizures and focal convulsions: midtemporal epileptic foci. In: Atlas of Electroencephalography, Vol. 2. Addison-Wesley, Boston, MA, 1952: 214-222. Gibbs, F.A. and Gibbs, E.L. Significance of a specific finding: midtemporal spike focus. In: Atlas of Electroencephalography, Vol. 3. Addison-Wesley, Boston, MA, 1964: 52-53. Green, J.B. Association of behaviour disorder with an electroencephalographic focus in children without seizures. Neurology, 1961, 11: 337-344. Gregory, D.L. and Wong, P.K. Topographical analysis of the centrotemporal discharges in benign rolandic epilepsy of childhood. Epilepsia, 1984, 25: 703-711. Gutierrez, A.R., Brick, J.F. and Bodensteiner, J. Dipole reversal: an ictal feature of benign partial epilepsy with centrotemporal spikes. Epilepsia, 1990, 31: 544-548. Lerman, P. Benign partial epilepsy with centro-temporal spikes. In: J. Roger, C. Dravet, M. Bureau, F.E. Dreifuss and P. Wolf (Eds.), Epileptic Syndromes in Infancy, Childhood and Adolescence. John Libbey Eurotext, London, 1985: 150-158. Lerman, P. and Kivity-Ephraim, S. Focal epileptic EEG discharges in children not suffering from clinical epilepsy: etiology, clinical significance and management. Epilepsia, 1981, 22: 551-558. Loiseau, P. and Beaussart, M. The seizures of benign childhood epilepsy with rolandic paroxysmal discharges. Epilepsia, 1973, 14: 381-389. Lombroso, C.S. Sylvian seizures and mid-temporal spike foci in children. Arch. Neurol., 1967, 17: 52-59. Liiders, H., Lesser, R.P., Dinner, D.S. and Morris, III, H.H. Benign focal epilepsy of childhood. In: H. Lfiders and R.P. Lesser (Eds.), Epilepsy. Electroclinical Syndromes. Springer, Berlin, 1987: 303346. Nayrac, P. and Beaussart, M. Les pointes-ondes pr6rolandiques: expression EEG tr~s particuli~re. Rev. Neurol., 1958, 99: 201-206. Penfield, W. and Jasper, H. Epilepsy and the Functional Anatomy of the Human Brain. Little, Brown and Co., Boston, MA, 1954. Roger, J., Bureau, M., Dravet, C., Dalla Bernardina, B., Tassinari, C.A., Revol, M. and Challamel, M.J. Les h(~mipl6gies c6r6brales infantiles. Rev. EEG Neurophysiol., 1972, 2: 5-28. Roger, J., Bureau, M., Genton, P. and Dravet, C. Idiopathic partial epilepsies. In: M. Dam and L. Gram (Eds.), Comprehensive Epileptology. Raven Press, New York, 1990: 155-170. Touwen, B.C.L. Examination of the child with minor neurological dysfunction. In: Clinics in Developmental Medicine. Spastic International Medical Publiciations. William Heineman Medical Books, London, 1977: 71. Trojaborg, W. Focal spike discharges in children, a longitudinal study. Dissertation, Copenhagen, 1966. Wong, P.K. Stability of source estimates in rolandic spikes. Brain Topogr., 1989, 2: 31-36.