The effect of sleep deprivation on the EEG in epilepsy

Electroencephalography and clinical Neurophysiology, 1983, 55:505-512 Elsevier Scientific Publishers Ireland, Ltd.

505

THE EFFECT OF SLEEP DEPRIVATION ON THE EEG IN EPILEPSY R. VELDHUIZEN, C.D. BINNIE and D.J. BEINTEMA lnstituut voor Epilepsiebestrijding Meer en Bosch/De Cruquiushoeve, 2100 AA Heemstede (The Netherlands) (Accepted for publication: December 21, 1982)

Upwards of 25 different published studies of EEG recording after sleep deprivation in persons with epilepsy are unanimous in claiming that this procedure increases the yield of epileptiform activity and is therefore by implication of diagnostic value. Estimates of the incidence of activation differ widely, from 6.9% (Broeker et al. 1973) to 83% (Spadetta 1971). Some authors appear to regard sleep deprivation (SD) chiefly as a convenient way of inducing sleep in the EEG laboratory and attribute the EEG effects simply to drowsiness and sleep (White et al. 1962; Broeker et al. 1973; Gajkowski and Zalejski 1976; Ritter et al. 1977; Scollo-Lavizzari et al. 1977). Others claim that there is a specific activating effect of sleep deprivation per se (amongst others, Pratt et al. 1968; Geller et al. 1969; Spadetta 1971; Bechinger et al. 1973; UI~ et al. 1976; Deisenhammer and Klingler 1978; Degen 1980; Tartara et al. 1980). Many simply note the increase in epileptiform activity during sleep recording after sleep deprivation without discussing the mechanism. Only 2 studies compare sleep after SD with that induced by drugs (Gajkowski and Zalejski 1976; Rumpl et al. 1977). In view of the known precipitation of seizures by prolonged sleeplessness in persons who have no previous history of epilepsy (Bennett et al. 1964, 1969; Gunderson et al. 1973), activation of epileptiform EEG activity by SD would not be unexpected. Nevertheless, though the literature is extensive and the number of patients reported exceeds 5500, the evidence for EEG activation by SD other than simply as a result of sleep induction is not robust. Virtually all reported studies were subject to the exigencies of routine clinical practice. Some authors clearly state that EEGs following SD were per-

formed only where an initial waking record had failed to support the diagnosis of epilepsy (Pratt et al. 1968; Geller et al. 1969; Spadetta 1971; Bechinger et al. 1973; Broeker et al. 1973; Domzal and Zalejski 1973; Rumpl et al. 1977; ScolloLavizzari et al. 1977). Others give no criteria for undertaking SD studies but report such a low incidence of abnormalities in the routine EEGs that it appears that they were following a similar practice. A few authors mention other indications: for instance Degen (1977) sometimes recorded after SD where advice was required concerning the choice of medication and 22% of his initial tracings were abnormal. Mattson et al. (1965) specifically included a group of patients with abnormal initial EEGs in their studies. Beck et al. (1977) apparently undertook an investigation specifically for research purposes in 10 children with known spike-wave discharges. Almost all authors appear to assume that a negative initial EEG in these selected patients would ordinarily have remained normal and that the appearance of epileptiform activity in a subsequent tracing was attributable to activation following SD. The evidence from monitoring studies of the extreme hour-to-hour fluctuation of the EEG in persons with epilepsy (Milligan and Richens 1982; Binnie 1982) has merely served to emphasize the general experience, documented long ago by Ajmone Marsan and Zivin (1970), that the findings in routine waking EEGs of persons with epilepsy are extremely variable. In their extensive series, some 25% of patients consistently exhibited epileptiform activity, even in 10 or more waking records. Some 15% were equally consistent in never exhibiting inter-ictal discharges, whereas the majority showed epileptiform activity in some rec-

0013-4649/83/0000-0000/$03.00 © 1983 Elsevier Scientific Publishers Ireland, Ltd.

506

ords and not in others. If one selects from such a population those patients not showing epileptiform activity in the initial EEG, a second recording must inevitably demonstrate a higher discharge incidence than the first, due to the statistical phenomenon of regression to the mean. This methodological problem appears to be appreciated by only one group of authors (Pratt et al. 1968) who repeated the routine EEG in those patients exhibiting epileptiform activity after SD, and concluded that 18% of their observec[ 41% incidence of activation was attributable to this effect. Geller et al. (1969) also repeated the routine EEG in 7 subjects, but with negative results. Another serious methodological problem arises from the fact that the sleep records were generally of considerably greater duration than the routine recordings and the probability of registering an intermittent phenomenon is obviously related to the period of observation. Bergonzi et al. (1970) employed quantitative measures of discharge rate, and 2 other studies (Welch and Stevens 1971; Bechinger et al. 1973) take some account of recording time in comparing the routine with SD records or contrasting wake and the various sleep stages. Mattson et al. (1965) distinguish qualitative and quantitative EEG changes after SD and cite discharge rates for 1 patient. Serial effects also need to be taken into account: it is unlikely that the mental state of an anxious patient undergoing his first EEG recording will be the same in the waking phase of a subsequent investigation after familiarization with the procedure. The action, usually suppressant, of arousal on epileptiform activity has long been recognized as a pitfall in quantitative studies of epilepsy (Margerison et al. 1967). In summary, to establish the effect, if any, of SD on epileptiform activity in persons with epilepsy it is necessary to perform studies in which routine and SD recordings are performed in random order and in which the incidence of epileptiform activity is quantified and corrected for the duration of each recording or stage of sleep or wakefulness. To distinguish possible activating effects of SD from those of sleep such a randomized design should also include recording during sleep induced by other means, for instance hypnotic drugs. Such a study is now reported.

R. V E L D H U I Z E N ET AL.

Material and Methods The observation that a colleague habitually requested a full series of routine, quinalbarbitone-induced sleep and sleep deprivation recordings in all patients, provided the opportunity for a randomized study. During 18 months all 72 patients admitted to one of the adult observation wards of the Instituut voor Epilepsiebestrijding were included, with the exclusion of three in whom the seizure incidence was so high that the risk of exacerbation by sleep deprivation was considered unacceptable. The principal characteristics of the group are listed in Table I. All the patients suffered from known or suspected epilepsy but the indications for admission were varied, including diagnosis, adjustment of medication, and evaluation of psycho-social problems. Sixty were taking antiepileptic drugs on admission. The initial EEG evaluation was performed without change of medication and generally within the space of 1 week. A routine EEG, a tracing with barbiturate-induced sleep (BIS) and a registration after a night's SD (that is 24-26 h wakefulness) were performed in random order. The routine record of approximately 30 min registration time included a period of hyperventilation and intermittent photic stimulation and sometimes spontaneous drowsiness or sleep. The BIS recording was obtained after oral administration of quinalbarbitone 200 mg and comprised typically 15 min waking inclusive of hyperventilation and some 25 min sleep. The recording after SD commenced at about 08.30 a.m. and included some 10 min waking, hyperventilation, approximately 30 min sleep and finally photic TABLE I Sex Age

31 male 38 female Mean 30.3 years_+ 11.5 years Reasons for admission

Diagnosis Seizure control Psychosocial

Principal indication

Secondary indication

52 15 2

19 32

507

SLEEP D E P R I V A T I O N IN EPILEPSY T A B L E II EEG conclusion

Final clinical diagnosis Secondary gen. epilepsy

Partial epilepsy

Total

Not epilepsy

Uncertain

Epilepsy unclassified

Primary gen. epilepsy

Not epilepsy Uncertain Epilepsy unclassified Primary gen. epilepsy Secondary gen. epilepsy Partial epilepsy

11 5

1 4

0 0

0 0

1 0

1 2

14 11

2

I

4

1

1

2

I1

0

()

0

1

0

0

1

0 1

(J 1

0 1

0 0

8 0

2 19

10 22

Total

19

7

5

2

10

26

69

stimulation. The EEGs were recorded on a 20- or 21-channel EEG machine. Chlorided silver stick-on electrodes were employed, positioned according to the 10/20 system with the addition as indicated of a non-standard anterior-temporal placement a n d / o r nasopharyngeal electrodes. The EEGs were interpreted for clinical purposes by one observer and subsequently analyzed in random order and without access to clinical information by another. For the purposes of this second analysis epileptiform activity (spikes, spikewave complexes or sharp waves) were classified as generalized or focal. The incidence of epileptiform activity (EA) was counted and subdivided for the various stages of wake or sleeping following the classification of Dement and Kleitman (1957) and expressed as events per minute. The final decision concerning the diagnosis of the presence or type of epilepsy was made at least 6 months after admission by a clinician not involved with the EEG investigations but who had had access to EEG reports. The final clinical diagnoses are compared with the EEG conclusions based on the initial 3 tracings in Table II. It will be seen that where the 'uncertain' categories are excluded the EEG supported the clinical diagnosis of the presence or absence of epilepsy in 50 out of 55 subjects.

Results Mean discharge rate The amounts of EA varied greatly within and across subjects and were not normally distributed. It was therefore necessary to use non-parametric methods for the statistical analysis. The mean discharge rate for each of the 3 recordings of any subject fell most often in the rank order sleep deprivation > barbiturate-induced sleep > routine and this effect was significant at the 0.001 level by the Friedman test. However, when the waking sections and those in the various sleep stages are separately analyzed no such rank order emerges (Table III). There is no evidence of EEG activation in the waking state after SD. The yield in stage I sleep is lower in the routine than in the other records but this may be an experimental artefact in that the spontaneous stage I of the routine record may have been less deep than that following quinalbarbitone or SD when the majority of patients passed through stage I to attain deeper sleep. Within-record paired comparisons of discharge rate (Table IV) indicated that the discharge rate was most often greatest in stage I.

508

R. V E L D H U I Z E N ET AL

T ABLE lII Paired comparisons of discharge rate. Conditions

Number of patients and significance level Total recording time

Wake

BIS B > R v B=R Routine B < R

39 7P<0.05 23

27 19 23

16 12 P < 0.05 7

SD S> R v S= R Routine S < R

46 3 P < 0.005 20

30 16 23

16 11 P < 0.065 8

SD v BIS

47 2 P < 0.005 20

28 16 25

28 11 26

S> B S= B S< B

Sleep stage IlI

Too

few data

6 7 P < 0.02 1

21 14 13

9 8 4

Sign test, single-tailed for total and waking, two-tailed for sleep stage comparison.

Presence or absence of epileptiform activity An alternative way of considering these data which may be of more direct relevance to practical E E G interpretation but unfortunately does not allow for the effect of the varying duration of different records or sleep stages is to consider only the presence or absence of EA (Table V). The results are similar to those above: the incidence of EA in the total records falls again in the rank order sleep deprivation > barbiturate-induced sleep > routine, but this effect disappears when the state of wake is separately considered (Table V above).

TABLE IV Within record paired comparisons of discharge rate. Wake/sleep

Wake W > I W= I W< I

v I

Wake

I1 Wake v

Mean discharge rate after exclusion of trivial findings It will be noted that although 19 patients were considered on clinical grounds not to have epilepsy there were only 2 subjects in whom no single epileptiform phenomenon was noted. The incidence of EA in those not thought to have epilepsy was, however, generally very low and confined to a few sporadic spikes or sharp waves. A clearer picture of the yield of new information likely to be of diagnostic importance may be obtained by considering only those patients in whom the discharge rate, either in waking or in one or olher sleep stage

Record

stage

IIl

W > II W = II W < II W>III W = III W
Wake W > IV

v IV

w = IV w < 1V

I v II

I > II I =ll I < lI

vl

I >III

III

Ill I < Ill I =

Routine

BIS

SD

10 12 14

13 15 P < 0.001 38

20 12 P < 0.001 36

17 12 23

18 11 P < 0.001 33

StagellI occurred in only 5 patients

11 8 7

17 10 22

Stage IV did not occur

Stage IV occurred in only 4 patients

5 5 6

6 7 3

-

7 4 7

25 11 16

27 10 25

16

19 I1 19

6 P < 0.001

4

Significance levels based on two-tailed sign test.

509

SLEEP DE PRIVATION IN EPILEPSY

Presence ( + ) or absence ( - ) of epileptiform activity.

of EA in stage II sleep after SD. In the waking state there are no significant effects.

Routine

Activation procedures

TABLE V

N

BIS -

SD +

-

BIS and SD +

+

24 45 69

7 3 10

Totalrecord 17 3 42 2 59 5

-

34 35 69

19 7 26

15 28 43

-

21 43 64

2 0 2

Wake +

16 11 27

18 24 42

11

4 15

in at least one of the records (Table VI) exceeded an arbitrary level of 1 event/min. Fifty-two patients met this criterion but it will be seen in Table VI that the pattern is essentially unchanged. On the basis of the total recording a majority of patients showed a higher mean discharge rate after SD than in the BIS tracing or the routine record, but these differences largely disappear when awake and the various sleep stages are considered separately. Apart from the higher incidence, already noted, of EA in stage I of BIS or SD tracings in comparison with routine, the only remaining significant finding is a greater incidence

Hyperventilation was accompanied by a significant increase in discharge rate over the resting level in all 3 recordings. This apparent activating effect was slightly but not significantly stronger after SD. A significantly greater incidence of epileptiform activity after hyperventilation than in stage I sleep was seen only after SD. In the routine record two patients exhibited a photoconvulsive response to intermittent photic stimulation. In the BIS record photic stimulation was not performed. After SD the same two patients remained photosensitive and two others exhibited a photoconvulsive response. The numbers are small but at least in accordance with previous reports (Scollo-Lavizzari and Scollo-Lavizzari 1974) that SD leads to an increased incidence of photosensitivity.

Type of epileptiform activity Table VIII lists the findings in wake and stage I sleep for focal and generalized discharges. It will be noted that in the waking state the incidence of focal discharges was actually greater in the routine recording than in either of the others and greater before BIS than after SD; these effects are not

TABLE VI Paired comparisons of discharge rate. Conditions

Number of patients and significance level Total recording time

Wake

BIS B> R v B=R routine B < R

34 1 P<0.05 17

25 6 21

16 5 P<0.05 6

4 6 1

SD S> R v S=R Routine S < R

37 1 P<(I.065 14

27 4 21

15 5 8

6 7P<0.05 0

SD v BIS

36 0 P < 0.005 16

26 4 22

23 4 23

S> B S= B S< B

Sleep stage I

Sign test, single-tailed for total and waking, two-tailed for sleep stage comparison.

II

19 7 11

III

]

Too few data

510

R. V E L D H U I Z E N

T A B L E VII W i t h i n r e c o r d p a i r e d c o m p a r i s o n o f d i s c h a r g e r a t e effect of HV. Record Routine

BIS

SD

Resting W < HV v W=HV HV W>HV

30 24P<0.005 12

31 24P<0.001 11

37 12P<0.001 10

Stage I v HV

12 12 9

19 13 25

30 11 P < 0.04 17

I < HV I=HV I > HV

significant. For generalized discharges, however, there is significant evidence of activation by SD in the waking state only. Where stage I occurred spontaneously during the routine record, the incidence of focal discharges was significantly less than in stage I of either BIS or SD records (but see comment on Table III above). Discussion

It is thought that the present study avoids some of the methodological weaknesses which render

T A B L E VIII P a i r e d c o m p a r i s o n s o f d i s c h a r g e rate. Conditions

N u m b e r o f p a t i e n t s a n d s i g n i f i c a n c e level Focal

Generalized

Wake

Stage I

BIS R > B v R=B Routine R < B

9 5 a 13

10 1b 3

5 7 a 3

3 2a 5

SD v

11 3a

9 1b

12 l c

5 0 ~

Routine R < N

13

3

2

SD v BIS

8 7 a 12

15 1 12

12 0 d 3

a

R>N R=N

B> N B=N B< N

n.s.

b p < 0.05. P < 0.002. d P < 0.01.

Wake

Stage I

5 6 2a 6

ET AL.

previous investigations of the effect of SD on the EEG in epilepsy difficult to interpret. A price has been paid, however, in that such investigation was possible only in a highly selected group of patients in a special centre for epilepsy. Primary generalized epilepsy is virtually unrepresented and the findings of the present study are therefore relevant only to partial and secondary generalized epilepsy. This is perhaps a less serious limitation than it may at first appear, as in those patients where uncertainty concerning the clinical diagnosis of epilepsy leads most reliance to be placed upon EEG findings, the attack pattern is generally incompatible with the primary generalized form and most often suggestive of complex partial seizures. The present findings, in an admittedly selected group of patients, confirm the already well-documented activating effect of sleep on the EEG in epilepsy. The mean discharge rates over the entire record after SD were significantly higher than in the BID and routine records. However, this finding was entirely explained by the greater proportion of recording time spent in sleep after SD. There was no evidence of an overall increase of discharge rate after SD when awake and the various sleep stages were separately compared with corresponding states in the routine and BID records. Specific activation by sleep deprivation, independent of sleep induction, was confined to generalized discharges, and indeed focal discharges showed a reduced incidence in waking after SD. It should perhaps be noted that the results of this study do not reflect the general utility of the EEG as a diagnostic aid in epilepsy. Though the level of agreement between electroencephalographer and neurologist concerning the final diagnosis was fairly close (Table II), the 52 patients admitted for diagnostic reasons included many referred because of an apparent discrepancy between the clinical picture and EEGs reported elsewhere. Moreover 7 patients subsequently underwent telemetric EEG and video monitoring (resuits not included here), the findings of which were conclusive for the diagnosis in 4. Sleep deprivation is disagreeable for patients and inconvenient for their families or nursing staff, and carries a small but undoubted risk of inducing tonic-clonic convulsions even in persons who have

SLEEP DEPRIVATION IN EPILEPSY never previously suffered this type of seizure. Though it is an effective way of inducing sleep in the EEG laboratory it has no general advantages over hypnotic drugs in persons with suspected partial epilepsy who form the majority of adult referrals for EEG investigation of possible seizure disorders.

Summaff Many published studies report an increased incidence of epileptiform EEG activity following sleep deprivation in persons with epilepsy in whom a previous routine EEG was normal or inconclusive. Few such studies, however, permit a clear distinction to be made between the effects of sleep deprivation per se, sleep induction following deprivation, or simply repeated EEG recording. Sixty-nine patients have been investigated in whom a routine waking record, a secobarbital-induced sleep recording and an EEG following 24 h sleep deprivation were obtained in random order irrespective of whether or not the initial EEG contained epileptiform activity. For each record the incidence of epileptiform activity if any, was quantified in terms of discharges per minute for wake, drowsiness and each sleep stage. The findings confirmed the marked activating effect of light sleep on the EEG but there was no evidence of an overall increase in discharge rate after sleep deprivation either in the waking state or in the various sleep stages when these were compared with secobarbital-induced or spontaneous sleep. There was a significant increase of generalized discharges in the waking state only, after sleep deprivation, but a decrease in incidence of focal epileptiform activity. It is concluded that although specific indications may exist for the use of sleep deprivation, as a general method of EEG activation in epilepsy it has no advantages over barbiturate-induced sleep to offset the greater inconvenience to patients.

511

R6sum6 Effet de la privation de sommeil sur I'EEG clans l'Opilepsie De nombreuses publications rapportent une augmentation de l'incidence d'une activit6 EEG 6pileptiforme apr6s une privation de sommeil chez des patients 6pileptiques dont un EEG pr6alable de routine 6tait normal ou non concluant. Parmi ces 6tudes, peu cependant ont permis de faire une distinction nette entre les effets de la privation de sommeil per se, de l'induction du sommeil qui suivait la privation, ou simplement de la r6p6tition de l'enregistrement EEG. Nous avons 6tudi6 69 patients chez lesquels nous avons pratiqu6 un enregistrement de veille de routine, un enregistrement de sommeil induit par le s6cobarbital et un EEG apr~s une privation de sommeil de 24 h, en ordre al6atoire et sans tenir compte du fait que I'EEG initial pr6sentait ou non une activit6 6pileptiforme. Pour chaque enregistrement on a quantifi6 l'incidence de l'activit6 6pileptiforme, lorsqu'elle existait, en termes de d6charges par minute pour la veille, l'assoupissement et chaque stade de sommeil. Les r6sultats ont confirm6 l'effet activateur marqu6 du sommeil 16ger sur I'EEG, mais ils n'ont pas mis en 6vidence une augmentation globale du taux de d6charge apr6s privation de sommeil, ni pendant la veille ni lors des diff6rents stades de sommeil compar6s ~t ces stades induits par le s6cobarbital ou lors du sommeil spontan6. Une augmentation significative des d6charges g6n6ralisbes ne survenait que pendant l'6tat de veille, apr6s privation de sommeil, tandis que baissait l'incidence de l'activit6 6pileptiforme focalis6e. On en conclut que, bien qu'il existe des indications sp6cifiques de l'utilisation de la privation de sommeil, celle-ci ne pr6sente pas en tant que m6thode g6n6rale d'activation de I'EEG dans l'6pilepsie, d'avantages suffisants par rapport au sommeil induit par les barbituriques, pour compenser l'inconfort qu'elle impose aux patients.

References Ajmone Marsan, C. and Zivin, L.S. Factors related to the occurrenceof typical paroxysmalabnormalities in the EEG

512 records of epileptic patients. Epilepsia, 1970, 11: 361-381. Bechinger, D., Kriebel, J. und Schlager, M. Das SchlafentzugEEG, ein wichtiges diagnostisches Hilfsmittel bei cerebralen Anfaelle. Z. Neurol., 1973, 205: 193-206. Beck, U., Wenzel, D. and Sauer, M. Selective deprivation of sleep in pycnoleptic children. Effects of deprivation of slow-wave sleep and REM sleep on the frequency and duration of petit real attacks. Arch. Psychiat. Nervenkr., 1977, 223: 107-115. Bennett, D.R., Mattson, R.H., Ziter, F.A., Calverley, J.R., Liske, E.A. and Pratt, K.L. Sleep deprivation: neurological and electroencephalographic effects. Aerospace Med., 1964, 35: 888-890. Bennett, D.R., Ziter, F.A. and Liske, E.A Electroencephalographic study of sleep deprivation in flying personnel. Neurology (Minneap.), 1969, 19: 375-377. Bergonzi, P., Cianchetti, C., Ferroni, A. and Marchesi, G. Phase V sleep deprivation in epileptic subjects: changes of discharges and of nocturnal sleep organization. Riv. Neurol., 1970, 40: 344-348. Binnie, C.D. The use of the inter-ictal EEG in the study of antiepileptic drugs. In: P. Buser, W.A. Cobb and T. Okuma (Eds.), Proceedings 10th International Congress of Electroencephalography and Clinical Neurophysiology, Kyoto, 1981. Electroenceph. clin. Neurophysiol., Suppl. 36. Elsevier, Amsterdam, 1982: 504-512. Broeker, H., Sack, G., Muller, D. und Muller, J. Schlaf-EEGUntersuchungen bei unklaren Anfalls2ustanden und episodischen Verhaltensstoerungen. Ein vorlaufiger Bericht. Psychiat. Neurol. Med. Psychol. (Lpz.), 1973, I1: 656-660. Degen, R. Die diagnostische Bedeutung des Schlafs nach Schlafentzug unter antiepileptische Therapie. Nervenarzt, 1977, 48: 314-320. Degen, R. A study of the diagnostic value of waking and sleep EEGs after sleep deprivation in epileptic patients on anticonvulsive therapy. Electroenceph. din. Neurophysiol., 1980, 49: 577-584. Deisenhammer, E. und Klingler, D. Das Schlafentzugs-EEG bei Kranken mit gesicherter und fraglicher Epilepsie sowie bei Kranken ohne epileptische Anfaelle. Z. EEG-EMG, 1978, 9: 38-42. Dement, W. and Kleitman, N. Cyclic variattons in EEG during sleep and their relations to eye movements, body motility and dreaming. Electroenceph. clin. Neurophysiol., 1957, 9: 637-690. Domzal, T. und Zalejski, S. EEG-Provokation durch Schlafentzug bei der Auswahl von Kandidaten fuer spezielle Berufe. Z. EEG-EMG, 1973, 4: 201-202. Gajkowski, K. and Zalejski, S. Comparative evaluation of seizure provocation methods with sleep deprivation and brietal in epileptics. Neurol. Neurochir. pol., 1976, 10: 267-272. Geller, M.R., Gourdji, N., Christoff, N. and Fox, E. The effects

R. VELDHUIZEN ET AL. of sleep deprivation on the EEGs of epileptic children. Develop. Med. Child Neurol., 1969, 11: 771-776. Gordon, N. Intractable epilepsy in childhood. Develop. Med. Child Neurol., 1980, 22: 97-100. Gunderson, C.H., Dunne, P.B. and Feher, T.L. Sleep deprivation seizures. Neurology (Minneap.), 1973, 23: 678-686. Margerison, J.H., St. John-Loe, p. and Binnie, C.D. Electroencephalography. In: P.H. Venables and I. Martin (Eds.), A Manual of Psychophysiological Methods. North-Holland, Amsterdam, 1967: 351-402. Mattson, R.H., Pratt, K.L. and Calverley, J.R. Electroencephalograms of epileptics following sleep deprivation. Arch. Neurol. (Chic.), 1965, 13: 310-315. Milligan, N. and Richens, A. Ambulatory monitoring of the EEG in the assessment of antiepileptic drugs. In: F.D. Stott, E.B. Raftery, D.L. Clement and S.L. Wright (Eds.), Proceedings of 4th International Symposium on Ambulatory Monitoring, Ghent, 1981. Academic Press, London, 1982: 224-233. Pratt, K.L., Mattson, R.H., Welkers, N.J. and Williams, R. EEG activation of epileptics following sleep deprivation: a prospective study of 114 cases. Electroenceph. clin. Neurophysiol., 1968, 24: 11-15. Ritter, G., Becker, A. und Duensing, F. Zum diagnostischen Weft des EEGs nach Schlafentzug. Nervenarzt, 1977, 48: 365-368. Rumpl, E., Lorenzi, E., Bauer, G. und Hengl, W. Zum diagnostischen Wert des EEGs nach Schlafentzug. Z. EEGEMG, 1977, 8: 205-209. Scollo-Lavizzari, G. and Scollo-Lavizzari, G.R. Sleep, sleep deprivation, photosensitivity and epilepsy. Europ. Neurol., 1974, 11: 1-21. Scollo-Lavizzari, G., Pralle, W. and Radue, E.W. Comparative study of efficacy of waking and sleep recordings following sleep deprivation as an activation method in the diagnosis of epilepsy. Europ. Neurol., 1977, 15: 121-123. Spadetta, V. La privazione di sonno nella diagnosi elettroencefalografica di epilessia. Acta neurol. (Napoli), 1971, 26: 7-13. Tartara, A., Moglia, A., Manni, R. and Corbellini, C. EEG findings and sleep deprivation. Europ. Neurol., 1980, 19: 330-334. UI~, M., Schwartzova, K., Vaclavickova, R. and Kleckova, H. The role of one night's wakefulness on provoking epileptic graphoelements in the EEG. Cs. Neurol. Neurochir., 1976, 39: 344-349. Welch, L.K. and Stevens, J.B. Clinical value of the electroencephalogram following sleep deprivation. Aerospace Med., 1971, 42: 349-351. White, P., Dyken, M., Grant, P. and Jackson, L. Electroencephalographic abnormalities during sleep as related to the temporal distribution of seizures. Epilepsia, 1962, 3: 167-174.