- Email: [email protected]
Combined influence of cyclic arousability and EEG synchrony on generalized interictal discharges within the sleep cycle
Epilepsy Research 44 (2001) 7 – 18 www.elsevier.com/locate/epilepsyres
Combined influence of cyclic arousability and EEG synchrony on generalized interictal discharges within the sleep cycle Liborio Parrino, Arianna Smerieri, Mario Giovanni Terzano * Istituto di Neurologia, Uni6ersita` de Parma, Via del Quartiere 4, 43100 Parma, Italy Received 19 April 2000; received in revised form 20 September 2000; accepted 26 September 2000
Abstract Purpose: to analyze the activating role of cyclic alternating pattern (CAP) and EEG synchrony on generalized interictal paroxysms in the first part of the night, when all sleep patterns are represented. Methods: nocturnal polysomnographic investigation was accomplished on a randomized series of 18 subjects with an active form of primary generalized epilepsy (PGE), but only six patients showed a complete and regular profile of the first two sleep cycles (SCs). Completeness and regularity of the selected SCs consisted in the absence of intervening wakefulness, in the presence of all sleep stages, and in the identification of three main units, (a) a descending branch, dominated by the build-up of EEG synchrony in the transition from light to deep non-rapid eye movement (NREM) sleep; (b) a trough, where the magnitude of EEG synchrony is greatest and gives rise to stages 3 and 4; (c) an ascending branch characterized by a decrease of EEG synchrony preceding the onset of rapid eye movement (REM) sleep. Generalized paroxysms were evaluated in terms of discharge rates (number of interictal bursts per minute of sleep) and distribution within the investigated sleep parameters. Results: the discharge rates decreased from SC1 to SC2, with higher values quantified during NREM sleep (mean, 2.8) compared with REM sleep (mean, 0.8). Both SCs showed a progressive decrease of activation across the three units, from the highest discharge rates reached during the descending branches (mean, 3.6) to the more attenuated discharge rates during the troughs (mean, 2.4) down to the lowest rates during the ascending limbs (mean, 1.1). The magnitude of activation during the descending branches was closely related to the CAP condition (mean, 5.2) and to the powerful effect of phase A (mean, 13.9). The great majority (82%) of EEG discharges occurring in phase A were distributed within the A1 subtypes (identified by sequences of k-complexes or delta bursts). Conclusions: within the first two SCs, the features of NREM sleep endowed with the major activating power on generalized bursts are represented by the rise of EEG synchrony (descending branch) and by the A phases of CAP involved in the regulation of its build-up. © 2001 Elsevier Science B.V. All rights reserved.
Abbre6iations: SC, sleep cycle; SC units, descending branch, trough, ascending branch; SC1, first sleep cycle; SC2, second sleep cycle; PGE, primary generalized epilepsy; SI, spike index; CAP, cyclic alternating pattern (arousal fluctuation); NCAP, non-cyclic alternating pattern (stable sleep); REM sleep, rapid eye movement sleep; NREMsleep, non-REM sleep; SWS, slow wave sleep (stages 3 and 4); ASDA, american sleep disorders association. * Corresponding author. Tel.: + 39-0521-287913; fax: + 39-0521-287913. E-mail address: [email protected] (M.G. Terzano). 0920-1211/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 0 - 1 2 1 1 ( 0 0 ) 0 0 1 9 2 - 3
8
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
Keywords: Cyclic alternating pattern; EEG; Generalized epilepsy; Interictal discharges; Sleep cycle
1. Introduction The physiological variations of arousal across the 24 h cycle represent an important factor in the modulation of epileptic phenomena. There is extensive agreement that epileptic manifestations more likely appear, when the level of arousal is reduced. In particular, generalized interictal discharges mostly occur in the resting non-rapid eye movement (NREM) sleep, when EEG rhythms present high amplitudes and slow frequencies (Steriade et al., 1990), and the number of generalized bursts is lowest during the more wake-like rapid eye movement (REM) sleep (Shouse et al., 1996). It has been postulated that generalized EEG discharges share common anatomical pathways with neural mechanisms that control EEG synchrony and arousal modulation during sleep (Terzano et al., 1992). The coexistence of physiological processes and pathological factors makes epilepsy a cerebral disorder in which the interictal manifestations may appear discontinuously, but not randomly. Within NREM sleep, the transitional phases, such as sleep onset, awakenings from sleep, stage transitions, activate interictal discharges more powerfully than the fully developed stages (Janz, 1962). These critical periods of poorly consolidated vigilance are characterized by EEG fluctuations and cyclic arousability. Quantification of these periodic changes is based on the scoring rules for coding the cyclic alternating pattern (CAP) (Terzano et al., 1985). CAP is an EEG phenomenon organized in sequences that occupy wide stretches of NREM sleep (Parrino et al., 1998). CAP sequences are composed of repetitive transient patterns lasting 8 – 15 s (phase A) separated by intervals of 15 – 20 s (phase B). The A phases are constituted by the stage-related arousal correlates (k-complexes, delta bursts, arousals), while the B phases are composed of the stage-related background activities. During
CAP, the EEG rhythms of sleep oscillate with periodic excitatory (phase A) and inhibitory (phase B) swings. It is during CAP that a number of epileptic events occurring in NREM sleep are seen with the highest frequency. In primary generalized epilepsy (PGE), the majority of interictal EEG abnormalities are associated with a CAP sequence and are triggered by an A phase (Terzano et al., 1989). The increase of generalized spike-wave paroxysms is apparent during CAP in all NREM stages, but no evaluation has been carried out on their position within the sleep cycle (SC). Conventionally, a SC is defined by the alternation of NREM and REM sleep. The NREM portion of the SC is composed of a succession of stages, which determine three distinct units characterized by different degrees of EEG synchrony. This succession involves a descending branch sloping from the more superficial to the deeper NREM stages, in which EEG synchrony undergoes a progressive increase ; a central trough of deep sleep, in which EEG synchrony is highest; an ascending branch where lighter NREM stages precede the onset of REM sleep, in which EEG synchrony decreases (Merica and Gaillard, 1986). CAP sequences interact with EEG synchrony showing different phase A patterns in the three units of NREM sleep (Ferrillo et al., 1997; Terzano et al., 2000). The present study aimed at assessing the combined influence of sleep patterns and arousal instability on PGE interictal discharges during the first part of the night, when the variations of EEG synchrony are more pronounced. For this purpose, we focused upon the initial hours of continuous sleep analyzing the amount of CAP sequences and generalized discharges within ‘ideal’ SCs, i.e. the ones uninterrupted by intervening wakefulness and, in which all stages are represented and linked in a regular succession of a descending branch, a trough and an ascending branch.
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
2. Material and methods The study was carried out on attended laboratory-recorded polysomnograms obtained at the Parma Sleep Disorders Center in subjects with an active form of PGE. Nocturnal investigation was accomplished on a randomized series of 18 subjects. For each recording, the total number of SCs could range between four and six. Aiming at assessing the behavior of generalized discharges, especially, when EEG synchrony processes operate more intensely, selection of SCs was restricted to the initial hours of sleep. However, a complete and regular profile of the first two SCs was found only in six recordings (see selective criteria). Diurnal EEG showed paroxysmal discharges consisting of generalized spike-and-waves, a regular background rhythm and no focal abnormality. All patients gave informed consent to the protocol.
2.1. Clinical data The patients were six subjects (two males and four females) with an age range between 10 and 25 years (17.49 5.9). Patients were affected by PGE with absence seizures (one subject), grand mal attacks associated with absence seizures (four subjects), and grand mal attacks (one subject). The diagnosis of epilepsy had been made from a mean of 5 years (range, 2 – 10 years). Medication regimen included either monotherapy (valproate, three subjects; carbamazepine, one subject) or polytherapy (valproate and phenitoin, one subject; valproate and carbamazepine, one subject). Serum levels of active compounds were within therapeutical ranges. No patient had taken either benzodiazepines or barbiturates in the past 6 months. Neurological and neuroradiological investigation was normal.
2.2. Sleep recordings Patients supplied an informed consent to undergo two consecutive polysomnographic recordings, the first one of which served as adaptation to the sleep lab and as control for screening other unrecognized sleep disorders. The second recording was used for analysis of polysomnographic measures.
9
On recording nights, subjects went to bed at their usual time, and they were asked to refrain from drinking beverages containing caffeine or alcohol during the previous afternoon and evening hours. PSG recordings were accomplished in a partially sound-proof recording chamber (sound pressure level below 30 dB Leq), under video-controlled supervision in a standard laboratory. Extensive bipolar montages (Fp1-F3, F3-C3, C3-P3, P3-O1; Fp2-F4, F4-C4, C4-P4, P4-O2) were integrated by monopolar derivation (C3/A2 or C4/A1) for the scoring of conventional sleep parameters. A calibration of 50 mV/cm was used for all EEG channels with a time constant of 0.1 s and a high-frequency filter in the 30 Hz range. Eye movements, heart rate, submental, deltoid and anterior tibialis muscles were recorded bipolarly. The paper speed was 15 mm/s. The total recording time for all PSG lasted 500 min.
2.3. Selecti6e criteria for SCs All recordings, in which the SCs showed sustained intervening wakefulness (\1 min), regressive shifts (abrupt variation from deep sleep to light sleep) during the descending branch or within the trough, or lack of internal modulation (composed only of stage 2 and REM sleep) were excluded from investigation. Only in six patients did the first two SCs meet the inclusive requirements. SCs started at the first epoch of NREM sleep and ended at the first epoch of stage 2 after a REM period had been completed. A REM period was considered completed, when the duration of NREM stages following the last epoch scored as stage REM exceeded 15 min. The NREM portion of the first SC (SC1) should contain three units, (a) a descending branch composed of stages 1, 2 and 3 in a rank order sequence and ending at the first epoch of stage 4; (b) a trough consisting of sustained slow wave sleep (SWS), basically characterized by prolonged periods of stage 4 and intervening epochs of stage 3; (c) an ascending branch identified between the last epoch of stage 4 and the onset of REM sleep.
10
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
The NREM portion of the second SC (SC2) should contain three units (descending branch, trough, ascending branch) with some SWS (at least 5 min of stage 4 or stage 3 or both) reached again with a rank order sequence from stage 1 or stage 2. Sleep staging (macrostructure) was based on standard guidelines (Rechtschaffen and Kales, 1968), while CAP parameters (microstructure) were coded according to the following rules. A CAP sequence is identified by repetitive clusters of stereotyped EEG features separated by time-equivalent intervals of background activ-
ity. Each CAP sequence includes at least two consecutive CAP cycles. The CAP cycle consists of a phase A (composed of transient EEG graphic elements) and a phase B (interval of theta/delta activity that separates two successive A phases). Each phase of CAP may last from 2 to 60 s. All CAP sequences begin with a phase A and end with a phase B (Terzano et al., 1985, 1988). CAP time is the temporal sum of CAP sequences. The percentage ratio of CAP time to sleep time is referred to as CAP rate. Three subtypes of A phases corresponding to different levels of neurophysiological activation can be distinguished (Fig. 1).
Fig. 1. Normal human sleep, succession of four A phases of CAP arising against the EEG background of stage 2 (B phases). Black arrows mark the boundaries of the CAP phases. The A1 subtypes are mainly composed of k-complexes, while phases A2 and A3 contain extensive amounts of EEG desynchronization. In particular, fast low-voltage rhythms are the dominant feature of subtype A3, whereas a balanced representation of EEG synchrony and desynchrony characterizes subtype A2. Note that the CAP phenomenon is not driven by any motor or respiratory disturbance. EOG, electro-oculogram; EMG, electromyogram (chin muscle); EKG, electrocardiogram; O-N PNG, oro-nasal airflow; THOR PNG, thoracic pneumogram; TIB ANT R, right anterior tibialis muscle; TIB ANT L, left anterior tibialis muscle; SDCP, Sleep Disorders Center of Parma.
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
Subtypes A1 (intermittent alpha rhythm in stage 1; sequences of k-complexes or delta bursts in the other NREM stages)-A phases consisting of synchronized EEG patterns (highvoltage low frequency rhythms). Subtypes A2 (k-complexes with alpha and beta activities, k-alpha, arousals with slow wave synchronization)-A phases composed of desynchronized EEG patterns (low-amplitude fast frequency rhythms) preceded by or mixed with slow high-voltage waves. Subtypes A3 (transient activation phases or arousals)-A phases with desynchronized EEG patterns alone or covering at least two-thirds of the phase A length. The EEG criteria for the identification of subtypes A3 and partially of subtypes A2 show extensive similarities with the ones proposed by the American Sleep Disorders Association (ASDA), (1992) for arousals. The ASDA arousals that occur in NREM sleep appear to be components of the CAP phenomenon (Boselli et al., 1998; Terzano and Parrino, 2000). Non-cyclic alternating pattern (NCAP) consists essentially of a rhythmic and stable EEG background, and it is scored whenever A phases are absent for \ 60 consecutive s. While NREM sleep can express periods of NCAP and periods of CAP, the latter is never detected, at least under physiological conditions, in REM sleep, where only isolated A3 subtypes (i.e. ASDA arousals) occur separated by intervals that do not meet the 2 – 60 s criteria for scoring cyclic EEG activities.
2.4. Sleep analysis The following variables were measured in each SC. Macrostructure Time and percentage of stage 1, 2, 3, 4, stages 3 and 4 SWS, NREM sleep, REM sleep. Duration of the basic SC units (descending branch, trough, ascending branch), and relative quantification of the sleep stages. Microstructure Total CAP Time. Total CAP Rate.
11
CAP Rate in each unit of the SC (descending branch, trough, ascending branch). Number of CAP cycles, of A phases (distinguished in subtypes A1, A2, A3), of B phases. Epileptic discharges Total number of generalized spikes and spikeand-waves. Discharge rate or spike index (SI, number of generalized spikes and spike-and-waves per minute) in each unit of the SC. Number of EEG discharges and SI during CAP, NCAP, phase A (including the subtype classification), phase B.
2.5. Statistical assessment As the six subjects supplied a complete series of the first two SCs meeting the inclusion/exclusion requirements, comparison of SI in NREM sleep, in CAP, in NCAP, in phase A and in phase B among the SC units (descending branch, trough, ascending branch) were analyzed by ANOVA for repeated measures, with pairwise contrasts tested by Bonferroni’s procedure (significance at PB 0.05). 3. Results
3.1. Sleep structure All the patients presented the basic pattern of EEG synchrony and EEG desynchrony, which characterizes the alternation of NREM and REM sleep within the SC (Fig. 2). The macrostructural values expressed in the first two SCs are illustrated in Table 1. Compared with SC1, SC2 had a longer duration mostly related to the increase of REM sleep. The stage composition of NREM sleep was in line with the well-known structure of sleep characterized by a declining concentration of SWS from SC1 (54 min) to SC2 (40.4 min) and by an increase of stage 2 from 16.6 min in SC1 to 38.6 min in SC2. At the microstructural level, the two SCs showed similar amounts of CAP time, CAP rate, and CAP cycles (Table 2). Subtypes A1 were more abundant in SC1, while subtypes A2 and A3 prevailed in SC2.
12
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
Fig. 2. Detailed histograms of the first two SCs expressed by the six patients with PGE. All SCs show a three-fold structure, a descending branch, sloping down progressively and uninterruptedly from wakefulness (or REM sleep) to the deeper NREM stages; a trough, composed of sustained stretches of SWS (stages 3 and/or 4); an ascending branch, characterized by a steep upward shift from SWS to the onset of REM sleep. The distribution of CAP sequences is indicated by the black lines placed over the continuous outline of the histogram profile. Wake, wakefulness; REM, rapid eye movement sleep; I, stage 1; II, stage 2; III, stage 3; IV, stage 4.
Table 3 reports the NREM stage composition of the SC units. The descending branch that characterizes the build-up of EEG synchrony was completed in 23.9 min in the course of SC1 and in 36.6 min during SC2. The trough that represents the climax of EEG synchrony lasted 47.4 min in SC1 and 35.4 min in SC2, while the reduction of SWS was an extremely accelerated process both in SC1 (3.8 min) and in SC2 (7.6 min). Table 4 depicts the microstructural variables in
the SC units. The values of CAP rate were higher in the descending branch of SC1 (59%) than of SC2 (38.8%). In contrast, SC2 presented greater amounts of CAP rate both in the troughs (+ 10.6%) and in the ascending branches (+ 43.2%). Within both SC1, subtypes A1 were prominent features of the descending branch and of the trough, while subtypes A2 and A3 were mostly found in the ascending branches of SC1 (n= 7) and of SC2 (n= 12).
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18 Table 1 Macrostructural parameters in the first two SCsa
13
Table 3 Macrostructural parameters in the main units (descending branch, trough, ascending branch) of the first two SCsa
SC1
SC2
Mean
86.8 (25.4)
101.6 (15.6)
94.2 (20.5)
SC1 Total Duration (min) S1 (min) (%) S2 (min) (%) S3 (min) (%) S4 (min) (%) SWS (min) (%) NREM (min) (%) REM (min) (%)
4.6 (3.4) 5.3 16.6 (5.6) 19.1 13.8 (9.2) 15.9 40.2 (20.1) 46.3 54 (17.1) 62.2 75.2 (20.2) 86.6 11.6 (5.3) 13.4
0.6 (1.3) 0.6 38.6 (14.6) 38 17.6 (6.4) 17.3 22.8 (23.6) 22.4 40.4 (21.9) 39.7 79.6 (12.2) 78.3 22 (9.6) 21.7
2.6 2.8 27.6 29.3 15.7 16.7 31.5 33.4 47.2 50.1 77.4 82.2 16.8 17.8
(2.8) (9.5) (7.7) (22.3) (19.6) (16.1) (7.3)
a SC1, first sleep cycle; SC2, second sleep cycle; S1, S2, S3, S4, stages 1, 2, 3, 4; NREM, non-REM sleep; REM, REM sleep; Percentages are all referred to the total duration of the sleep cycle; S.Ds in parentheses.
3.2. Acti6ation of EEG paroxysms A total number of 2724 EEG discharges was quantified within the first two SCs. Table 5 shows that SI indexes were higher in SC1 than in SC2. In both SCs, activation of EEG paroxysms was more powerful in NREM than in REM sleep. Throughout NREM sleep (mean SI, 2.8), the discharge rate was lower in NCAP (mean SI, 1.9)
Table 2 CAP parameters in the first two SCsa
CAP time (min) CAP rate (%) CAP cycles (n) A1 (n) A2 (n) A3 (n)
SC1
SC2
Mean
44.8 (19) 59.8 (16.6) 87 (32) 76 (26) 9 (10) 2 (1)
46.2 (17) 58 (17.6) 86 (38) 67 (35) 15 (9) 4 (4)
45.5 (18) 58.9 (17) 86.5 (34) 71.5 (29) 12 (9) 3 (2)
a CAP, cyclic alternating pattern; A1, phase A1 subtype; A2, phase A2 subtype; A3, phase A3 subtype; S.Ds in parentheses.
SC2
Descending branch (min) S1 (min) S2 (min) S3 (min) S4 (min)
23.9 4.6 12.7 6.6 0
(11.3) (3.4) (5) (5.4)
36.6 (18.4) 0.6 (1.3) 31 (16.3) 5 (6.3) 0
Trough (min) S1 (min) S2 (min) S3 (min) S4 (min)
47.4 (6.9) 0 0 7.2 (6.3) 40.2 (20.1)
35.4 (19) 0 0 12.6 (7.2) 22.8 (23.6)
3.9 (0.8) 0 3.9 (1.1) 0 0
7.6 (6.4) 0 7.6 (6.4) 0 0
Ascending branch (min) S1 (min) S2 (min) S3 (min) S4 (min)
a SC1, first sleep cycle; SC2, second sleep cycle; S1, S2, S3, S4, stages 1, 2, 3, 4; NREM, non-REM sleep; REM, REM sleep; standard deviations in parentheses.
and higher in CAP (mean SI, 3.4). The CAP-related activation of epileptic events was linked to the single A phases (Fig. 3), while the B phases exerted a shallow action, even compared with the NCAP condition. EEG paroxysms were 1302 in the descending branches, 1188 in the troughs and 72 in the ascending branches. The distribution of interictal bursts in the two SCs is reported in Table 6. Both SCs showed a progressive decrease of activation across the three units, from the highest SI values reached during the descending branches (mean SI, 3.6) to the more attenuated SI values during the troughs (mean SI, 2.4) down to the lowest values during the ascending limbs (mean SI, 1.1). The magnitude of activation during the descending branches was closely related to the CAP condition (mean SI, 5.2) and to the powerful effect of phase A (mean SI, 13.9) (Fig. 4). In each unit, the SI quantified during the A phases of CAP was three-fold higher compared with the values expressed throughout the entire unit. About 82% of all the EEG discharges occurring in phase A were distributed within a subtype A1.
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
14
Table 4 CAP parameters in the main units (descending branch, trough, ascending branch) of the first two SCsa SC1
SC2
Descending branch (min) CAP (min) CAP rate (%) A1 (n) A2 (n) A3 (n)
23.9 (11.3) 14.1 (10.8) 59 21 (13) 3 (2) 0
36.6 (18.4) 14.2 (14.3) 38.8 19 (9) 4 (2) 0
Trough (min) CAP (min) CAP rate (%) A1 (n) A2 (n) A3 (n)
47.4 (6.9) 29 (11.5) 61.1 54 (21) 1 (1) 1 (1)
35.4 (19) 25.4 (11) 71.7 48 (33) 3 (2) 0
Ascending branch (min) CAP (min) CAP rate (%) A1 (n) A2 (n) A3 (n)
3.9 (0.8) 1.7 (0.9) 43.6 1 (1) 5 (2) 2 (1)
7.6 (6.4) 6.6 (3.1) 86.8 0 8 (3) 4 (2)
a SC1, first sleep cycle; SC2, second sleep cycle; CAP, cyclic alternating pattern; CAP rate, ratio of CAP time to NREM sleep time throughout the SC; A1, A2, A3, phase A subtypes of CAP; S.D.s in parentheses.
4. Discussion There is consolidated evidence that states of extreme waking or deep sleep are much less conducive to interictal bursting compared with intensive fluctuations of the vigilance level (Halasz, Table 5 Spike indexes referred to the first two SCsa SC1 TST NREM REM CAP NCAP Phase A Phase B
3.1 3.4 0.9 4.3 2.1 10.7 0.5
(1.1) (1.6) (0.6) (2.2) (1) (6.2) (0.2)
SC2
Mean
1.9 2.1 0.8 2.4 1.7 6.6 0.1
2.5 (0.8) 2.8 (1.4) 0.8 (0.5) 3.4 (1.4) 1.9 (0.9) 8.7(4.9) 0.3 (0.2)
(0.7) (1) (0.5) (1.2) (0.8) (3.8) (0.1)
a SC1, first sleep cycle; SC2, second sleep cycle; TST, total SC time; NREM, NREM sleep time; REM, REM sleep time; CAP, cyclic alternating pattern; NCAP, non-CAP; S.D.s in parentheses.
1991). CAP is a major marker of arousal instability that accompanies the transitional phases (Terzano et al., 1988) and, therefore, acts as a convulsive-promoting agent (Parrino et al., 2000). In the current study, activation of generalized EEG abnormalities was higher during CAP compared with the stable NCAP sleep condition. Confirming the data of previous reports (Terzano et al., 1989, 1991a,b, 1992, 1997; Gigli et al., 1992; Oldani et al., 1998), the A phases of CAP operated as permissive windows of epileptic events in both SCs and in each SC unit. Overall, activation was about 30-fold stronger in phase A compared with phase B. The great majority of interictal discharges occurred during subtypes A1, suggesting a crucial association between generalized bursts and the slow arousal components of CAP (Terzano and Parrino, 2000). It is known that the synchronizing activity of sleep decreases homeostatically across the successive SCs (Borbely, 1982). Like EEG synchrony, activation of EEG paroxysms diminishes as sleep progresses. This effect is mostly evident over the first two SCs (Kellaway et al. 1980). In accordance with these findings, the discharge rates in our patients with PGE decreased from SC1 to SC2, with higher values quantified during NREM sleep compared with REM sleep. The stronger influence of NREM sleep upon interictal discharges is attributed to the greater synchronization of its EEG activities. In the light of this, the maximum of epileptic activation within the SC should coincide with the climax of EEG synchrony. The present findings indicate, however, that the bursting activity in PGE is higher before the onset of stage 4 that represents the maximum expression of EEG synchrony within the SC. The SC is organized in regular patterns based upon the alternation of NREM and REM sleep, which is repeated several times during the night. This two-fold nature of sleep within the SC framework is regulated by the reciprocal activation of antagonistic cell groups, the REM-off neurons that operate during the genesis and consolidation of deep NREM sleep, and the REM-on neurons that fire immediately before and during REM sleep (McCarley and Hobson, 1975). There is experimental evidence
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
15
Fig. 3. Bottom: enhancement of one of the histograms reported in Fig. 1 allowing a clearer representation of CAP sequences (here depicted by black dots) across the first two SCs. Middle: number of A phases (Ph. A) per 5 min of sleep. The vertical black dots coincide to subtypes A1, the vertical black lines correspond to subtypes A2 and A3. Top: number of spike-and-waves (P.O.) per 5 min of sleep. Compared with the bursting activity expressed during stable sleep (without CAP sequences), generalized discharges occur more frequently, when they arise in association with a phase A of CAP.
that paroxysmal activities in corticothalamic networks develop progressively through increasing levels of neuronal synchronization during sleep EEG patterns (Steriade and Contreras, 1995). Accordingly, activation of generalized discharges was three-fold higher in the descending branch (mean SI, 3.6) and two-fold higher in the trough (mean SI, 2.4) compared with the ascending branch (mean SI, 1.1), which showed discharge rates similar to the ones measured in REM sleep (mean SI, 0.9). A further promoting factor of bursting was related to the time required to achieve sustained SWS. Comparing the two descending branches, activation of interictal bursts was greater in SC1,
that more rapidly reached full EEG synchrony due to the stronger initial sleep pressure (Borbely, 1982). In other words, the steeper the descending slope the more potent the activation. This can explain why sleep recovery after sleep deprivation, that actually accelerates the accomplishment of deep NREM sleep, is effective in provoking generalized discharges (Shouse et al., 1996). The stronger activating influence of rising EEG synchrony is not limited to PGE. Investigation carried out on children with partial epilepsy (either symptomatic or cryptognetic) confirmed the strong activating properties of NREM sleep on interictal bursts, with higher discharge rates measured in stages 2 and 3 of the descending branches
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
16
Table 6 Spike indexes in the different units of the first twoSCs Descending branch (DB)
Trough (T)
Ascending branch (AB)
P-value
Statistical comparisons
(a) NREM component of SC1 NREM 5.2 (2) CAP 6.6 (1.7) NCAP 3.2 (1.5) Phase A 17.4 (10) Phase B 0.5 (0.2)
2.7 3.4 1.6 8.2 0.4
(1.8) (1.9) (0.6) (4.7) (0.2)
1.3 1.4 1.3 3.7 0.2
0.003 0.0001 0.01 0.005 0.02
DB\AB DB\T= AB DB\T= AB DB\AB DB\AB
(b) NREM component of SC2 NREM 2.5 (1.7) CAP 3.7 (1.3) NCAP 1.7 (0.8) Phase A 10.4 (5.6) Phase B 0.2 (0.1)
2.1 2.3 1.3 5.9 0.1
(1.3) (1.4) (0.6) (4.1) (0.1)
0.9 (0.7) 0.9 (0.5) 1 (0.4) 3.1 (2) 0
(Nobili et al., 1999). In adults, affected by partial epilepsy, interictal EEG discharges were more frequent just before sleep reached the climax of EEG synchrony within each SC (Malow et al., 1998). In that study, the smoothing of the delta power curves obscured the short-lasting fluctuations of sleep depth that are known to precede the peak of EEG synchrony within the SC (Feinberg, 1989). The achievement of EEG synchrony is not a stepwise process, but it is shaped progressively through the recurring CAP oscillations. Spectral analysis has shown that, in the evolution from light to deep NREM sleep, the total power of the oscillation increases, while the amplitude gaps decline between the maximum (phase A) and minimum (phase B) values of the oscillation (Ferrillo et al., 1997). In other words, in the generation of EEG synchrony CAP cycles undergo considerable power variations. The major activating role played by CAP in the descending branches indicates that it is not massive EEG synchrony, but broad amplitude shifting of delta power that triggers interictal responses in PGE. Once the climax of EEG synchrony is reached, the swings of arousal level take place within a narrower range and activation is attenuated. In effect, the troughs of our PGE patients presented elevated values of CAP rate (for comparison with normative data see Terzano et al., 2000), but still their bursting activity was lower compared with the descending branches.
(1) (1.4) (0.7) (2.1) (0.1)
NS 0.003 NS 0.02 0.003
DB\AB DB\AB DB\AB
While the generation of EEG synchrony promoted interictal firing, the ascending branches, under the neurophysiological domain of REM sleep, operated as inhibitory attractors, as reflected by the convergent SI values measured during CAP and NCAP (Fig. 4). Attenuation of bursting activity across the SCs affected the pro-
Fig. 4. Number of EEG bursts per minute of sleep (SI) during NREM sleep (NREM), CAP, non-cyclic alternating pattern (NCAP), phase A and phase B. In spite of the overlapping values expressed by CAP and NCAP in the ascending branch, the activating effect of phase A and the inhibitory action of phase B still persist in this SC unit.
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
moting impact of A phases that, however, expressed in each unit discharge rates well over the ones related to any other sleep variable. The powerful inhibitory influence of phase B mirrored the declining trend of phase A. In conclusion, the present study confirms the hypothesis that generalized epileptic events share common neurophysiological pathways with the mechanisms involved in EEG synchrony. As CAP participates in the production and attenuation of EEG synchrony, it is not surprising that the modulatory effect of CAP on generalized interictal discharges varies in relation to the dynamics of the SC. Within the first two SCs, the features of NREM sleep endowed with the major activating power on generalized bursts are represented by the generation of EEG synchrony and by the A phases of CAP that regulate its build-up.
References American Sleep Disorders Association (ASDA) 1992. EEG arousals: scoring rules and examples. A preliminary report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association. Sleep 15, 174–84. Borbely, A.A., 1982. A two process model of sleep regulation. Hum. Neurobiol. 1, 195–204. Boselli, M., Parrino, L., Smerieri, A., Terzano, M.G., 1998. Effect of age on EEG arousals in normal sleep. Sleep 21, 361– 367. Feinberg, I., 1989. Effects of maturation and aging on slowwave sleep in man: implications for neurobiology. In: Wauquier, A., Dugovic, C., Radulovacki, M. (Eds.), Slow Wave Sleep: Physiological, Pathophysiological and Functional Aspects. Raven Press, New York, pp. 31–48. Ferrillo, F., Gabarra, M., Nobili, L., Parrino, L., Schiavi, G., Stubinski, B., Terzano, M.G., 1997. Comparison between visual scoring of cyclic alternating pattern (CAP) and computerized assessment of slow EEG oscillations in the transition from light to deep non-REM sleep. J. Clin. Neurophysiol. 14, 210–216. Gigli, G.L., Calia, E., Marciani, M.G., Mazza, S., Mennuni, G., Diomedi, M., Terzano, M.G., Janz, D., 1992. Sleep microstructure and EEG epileptiform activity in patients with juvenile myoclonic epilepsy. Epilepsia 33, 799–804. Halasz, P., 1991. Sleep, arousal and electroclinical manifestations of generalized epilepsy with spike wave pattern. In: Degen, R., Rodin, E.A. (Eds.), Epilepsy, Sleep and Sleep Deprivation. Elsevier, Amsterdam, pp. 43–48. Janz, D., 1962. The grand-mal epilepsies and the sleeping-waking cycle. Epilepsia 3, 69–109.
17
Kellaway, P., Frost, J.D., Crawley, J.W., 1980. Time modulation of spike-and-wave activity in generalized epilepsy. Ann. Neurol. 8, 491– 500. Malow, B.A., Lin, X., Kushwaha, R., Aldrich, M.S., 1998. Interictal spiking increases with sleep depth in temporal lobe epilepsy. Epilepsia 39, 1309– 1316. McCarley, R.W., Hobson, J.A., 1975. Neuronal excitability modulation over the sleep cycle: a structural and mathematical model. Science 189, 58 – 60. Merica, H., Gaillard, J.-M., 1986. Internal structure of sleep cycles in a healthy population. Sleep 9, 502– 513. Nobili, L., Baglietto, M.G., Beelke, M., De Carli, F., De Negri, E., Rosadini, G., De Negri, M., Ferrillo, F., 1999. Modulation of sleep interictal epileptiform discharges in partial epilepsy of childhood. Clin. Neurophysiol. 110, 839– 845. Oldani, A., Zucconi, M., Asselta, R., Modugno, M., Bonati, M.T., Dalpra`, L., Malcovati, M., Tenchini, M.L., Smirne, S., Ferini-Strambi, L., 1998. Autosomal dominant nocturnal frontal lobe epilepsy. A video-polysomnographic and genetic appraisal of 40 patients and delineation of the epileptic syndrome. Brain 121, 205– 223. Parrino L, Smerieri A, Spaggiari MC, Terzano M.G., 2000. Cyclic alternating pattern (CAP) and epilepsy during sleep: how a physiological rhythm modulates a pathological event. Clin Neurophysiol, in press. Parrino, L., Boselli, M., Spaggiari, M.C., Smerieri, A., Terzano, M.G., 1998. Cyclic alternating pattern (CAP) in normal sleep: polysomnographic parameters in different age groups. Electroencephalogr. Clin. Neurophysiol. 107, 439– 450. Rechtschaffen A, Kales A., 1968. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Brain Information Service/Brain Research Institute, University of California at Los Angeles, Los Angeles. Shouse, M.N., Martins da Silva, A., Sammaritano, M., 1996. Circadian rhythms, sleep, and epilepsy. J. Clin. Neurophysiol. 13, 32 – 50. Steriade, M., Contreras, D., 1995. Relations between cortical and thalamic cellular events during transition from sleep patterns to paroxysmal activity. J. Neurosci. 15, 623– 642. Steriade, M., Gloor, P., Llinas, R.R., Lopes da Silav, F.H., Mesulam, M.-M., 1990. Basic mechanisms of cerebral activities. Electroencephalogr. Clin. Neurophysiol. 76, 481– 508. Terzano, M.G., Parrino, L., 2000. Origin and significance of cyclic alternating pattern (CAP). Sleep Med. Rev. 4, 101– 123. Terzano, M.G., Mancia, D., Salati, M.R., Costani, G., Decembrino, A., Parrino, L., 1985. The cyclic alternating pattern as a physiologic component of normal NREM sleep. Sleep 8, 137– 145. Terzano, M.G., Parrino, L., Spaggiari, M.C., 1988. The cyclic alternating pattern sequences in the dynamic organization of sleep. Electroencephalogr. Clin. Neurophysiol. 69, 437– 447.
18
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
Terzano, M.G., Parrino, L., Anelli, S., Halasz, P., 1989. Modulation of generalized spike-and-wave discharges during sleep by cyclic alternating pattern. Epilepsia 30, 772– 781. Terzano, M.G., Parrino, L., Garofalo, P.G., Durisotti, C., Filati-Roso, C., 1991a. Activation of motor seizures during cyclic alternating pattern in sleep. Epilepsy Res. 10, 166– 173. Terzano, M.G., Parrino, L., Spaggiari, M.C., Barusi, R., Simeoni, S., 1991b. Discriminatory effect of cyclic alternating pattern in focal lesional and benign rolandic interictal spikes during sleep. Epilepsia 32, 616–628.
Terzano, M.G., Parrino, L., Anelli, S., Boselli, M., Clemens, B., 1992. Effects of generalized interictal EEG discharges on sleep stability: assessment by means of cyclic alternating pattern. Epilepsia 33, 317– 326. Terzano, M.G., Monge-Strauss, M.F., Mikol, F., Spaggiari, M.C., Parrino, L., 1997. Cyclic alternating pattern as a provocative factor in nocturnal paroxysmal dystonia. Epilepsia 38, 1015– 1025. Terzano, M.G., Parrino, L., Boselli, M., Smerieri, A., Spaggiari, M.C., 2000. CAP components and EEG synchronization in the first three sleep cycles. Clin. Neurophysiol. 111, 283– 290.
.
Combined influence of cyclic arousability and EEG synchrony on generalized interictal discharges within the sleep cycle Liborio Parrino, Arianna Smerieri, Mario Giovanni Terzano * Istituto di Neurologia, Uni6ersita` de Parma, Via del Quartiere 4, 43100 Parma, Italy Received 19 April 2000; received in revised form 20 September 2000; accepted 26 September 2000
Abstract Purpose: to analyze the activating role of cyclic alternating pattern (CAP) and EEG synchrony on generalized interictal paroxysms in the first part of the night, when all sleep patterns are represented. Methods: nocturnal polysomnographic investigation was accomplished on a randomized series of 18 subjects with an active form of primary generalized epilepsy (PGE), but only six patients showed a complete and regular profile of the first two sleep cycles (SCs). Completeness and regularity of the selected SCs consisted in the absence of intervening wakefulness, in the presence of all sleep stages, and in the identification of three main units, (a) a descending branch, dominated by the build-up of EEG synchrony in the transition from light to deep non-rapid eye movement (NREM) sleep; (b) a trough, where the magnitude of EEG synchrony is greatest and gives rise to stages 3 and 4; (c) an ascending branch characterized by a decrease of EEG synchrony preceding the onset of rapid eye movement (REM) sleep. Generalized paroxysms were evaluated in terms of discharge rates (number of interictal bursts per minute of sleep) and distribution within the investigated sleep parameters. Results: the discharge rates decreased from SC1 to SC2, with higher values quantified during NREM sleep (mean, 2.8) compared with REM sleep (mean, 0.8). Both SCs showed a progressive decrease of activation across the three units, from the highest discharge rates reached during the descending branches (mean, 3.6) to the more attenuated discharge rates during the troughs (mean, 2.4) down to the lowest rates during the ascending limbs (mean, 1.1). The magnitude of activation during the descending branches was closely related to the CAP condition (mean, 5.2) and to the powerful effect of phase A (mean, 13.9). The great majority (82%) of EEG discharges occurring in phase A were distributed within the A1 subtypes (identified by sequences of k-complexes or delta bursts). Conclusions: within the first two SCs, the features of NREM sleep endowed with the major activating power on generalized bursts are represented by the rise of EEG synchrony (descending branch) and by the A phases of CAP involved in the regulation of its build-up. © 2001 Elsevier Science B.V. All rights reserved.
Abbre6iations: SC, sleep cycle; SC units, descending branch, trough, ascending branch; SC1, first sleep cycle; SC2, second sleep cycle; PGE, primary generalized epilepsy; SI, spike index; CAP, cyclic alternating pattern (arousal fluctuation); NCAP, non-cyclic alternating pattern (stable sleep); REM sleep, rapid eye movement sleep; NREMsleep, non-REM sleep; SWS, slow wave sleep (stages 3 and 4); ASDA, american sleep disorders association. * Corresponding author. Tel.: + 39-0521-287913; fax: + 39-0521-287913. E-mail address: [email protected] (M.G. Terzano). 0920-1211/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 9 2 0 - 1 2 1 1 ( 0 0 ) 0 0 1 9 2 - 3
8
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
Keywords: Cyclic alternating pattern; EEG; Generalized epilepsy; Interictal discharges; Sleep cycle
1. Introduction The physiological variations of arousal across the 24 h cycle represent an important factor in the modulation of epileptic phenomena. There is extensive agreement that epileptic manifestations more likely appear, when the level of arousal is reduced. In particular, generalized interictal discharges mostly occur in the resting non-rapid eye movement (NREM) sleep, when EEG rhythms present high amplitudes and slow frequencies (Steriade et al., 1990), and the number of generalized bursts is lowest during the more wake-like rapid eye movement (REM) sleep (Shouse et al., 1996). It has been postulated that generalized EEG discharges share common anatomical pathways with neural mechanisms that control EEG synchrony and arousal modulation during sleep (Terzano et al., 1992). The coexistence of physiological processes and pathological factors makes epilepsy a cerebral disorder in which the interictal manifestations may appear discontinuously, but not randomly. Within NREM sleep, the transitional phases, such as sleep onset, awakenings from sleep, stage transitions, activate interictal discharges more powerfully than the fully developed stages (Janz, 1962). These critical periods of poorly consolidated vigilance are characterized by EEG fluctuations and cyclic arousability. Quantification of these periodic changes is based on the scoring rules for coding the cyclic alternating pattern (CAP) (Terzano et al., 1985). CAP is an EEG phenomenon organized in sequences that occupy wide stretches of NREM sleep (Parrino et al., 1998). CAP sequences are composed of repetitive transient patterns lasting 8 – 15 s (phase A) separated by intervals of 15 – 20 s (phase B). The A phases are constituted by the stage-related arousal correlates (k-complexes, delta bursts, arousals), while the B phases are composed of the stage-related background activities. During
CAP, the EEG rhythms of sleep oscillate with periodic excitatory (phase A) and inhibitory (phase B) swings. It is during CAP that a number of epileptic events occurring in NREM sleep are seen with the highest frequency. In primary generalized epilepsy (PGE), the majority of interictal EEG abnormalities are associated with a CAP sequence and are triggered by an A phase (Terzano et al., 1989). The increase of generalized spike-wave paroxysms is apparent during CAP in all NREM stages, but no evaluation has been carried out on their position within the sleep cycle (SC). Conventionally, a SC is defined by the alternation of NREM and REM sleep. The NREM portion of the SC is composed of a succession of stages, which determine three distinct units characterized by different degrees of EEG synchrony. This succession involves a descending branch sloping from the more superficial to the deeper NREM stages, in which EEG synchrony undergoes a progressive increase ; a central trough of deep sleep, in which EEG synchrony is highest; an ascending branch where lighter NREM stages precede the onset of REM sleep, in which EEG synchrony decreases (Merica and Gaillard, 1986). CAP sequences interact with EEG synchrony showing different phase A patterns in the three units of NREM sleep (Ferrillo et al., 1997; Terzano et al., 2000). The present study aimed at assessing the combined influence of sleep patterns and arousal instability on PGE interictal discharges during the first part of the night, when the variations of EEG synchrony are more pronounced. For this purpose, we focused upon the initial hours of continuous sleep analyzing the amount of CAP sequences and generalized discharges within ‘ideal’ SCs, i.e. the ones uninterrupted by intervening wakefulness and, in which all stages are represented and linked in a regular succession of a descending branch, a trough and an ascending branch.
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
2. Material and methods The study was carried out on attended laboratory-recorded polysomnograms obtained at the Parma Sleep Disorders Center in subjects with an active form of PGE. Nocturnal investigation was accomplished on a randomized series of 18 subjects. For each recording, the total number of SCs could range between four and six. Aiming at assessing the behavior of generalized discharges, especially, when EEG synchrony processes operate more intensely, selection of SCs was restricted to the initial hours of sleep. However, a complete and regular profile of the first two SCs was found only in six recordings (see selective criteria). Diurnal EEG showed paroxysmal discharges consisting of generalized spike-and-waves, a regular background rhythm and no focal abnormality. All patients gave informed consent to the protocol.
2.1. Clinical data The patients were six subjects (two males and four females) with an age range between 10 and 25 years (17.49 5.9). Patients were affected by PGE with absence seizures (one subject), grand mal attacks associated with absence seizures (four subjects), and grand mal attacks (one subject). The diagnosis of epilepsy had been made from a mean of 5 years (range, 2 – 10 years). Medication regimen included either monotherapy (valproate, three subjects; carbamazepine, one subject) or polytherapy (valproate and phenitoin, one subject; valproate and carbamazepine, one subject). Serum levels of active compounds were within therapeutical ranges. No patient had taken either benzodiazepines or barbiturates in the past 6 months. Neurological and neuroradiological investigation was normal.
2.2. Sleep recordings Patients supplied an informed consent to undergo two consecutive polysomnographic recordings, the first one of which served as adaptation to the sleep lab and as control for screening other unrecognized sleep disorders. The second recording was used for analysis of polysomnographic measures.
9
On recording nights, subjects went to bed at their usual time, and they were asked to refrain from drinking beverages containing caffeine or alcohol during the previous afternoon and evening hours. PSG recordings were accomplished in a partially sound-proof recording chamber (sound pressure level below 30 dB Leq), under video-controlled supervision in a standard laboratory. Extensive bipolar montages (Fp1-F3, F3-C3, C3-P3, P3-O1; Fp2-F4, F4-C4, C4-P4, P4-O2) were integrated by monopolar derivation (C3/A2 or C4/A1) for the scoring of conventional sleep parameters. A calibration of 50 mV/cm was used for all EEG channels with a time constant of 0.1 s and a high-frequency filter in the 30 Hz range. Eye movements, heart rate, submental, deltoid and anterior tibialis muscles were recorded bipolarly. The paper speed was 15 mm/s. The total recording time for all PSG lasted 500 min.
2.3. Selecti6e criteria for SCs All recordings, in which the SCs showed sustained intervening wakefulness (\1 min), regressive shifts (abrupt variation from deep sleep to light sleep) during the descending branch or within the trough, or lack of internal modulation (composed only of stage 2 and REM sleep) were excluded from investigation. Only in six patients did the first two SCs meet the inclusive requirements. SCs started at the first epoch of NREM sleep and ended at the first epoch of stage 2 after a REM period had been completed. A REM period was considered completed, when the duration of NREM stages following the last epoch scored as stage REM exceeded 15 min. The NREM portion of the first SC (SC1) should contain three units, (a) a descending branch composed of stages 1, 2 and 3 in a rank order sequence and ending at the first epoch of stage 4; (b) a trough consisting of sustained slow wave sleep (SWS), basically characterized by prolonged periods of stage 4 and intervening epochs of stage 3; (c) an ascending branch identified between the last epoch of stage 4 and the onset of REM sleep.
10
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
The NREM portion of the second SC (SC2) should contain three units (descending branch, trough, ascending branch) with some SWS (at least 5 min of stage 4 or stage 3 or both) reached again with a rank order sequence from stage 1 or stage 2. Sleep staging (macrostructure) was based on standard guidelines (Rechtschaffen and Kales, 1968), while CAP parameters (microstructure) were coded according to the following rules. A CAP sequence is identified by repetitive clusters of stereotyped EEG features separated by time-equivalent intervals of background activ-
ity. Each CAP sequence includes at least two consecutive CAP cycles. The CAP cycle consists of a phase A (composed of transient EEG graphic elements) and a phase B (interval of theta/delta activity that separates two successive A phases). Each phase of CAP may last from 2 to 60 s. All CAP sequences begin with a phase A and end with a phase B (Terzano et al., 1985, 1988). CAP time is the temporal sum of CAP sequences. The percentage ratio of CAP time to sleep time is referred to as CAP rate. Three subtypes of A phases corresponding to different levels of neurophysiological activation can be distinguished (Fig. 1).
Fig. 1. Normal human sleep, succession of four A phases of CAP arising against the EEG background of stage 2 (B phases). Black arrows mark the boundaries of the CAP phases. The A1 subtypes are mainly composed of k-complexes, while phases A2 and A3 contain extensive amounts of EEG desynchronization. In particular, fast low-voltage rhythms are the dominant feature of subtype A3, whereas a balanced representation of EEG synchrony and desynchrony characterizes subtype A2. Note that the CAP phenomenon is not driven by any motor or respiratory disturbance. EOG, electro-oculogram; EMG, electromyogram (chin muscle); EKG, electrocardiogram; O-N PNG, oro-nasal airflow; THOR PNG, thoracic pneumogram; TIB ANT R, right anterior tibialis muscle; TIB ANT L, left anterior tibialis muscle; SDCP, Sleep Disorders Center of Parma.
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
Subtypes A1 (intermittent alpha rhythm in stage 1; sequences of k-complexes or delta bursts in the other NREM stages)-A phases consisting of synchronized EEG patterns (highvoltage low frequency rhythms). Subtypes A2 (k-complexes with alpha and beta activities, k-alpha, arousals with slow wave synchronization)-A phases composed of desynchronized EEG patterns (low-amplitude fast frequency rhythms) preceded by or mixed with slow high-voltage waves. Subtypes A3 (transient activation phases or arousals)-A phases with desynchronized EEG patterns alone or covering at least two-thirds of the phase A length. The EEG criteria for the identification of subtypes A3 and partially of subtypes A2 show extensive similarities with the ones proposed by the American Sleep Disorders Association (ASDA), (1992) for arousals. The ASDA arousals that occur in NREM sleep appear to be components of the CAP phenomenon (Boselli et al., 1998; Terzano and Parrino, 2000). Non-cyclic alternating pattern (NCAP) consists essentially of a rhythmic and stable EEG background, and it is scored whenever A phases are absent for \ 60 consecutive s. While NREM sleep can express periods of NCAP and periods of CAP, the latter is never detected, at least under physiological conditions, in REM sleep, where only isolated A3 subtypes (i.e. ASDA arousals) occur separated by intervals that do not meet the 2 – 60 s criteria for scoring cyclic EEG activities.
2.4. Sleep analysis The following variables were measured in each SC. Macrostructure Time and percentage of stage 1, 2, 3, 4, stages 3 and 4 SWS, NREM sleep, REM sleep. Duration of the basic SC units (descending branch, trough, ascending branch), and relative quantification of the sleep stages. Microstructure Total CAP Time. Total CAP Rate.
11
CAP Rate in each unit of the SC (descending branch, trough, ascending branch). Number of CAP cycles, of A phases (distinguished in subtypes A1, A2, A3), of B phases. Epileptic discharges Total number of generalized spikes and spikeand-waves. Discharge rate or spike index (SI, number of generalized spikes and spike-and-waves per minute) in each unit of the SC. Number of EEG discharges and SI during CAP, NCAP, phase A (including the subtype classification), phase B.
2.5. Statistical assessment As the six subjects supplied a complete series of the first two SCs meeting the inclusion/exclusion requirements, comparison of SI in NREM sleep, in CAP, in NCAP, in phase A and in phase B among the SC units (descending branch, trough, ascending branch) were analyzed by ANOVA for repeated measures, with pairwise contrasts tested by Bonferroni’s procedure (significance at PB 0.05). 3. Results
3.1. Sleep structure All the patients presented the basic pattern of EEG synchrony and EEG desynchrony, which characterizes the alternation of NREM and REM sleep within the SC (Fig. 2). The macrostructural values expressed in the first two SCs are illustrated in Table 1. Compared with SC1, SC2 had a longer duration mostly related to the increase of REM sleep. The stage composition of NREM sleep was in line with the well-known structure of sleep characterized by a declining concentration of SWS from SC1 (54 min) to SC2 (40.4 min) and by an increase of stage 2 from 16.6 min in SC1 to 38.6 min in SC2. At the microstructural level, the two SCs showed similar amounts of CAP time, CAP rate, and CAP cycles (Table 2). Subtypes A1 were more abundant in SC1, while subtypes A2 and A3 prevailed in SC2.
12
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
Fig. 2. Detailed histograms of the first two SCs expressed by the six patients with PGE. All SCs show a three-fold structure, a descending branch, sloping down progressively and uninterruptedly from wakefulness (or REM sleep) to the deeper NREM stages; a trough, composed of sustained stretches of SWS (stages 3 and/or 4); an ascending branch, characterized by a steep upward shift from SWS to the onset of REM sleep. The distribution of CAP sequences is indicated by the black lines placed over the continuous outline of the histogram profile. Wake, wakefulness; REM, rapid eye movement sleep; I, stage 1; II, stage 2; III, stage 3; IV, stage 4.
Table 3 reports the NREM stage composition of the SC units. The descending branch that characterizes the build-up of EEG synchrony was completed in 23.9 min in the course of SC1 and in 36.6 min during SC2. The trough that represents the climax of EEG synchrony lasted 47.4 min in SC1 and 35.4 min in SC2, while the reduction of SWS was an extremely accelerated process both in SC1 (3.8 min) and in SC2 (7.6 min). Table 4 depicts the microstructural variables in
the SC units. The values of CAP rate were higher in the descending branch of SC1 (59%) than of SC2 (38.8%). In contrast, SC2 presented greater amounts of CAP rate both in the troughs (+ 10.6%) and in the ascending branches (+ 43.2%). Within both SC1, subtypes A1 were prominent features of the descending branch and of the trough, while subtypes A2 and A3 were mostly found in the ascending branches of SC1 (n= 7) and of SC2 (n= 12).
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18 Table 1 Macrostructural parameters in the first two SCsa
13
Table 3 Macrostructural parameters in the main units (descending branch, trough, ascending branch) of the first two SCsa
SC1
SC2
Mean
86.8 (25.4)
101.6 (15.6)
94.2 (20.5)
SC1 Total Duration (min) S1 (min) (%) S2 (min) (%) S3 (min) (%) S4 (min) (%) SWS (min) (%) NREM (min) (%) REM (min) (%)
4.6 (3.4) 5.3 16.6 (5.6) 19.1 13.8 (9.2) 15.9 40.2 (20.1) 46.3 54 (17.1) 62.2 75.2 (20.2) 86.6 11.6 (5.3) 13.4
0.6 (1.3) 0.6 38.6 (14.6) 38 17.6 (6.4) 17.3 22.8 (23.6) 22.4 40.4 (21.9) 39.7 79.6 (12.2) 78.3 22 (9.6) 21.7
2.6 2.8 27.6 29.3 15.7 16.7 31.5 33.4 47.2 50.1 77.4 82.2 16.8 17.8
(2.8) (9.5) (7.7) (22.3) (19.6) (16.1) (7.3)
a SC1, first sleep cycle; SC2, second sleep cycle; S1, S2, S3, S4, stages 1, 2, 3, 4; NREM, non-REM sleep; REM, REM sleep; Percentages are all referred to the total duration of the sleep cycle; S.Ds in parentheses.
3.2. Acti6ation of EEG paroxysms A total number of 2724 EEG discharges was quantified within the first two SCs. Table 5 shows that SI indexes were higher in SC1 than in SC2. In both SCs, activation of EEG paroxysms was more powerful in NREM than in REM sleep. Throughout NREM sleep (mean SI, 2.8), the discharge rate was lower in NCAP (mean SI, 1.9)
Table 2 CAP parameters in the first two SCsa
CAP time (min) CAP rate (%) CAP cycles (n) A1 (n) A2 (n) A3 (n)
SC1
SC2
Mean
44.8 (19) 59.8 (16.6) 87 (32) 76 (26) 9 (10) 2 (1)
46.2 (17) 58 (17.6) 86 (38) 67 (35) 15 (9) 4 (4)
45.5 (18) 58.9 (17) 86.5 (34) 71.5 (29) 12 (9) 3 (2)
a CAP, cyclic alternating pattern; A1, phase A1 subtype; A2, phase A2 subtype; A3, phase A3 subtype; S.Ds in parentheses.
SC2
Descending branch (min) S1 (min) S2 (min) S3 (min) S4 (min)
23.9 4.6 12.7 6.6 0
(11.3) (3.4) (5) (5.4)
36.6 (18.4) 0.6 (1.3) 31 (16.3) 5 (6.3) 0
Trough (min) S1 (min) S2 (min) S3 (min) S4 (min)
47.4 (6.9) 0 0 7.2 (6.3) 40.2 (20.1)
35.4 (19) 0 0 12.6 (7.2) 22.8 (23.6)
3.9 (0.8) 0 3.9 (1.1) 0 0
7.6 (6.4) 0 7.6 (6.4) 0 0
Ascending branch (min) S1 (min) S2 (min) S3 (min) S4 (min)
a SC1, first sleep cycle; SC2, second sleep cycle; S1, S2, S3, S4, stages 1, 2, 3, 4; NREM, non-REM sleep; REM, REM sleep; standard deviations in parentheses.
and higher in CAP (mean SI, 3.4). The CAP-related activation of epileptic events was linked to the single A phases (Fig. 3), while the B phases exerted a shallow action, even compared with the NCAP condition. EEG paroxysms were 1302 in the descending branches, 1188 in the troughs and 72 in the ascending branches. The distribution of interictal bursts in the two SCs is reported in Table 6. Both SCs showed a progressive decrease of activation across the three units, from the highest SI values reached during the descending branches (mean SI, 3.6) to the more attenuated SI values during the troughs (mean SI, 2.4) down to the lowest values during the ascending limbs (mean SI, 1.1). The magnitude of activation during the descending branches was closely related to the CAP condition (mean SI, 5.2) and to the powerful effect of phase A (mean SI, 13.9) (Fig. 4). In each unit, the SI quantified during the A phases of CAP was three-fold higher compared with the values expressed throughout the entire unit. About 82% of all the EEG discharges occurring in phase A were distributed within a subtype A1.
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
14
Table 4 CAP parameters in the main units (descending branch, trough, ascending branch) of the first two SCsa SC1
SC2
Descending branch (min) CAP (min) CAP rate (%) A1 (n) A2 (n) A3 (n)
23.9 (11.3) 14.1 (10.8) 59 21 (13) 3 (2) 0
36.6 (18.4) 14.2 (14.3) 38.8 19 (9) 4 (2) 0
Trough (min) CAP (min) CAP rate (%) A1 (n) A2 (n) A3 (n)
47.4 (6.9) 29 (11.5) 61.1 54 (21) 1 (1) 1 (1)
35.4 (19) 25.4 (11) 71.7 48 (33) 3 (2) 0
Ascending branch (min) CAP (min) CAP rate (%) A1 (n) A2 (n) A3 (n)
3.9 (0.8) 1.7 (0.9) 43.6 1 (1) 5 (2) 2 (1)
7.6 (6.4) 6.6 (3.1) 86.8 0 8 (3) 4 (2)
a SC1, first sleep cycle; SC2, second sleep cycle; CAP, cyclic alternating pattern; CAP rate, ratio of CAP time to NREM sleep time throughout the SC; A1, A2, A3, phase A subtypes of CAP; S.D.s in parentheses.
4. Discussion There is consolidated evidence that states of extreme waking or deep sleep are much less conducive to interictal bursting compared with intensive fluctuations of the vigilance level (Halasz, Table 5 Spike indexes referred to the first two SCsa SC1 TST NREM REM CAP NCAP Phase A Phase B
3.1 3.4 0.9 4.3 2.1 10.7 0.5
(1.1) (1.6) (0.6) (2.2) (1) (6.2) (0.2)
SC2
Mean
1.9 2.1 0.8 2.4 1.7 6.6 0.1
2.5 (0.8) 2.8 (1.4) 0.8 (0.5) 3.4 (1.4) 1.9 (0.9) 8.7(4.9) 0.3 (0.2)
(0.7) (1) (0.5) (1.2) (0.8) (3.8) (0.1)
a SC1, first sleep cycle; SC2, second sleep cycle; TST, total SC time; NREM, NREM sleep time; REM, REM sleep time; CAP, cyclic alternating pattern; NCAP, non-CAP; S.D.s in parentheses.
1991). CAP is a major marker of arousal instability that accompanies the transitional phases (Terzano et al., 1988) and, therefore, acts as a convulsive-promoting agent (Parrino et al., 2000). In the current study, activation of generalized EEG abnormalities was higher during CAP compared with the stable NCAP sleep condition. Confirming the data of previous reports (Terzano et al., 1989, 1991a,b, 1992, 1997; Gigli et al., 1992; Oldani et al., 1998), the A phases of CAP operated as permissive windows of epileptic events in both SCs and in each SC unit. Overall, activation was about 30-fold stronger in phase A compared with phase B. The great majority of interictal discharges occurred during subtypes A1, suggesting a crucial association between generalized bursts and the slow arousal components of CAP (Terzano and Parrino, 2000). It is known that the synchronizing activity of sleep decreases homeostatically across the successive SCs (Borbely, 1982). Like EEG synchrony, activation of EEG paroxysms diminishes as sleep progresses. This effect is mostly evident over the first two SCs (Kellaway et al. 1980). In accordance with these findings, the discharge rates in our patients with PGE decreased from SC1 to SC2, with higher values quantified during NREM sleep compared with REM sleep. The stronger influence of NREM sleep upon interictal discharges is attributed to the greater synchronization of its EEG activities. In the light of this, the maximum of epileptic activation within the SC should coincide with the climax of EEG synchrony. The present findings indicate, however, that the bursting activity in PGE is higher before the onset of stage 4 that represents the maximum expression of EEG synchrony within the SC. The SC is organized in regular patterns based upon the alternation of NREM and REM sleep, which is repeated several times during the night. This two-fold nature of sleep within the SC framework is regulated by the reciprocal activation of antagonistic cell groups, the REM-off neurons that operate during the genesis and consolidation of deep NREM sleep, and the REM-on neurons that fire immediately before and during REM sleep (McCarley and Hobson, 1975). There is experimental evidence
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
15
Fig. 3. Bottom: enhancement of one of the histograms reported in Fig. 1 allowing a clearer representation of CAP sequences (here depicted by black dots) across the first two SCs. Middle: number of A phases (Ph. A) per 5 min of sleep. The vertical black dots coincide to subtypes A1, the vertical black lines correspond to subtypes A2 and A3. Top: number of spike-and-waves (P.O.) per 5 min of sleep. Compared with the bursting activity expressed during stable sleep (without CAP sequences), generalized discharges occur more frequently, when they arise in association with a phase A of CAP.
that paroxysmal activities in corticothalamic networks develop progressively through increasing levels of neuronal synchronization during sleep EEG patterns (Steriade and Contreras, 1995). Accordingly, activation of generalized discharges was three-fold higher in the descending branch (mean SI, 3.6) and two-fold higher in the trough (mean SI, 2.4) compared with the ascending branch (mean SI, 1.1), which showed discharge rates similar to the ones measured in REM sleep (mean SI, 0.9). A further promoting factor of bursting was related to the time required to achieve sustained SWS. Comparing the two descending branches, activation of interictal bursts was greater in SC1,
that more rapidly reached full EEG synchrony due to the stronger initial sleep pressure (Borbely, 1982). In other words, the steeper the descending slope the more potent the activation. This can explain why sleep recovery after sleep deprivation, that actually accelerates the accomplishment of deep NREM sleep, is effective in provoking generalized discharges (Shouse et al., 1996). The stronger activating influence of rising EEG synchrony is not limited to PGE. Investigation carried out on children with partial epilepsy (either symptomatic or cryptognetic) confirmed the strong activating properties of NREM sleep on interictal bursts, with higher discharge rates measured in stages 2 and 3 of the descending branches
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
16
Table 6 Spike indexes in the different units of the first twoSCs Descending branch (DB)
Trough (T)
Ascending branch (AB)
P-value
Statistical comparisons
(a) NREM component of SC1 NREM 5.2 (2) CAP 6.6 (1.7) NCAP 3.2 (1.5) Phase A 17.4 (10) Phase B 0.5 (0.2)
2.7 3.4 1.6 8.2 0.4
(1.8) (1.9) (0.6) (4.7) (0.2)
1.3 1.4 1.3 3.7 0.2
0.003 0.0001 0.01 0.005 0.02
DB\AB DB\T= AB DB\T= AB DB\AB DB\AB
(b) NREM component of SC2 NREM 2.5 (1.7) CAP 3.7 (1.3) NCAP 1.7 (0.8) Phase A 10.4 (5.6) Phase B 0.2 (0.1)
2.1 2.3 1.3 5.9 0.1
(1.3) (1.4) (0.6) (4.1) (0.1)
0.9 (0.7) 0.9 (0.5) 1 (0.4) 3.1 (2) 0
(Nobili et al., 1999). In adults, affected by partial epilepsy, interictal EEG discharges were more frequent just before sleep reached the climax of EEG synchrony within each SC (Malow et al., 1998). In that study, the smoothing of the delta power curves obscured the short-lasting fluctuations of sleep depth that are known to precede the peak of EEG synchrony within the SC (Feinberg, 1989). The achievement of EEG synchrony is not a stepwise process, but it is shaped progressively through the recurring CAP oscillations. Spectral analysis has shown that, in the evolution from light to deep NREM sleep, the total power of the oscillation increases, while the amplitude gaps decline between the maximum (phase A) and minimum (phase B) values of the oscillation (Ferrillo et al., 1997). In other words, in the generation of EEG synchrony CAP cycles undergo considerable power variations. The major activating role played by CAP in the descending branches indicates that it is not massive EEG synchrony, but broad amplitude shifting of delta power that triggers interictal responses in PGE. Once the climax of EEG synchrony is reached, the swings of arousal level take place within a narrower range and activation is attenuated. In effect, the troughs of our PGE patients presented elevated values of CAP rate (for comparison with normative data see Terzano et al., 2000), but still their bursting activity was lower compared with the descending branches.
(1) (1.4) (0.7) (2.1) (0.1)
NS 0.003 NS 0.02 0.003
DB\AB DB\AB DB\AB
While the generation of EEG synchrony promoted interictal firing, the ascending branches, under the neurophysiological domain of REM sleep, operated as inhibitory attractors, as reflected by the convergent SI values measured during CAP and NCAP (Fig. 4). Attenuation of bursting activity across the SCs affected the pro-
Fig. 4. Number of EEG bursts per minute of sleep (SI) during NREM sleep (NREM), CAP, non-cyclic alternating pattern (NCAP), phase A and phase B. In spite of the overlapping values expressed by CAP and NCAP in the ascending branch, the activating effect of phase A and the inhibitory action of phase B still persist in this SC unit.
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
moting impact of A phases that, however, expressed in each unit discharge rates well over the ones related to any other sleep variable. The powerful inhibitory influence of phase B mirrored the declining trend of phase A. In conclusion, the present study confirms the hypothesis that generalized epileptic events share common neurophysiological pathways with the mechanisms involved in EEG synchrony. As CAP participates in the production and attenuation of EEG synchrony, it is not surprising that the modulatory effect of CAP on generalized interictal discharges varies in relation to the dynamics of the SC. Within the first two SCs, the features of NREM sleep endowed with the major activating power on generalized bursts are represented by the generation of EEG synchrony and by the A phases of CAP that regulate its build-up.
References American Sleep Disorders Association (ASDA) 1992. EEG arousals: scoring rules and examples. A preliminary report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association. Sleep 15, 174–84. Borbely, A.A., 1982. A two process model of sleep regulation. Hum. Neurobiol. 1, 195–204. Boselli, M., Parrino, L., Smerieri, A., Terzano, M.G., 1998. Effect of age on EEG arousals in normal sleep. Sleep 21, 361– 367. Feinberg, I., 1989. Effects of maturation and aging on slowwave sleep in man: implications for neurobiology. In: Wauquier, A., Dugovic, C., Radulovacki, M. (Eds.), Slow Wave Sleep: Physiological, Pathophysiological and Functional Aspects. Raven Press, New York, pp. 31–48. Ferrillo, F., Gabarra, M., Nobili, L., Parrino, L., Schiavi, G., Stubinski, B., Terzano, M.G., 1997. Comparison between visual scoring of cyclic alternating pattern (CAP) and computerized assessment of slow EEG oscillations in the transition from light to deep non-REM sleep. J. Clin. Neurophysiol. 14, 210–216. Gigli, G.L., Calia, E., Marciani, M.G., Mazza, S., Mennuni, G., Diomedi, M., Terzano, M.G., Janz, D., 1992. Sleep microstructure and EEG epileptiform activity in patients with juvenile myoclonic epilepsy. Epilepsia 33, 799–804. Halasz, P., 1991. Sleep, arousal and electroclinical manifestations of generalized epilepsy with spike wave pattern. In: Degen, R., Rodin, E.A. (Eds.), Epilepsy, Sleep and Sleep Deprivation. Elsevier, Amsterdam, pp. 43–48. Janz, D., 1962. The grand-mal epilepsies and the sleeping-waking cycle. Epilepsia 3, 69–109.
17
Kellaway, P., Frost, J.D., Crawley, J.W., 1980. Time modulation of spike-and-wave activity in generalized epilepsy. Ann. Neurol. 8, 491– 500. Malow, B.A., Lin, X., Kushwaha, R., Aldrich, M.S., 1998. Interictal spiking increases with sleep depth in temporal lobe epilepsy. Epilepsia 39, 1309– 1316. McCarley, R.W., Hobson, J.A., 1975. Neuronal excitability modulation over the sleep cycle: a structural and mathematical model. Science 189, 58 – 60. Merica, H., Gaillard, J.-M., 1986. Internal structure of sleep cycles in a healthy population. Sleep 9, 502– 513. Nobili, L., Baglietto, M.G., Beelke, M., De Carli, F., De Negri, E., Rosadini, G., De Negri, M., Ferrillo, F., 1999. Modulation of sleep interictal epileptiform discharges in partial epilepsy of childhood. Clin. Neurophysiol. 110, 839– 845. Oldani, A., Zucconi, M., Asselta, R., Modugno, M., Bonati, M.T., Dalpra`, L., Malcovati, M., Tenchini, M.L., Smirne, S., Ferini-Strambi, L., 1998. Autosomal dominant nocturnal frontal lobe epilepsy. A video-polysomnographic and genetic appraisal of 40 patients and delineation of the epileptic syndrome. Brain 121, 205– 223. Parrino L, Smerieri A, Spaggiari MC, Terzano M.G., 2000. Cyclic alternating pattern (CAP) and epilepsy during sleep: how a physiological rhythm modulates a pathological event. Clin Neurophysiol, in press. Parrino, L., Boselli, M., Spaggiari, M.C., Smerieri, A., Terzano, M.G., 1998. Cyclic alternating pattern (CAP) in normal sleep: polysomnographic parameters in different age groups. Electroencephalogr. Clin. Neurophysiol. 107, 439– 450. Rechtschaffen A, Kales A., 1968. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. Brain Information Service/Brain Research Institute, University of California at Los Angeles, Los Angeles. Shouse, M.N., Martins da Silva, A., Sammaritano, M., 1996. Circadian rhythms, sleep, and epilepsy. J. Clin. Neurophysiol. 13, 32 – 50. Steriade, M., Contreras, D., 1995. Relations between cortical and thalamic cellular events during transition from sleep patterns to paroxysmal activity. J. Neurosci. 15, 623– 642. Steriade, M., Gloor, P., Llinas, R.R., Lopes da Silav, F.H., Mesulam, M.-M., 1990. Basic mechanisms of cerebral activities. Electroencephalogr. Clin. Neurophysiol. 76, 481– 508. Terzano, M.G., Parrino, L., 2000. Origin and significance of cyclic alternating pattern (CAP). Sleep Med. Rev. 4, 101– 123. Terzano, M.G., Mancia, D., Salati, M.R., Costani, G., Decembrino, A., Parrino, L., 1985. The cyclic alternating pattern as a physiologic component of normal NREM sleep. Sleep 8, 137– 145. Terzano, M.G., Parrino, L., Spaggiari, M.C., 1988. The cyclic alternating pattern sequences in the dynamic organization of sleep. Electroencephalogr. Clin. Neurophysiol. 69, 437– 447.
18
L. Parrino et al. / Epilepsy Research 44 (2001) 7–18
Terzano, M.G., Parrino, L., Anelli, S., Halasz, P., 1989. Modulation of generalized spike-and-wave discharges during sleep by cyclic alternating pattern. Epilepsia 30, 772– 781. Terzano, M.G., Parrino, L., Garofalo, P.G., Durisotti, C., Filati-Roso, C., 1991a. Activation of motor seizures during cyclic alternating pattern in sleep. Epilepsy Res. 10, 166– 173. Terzano, M.G., Parrino, L., Spaggiari, M.C., Barusi, R., Simeoni, S., 1991b. Discriminatory effect of cyclic alternating pattern in focal lesional and benign rolandic interictal spikes during sleep. Epilepsia 32, 616–628.
Terzano, M.G., Parrino, L., Anelli, S., Boselli, M., Clemens, B., 1992. Effects of generalized interictal EEG discharges on sleep stability: assessment by means of cyclic alternating pattern. Epilepsia 33, 317– 326. Terzano, M.G., Monge-Strauss, M.F., Mikol, F., Spaggiari, M.C., Parrino, L., 1997. Cyclic alternating pattern as a provocative factor in nocturnal paroxysmal dystonia. Epilepsia 38, 1015– 1025. Terzano, M.G., Parrino, L., Boselli, M., Smerieri, A., Spaggiari, M.C., 2000. CAP components and EEG synchronization in the first three sleep cycles. Clin. Neurophysiol. 111, 283– 290.
.