Background Automatisms are well recognized to occur in complex partial seizures; however, their occurrence in generalized epilepsies is not always appreciated. There has been considerable debate regarding the nature, triggers, and timing of automatisms in absence seizures.
Objectives To examine the frequency and nature of automatisms in new-onset absence seizures and assess the influence of the state of arousal, provocation, age, and epilepsy syndrome on the presence and type of automatisms.
Design Analysis of absence seizures through video electroencephalogram (EEG) recordings.
Setting British Columbia's Children's Hospital, Vancouver, British Columbia, Canada.
Patients Seventy consecutive children with new-onset untreated absence seizures in idiopathic generalized epilepsy recruited between January 1, 1992, and June 30, 1997.
Main Outcome Measures Each seizure was analyzed for the presence and characteristics of automatisms. The influence of the following variables on the presence of automatisms was statistically analyzed: state of arousal (awake, drowsy, asleep), provocation (hyperventilation, photic stimulation), age, and epilepsy syndrome.
Results Automatisms occurred in 163 of 405 seizures (40%) in 53 of 70 children (76%). Automatisms were more likely in longer seizures and hyperventilation. Only 23% of spontaneous awake seizures had automatisms. Automatisms were similar for an individual child; however, automatisms were not present in all their seizures. Age, epilepsy syndrome, or state of alertness had no effect on the presence of automatisms.
Conclusions Automatisms are frequently seen during childhood absence seizures. The high frequency of automatisms during EEG recordings is predominantly due to the effect of hyperventilation. Their preponderance during longer seizures may relate to opportunity for automatisms to occur. The characteristic pattern of automatisms suggests a reactive phenomenon to internal and external stimuli.
Automatisms are semicoordinated, repetitive motor activities that are associated with impaired awareness and occur in both focal and generalized seizures.1 Reports2-4 suggest that automatisms occur in all children with absence seizures but not in all seizures. Automatisms may be perseverative, when there is continuation of an activity that commenced before the seizure, or de novo, when automatisms start during the seizure.
There has been debate regarding whether automatisms are reactive or an intrinsic property of an absence seizure. The observation that automatisms can be evoked suggests that automatisms can be influenced by external stimuli. De novo automatisms can be induced by an external stimulus, for example, touching the patient with cotton wool, which results in scratching during the seizure.2,3 It has been suggested that scratching spontaneously during an automatism may be a response to an internal stimulus or an unseen stimulus.3 Thus, it was proposed that automatisms could be reactive to the internal and external environment and that they may reflect the circumstance of the individual rather than an intrinsic property of the absence seizure. This theory is consistent with reports that automatisms vary greatly in complexity, location, and character from seizure to seizure within a single patient.2
A contrasting view is that automatisms are an intrinsic component of the seizure and are not random events.5,6 This draws on the observation that the clinical features of an absence seizure follow a stereotyped progression within an individual, although not all clinical features are present in every seizure.5,7,8 The earliest features tend to be irregular respiration and rhythmic myoclonic movements of the eyelids, followed by oral automatisms and finally other motor automatisms. In addition, when clinical features occur, they tend to do so at constant times during the seizure.5
The objective of this study is to describe the clinical features of automatisms during absence seizures in an unselected group of children with new-onset idiopathic generalized epilepsy before the initiation of antiepileptic therapy. We used a statistical model to examine the effect of various factors that could influence the presence of automatisms, such as age, epilepsy syndrome, state of alertness, and type of provocation.
The electroencephalography department of British Columbia's Children's Hospital is a referral center for electroencephalographic (EEG) studies requested by family physicians, pediatricians, and pediatric neurologists. The departmental database of 14 452 EEG recordings was searched for consecutive patients younger than 18 years who had at least 1 absence seizure that was captured during a routine video EEG recording during a 5½-year period between January 1, 1992, and June 30, 1997. The EEG studies were of sleep-deprived patients and included periods of hyperventilation and intermittent photic stimulation (IPS). An absence seizure was defined as a clinical change associated with generalized spike and slow wave or multiple spike and slow wave with a frequency of greater than 2.5 Hz at the onset. The hospital records of patients with an absence seizure were reviewed. Inclusion criteria for the study included normal intelligence, no previous or present use of antiepileptic drugs, and a normal interictal EEG background (excluding epileptiform discharges). Formal consent was obtained for review of the video EEG recordings for research purposes.
Two epileptologists (L.G.S. and K.F.), who were masked to the epilepsy syndrome of the child, independently reviewed the video of each seizure at least 5 times. Only seizures with a clinical sign in which clear video information was available were assessed for automatisms.
Automatisms were classified as oral, limb, or other. Oral and limb automatisms were further classified as either simple or complex. Simple oral automatisms were simple movements of the mouth or tongue that were repeated no more than twice during the seizure. Oral automatisms were complex if they were a combination of tongue and mouth movements or at least 3 simple oral automatisms occurred in a seizure. Limb automatisms were movements of the arms, hands, fingers, or legs. These automatisms were classified as simple if they were brief and the child was only doing a single action, such as picking at his or her clothes. Complex limb automatisms were combinations of movements. Leg involvement could not be assessed in all seizures because only the upper part of the body was videotaped in most seizures. Other automatisms, such as humming, were classified as other. No attempt was made to induce automatisms.
Seizures with clinical signs were assessed for whether the clinical features were consistent with the temporal evolution described by Stefan and colleagues.5 The clinical features were considered consistent with this theory if the order of events comprised (1) eye features (eye opening or abnormal movement of the eye or eyeball), (2) oral automatisms, and then (3) limb automatisms. They were also considered consistent if there were only fragments of this temporal association: for example, if only oral automatisms were seen, then the seizure was included. If the order did not follow this progression, then the seizures were considered to be inconsistent with the temporal hypothesis.
The epilepsy syndrome was classified according to the International League Against Epilepsy 1989 proposal for the revised classification of epilepsies and epileptic syndromes.9 The children with childhood absence epilepsy were further categorized based on the presence or absence of a photoparoxysmal response.
For each patient, data on multiple seizures were recorded. The seizure was the basic unit of analysis. A statistical model accounting for multiple seizures per child and for variables that could potentially influence the presence of automatisms was used. The variables considered were epilepsy syndrome, age, arousal (awake, drowsy, asleep), and provocation (hyperventilation and IPS).
Because the occurrence of automatisms in different seizures within the same patient was not likely to be independent, a random-effects logistic regression model was used with a random patient effect. The log odds ratio of the patient effects was assumed normally distributed. The effects of variables are reported as odds ratio. The between-patient variation modeled by the random effect is reported as the median odds ratio of the presence of automatisms between seizures from 2 randomly chosen patients with identical age, syndrome, state, and provocation.
The study population consisted of 70 children (35 girls) (Table 1) in whom 405 absence seizures were analyzed. Automatisms were seen in 163 of the 405 seizures (40%) in 53 of the 70 children (76%). Complex automatisms occurred in 61 seizures (15%) and 30 children (43%). Oral automatisms were observed in 130 seizures (32%) and limb automatisms in 80 seizures (20%). Eighty-five of the 130 oral automatisms (65%) and 56 of the 80 limb automatisms (70%) were simple. The specific types of automatisms are listed in Table 2 and Table 3. Humming was the only other type of automatism observed and only occurred in 1 seizure in 2 children.
Although the clinical features of automatisms varied greatly from seizure to seizure for each child, there was a tendency for a child to have either oral or manual automatisms. Automatisms in an individual tended to involve the same area of the body. For example, an individual might scratch his ear in the first seizure, pick the same ear in the second seizure, and scratch his head in the third seizure.
Only duration of the seizure and provocation (hyperventilation and IPS) had an effect on the presence of automatisms (Figure 1). Increasing the duration of the seizure by 10 seconds doubled the likelihood of automatisms occurring. Seizures recorded during IPS were 10 times less likely to have automatisms than spontaneous attacks, whereas seizures during hyperventilation were 6 times more likely to have automatisms than those occurring spontaneously. Oral and limb automatisms were more likely to be seen during hyperventilation, with oral automatisms being more frequent (Figure 2). State of arousal or epilepsy syndrome had no effect on the presence of automatisms (Figures 1 and 2).
Automatisms were an inconsistent feature. Of the 70 children, 16 (23%) never had an automatism and 15 (21%) had automatisms in every seizure (Figure 3). The odds ratio for the occurrence of automatisms between 2 randomly selected seizures of the same duration in the same state in 2 children of the same age and syndrome is 4.2. This value is of equivalent magnitude to the effect observed for duration and hyperventilation, which implies that other factors, apart from the variables studied, influence the presence of automatisms.
The average time to first automatism (either oral or manual) was 3.9 seconds (median, 3.0 seconds; SD, 3.6 seconds; range, 0.1-26.0 seconds) (Table 4). Oral automatisms occurred significantly earlier than manual ones (3.04 seconds; P < .001). The duration of the seizure also influenced the time of onset of the first automatism. With every increase of 1 second in duration of the clinical seizure, the time to the first automatism is 0.2 second later (P < .001) (eg, a clinical seizure that is 10 seconds longer will have the first automatism 2 seconds later).
For the 163 absence seizures in which automatisms occurred, 64 (39%) did not follow the sequence suggested by Stefan et al.5 The way in which the progression of features differed from the hypothesis of Stefan et al is detailed in Table 5.
We previously analyzed the clinical and electroencephalographic features of absence seizures in a group of children with new onset idiopathic generalized epilepsy, but did not consider automatisms.10 We assessed the influence of state of arousal, provocation (hyperventilation and IPS), age, and epilepsy syndrome on seizure duration, eye opening, eyelid movements, and level of awareness. We found that the variation in the clinical features was determined by a complex interaction of these variables and undetermined factors specific to the child.10 Herein, we present a study of automatisms in absence seizures. Previous studies2,3,5,6,11 have highlighted the variability in automatisms, which are influenced by both environmental and intrinsic factors. Hence, we considered that automatisms were worthy of more detailed analysis.
The incidence of automatisms in our study is lower than the reported incidence of 88% to 100% of children in other studies.2-4,6,11-13 This variation probably relates to differences in the study populations. Previous studies2-4,6,11-13 include both children and adults with symptomatic generalized epilepsy and intractable absence seizures who were already receiving antiepileptic drugs. In their study, Penry and colleagues3 attempted to induce automatisms by stimulating their patients during a seizure; we did not attempt to elicit automatisms. In contrast, our study is of an unselected cohort of children with new-onset idiopathic generalized epilepsy, who were referred for an EEG because of staring spells, and who were not yet taking medication. The lower incidence of automatisms in our study may reflect these different populations and methods. On the other hand, our study is predicated on the patient having absence seizures during a routine EEG, which means that patients with infrequent seizures may not have been included.
The clinical features of automatisms observed in our study were similar to those detailed in previous studies.2,3,6,11-13 For 56% of children, automatisms were inconsistently present or varied in nature in their seizures; these findings confirmed the observations of other workers.2,3,11
The nature of automatisms may be similar, but not identical, in the seizures of a child. For example, a child might have a variety of manual automatisms, such as rubbing, picking, or scratching at her ear, in a succession of seizures. The automatisms, particularly those involving the upper limbs, were often directed toward an object, such as an electrode, that may have been an irritant. However, more often an environmental trigger was not apparent.
The impact of level of arousal, hyperventilation, and photic stimulation on automatisms has not been reported previously. Automatisms were 6 times more likely to occur in seizures during hyperventilation than in spontaneous seizures in the awake state. Specifically, oral automatisms occurred in 28% of seizures in hyperventilation compared with 17% of spontaneous seizures, and limb automatisms in 56% of seizures in hyperventilation compared with 9% of spontaneous seizures. The widely held view that automatisms, particularly oral ones, are common in absence seizures2,3,6,13 may have arisen because most observed absence seizures occur when elicited by hyperventilation in the clinic and during an EEG recording. Several factors may contribute to the increased incidence of automatisms in hyperventilation. Although it is possible that the pathophysiology of absence seizures in hyperventilation is different, it is more likely that hyperventilation itself influences the clinical features. Hyperventilation is an unnatural state in which people feel dizzy, “tingly,” and uncomfortable and often develop dry lips and mouth. Thus, the automatisms may be reactive to these sensations and consequently may occur more frequently during hyperventilation.
Penry and Dreifuss11 suggested that automatisms are phenomena that are reactive to internal and external stimuli rather than innate manifestations of absence seizures. This theory is based on the finding that automatisms can be induced and modified by external stimuli and are not constantly present. It is supported by the increased incidence of automatisms during hyperventilation and the similar nature of automatisms if there is a clear-cut irritant, such as an electrode. The increased incidence of automatisms in longer seizures may reflect more opportunity for a reactive response. Automatisms were found to have no association with the inherent factors of age and epilepsy syndrome, a finding which is consistent with that of a previous report.2,14
An alternative view suggested that automatisms were innate rather than reactive phenomena. Stefan et al5 proposed that automatisms occur in a relatively consistent sequence: ocular, then oral, and finally manual or pedal automatisms. The onset of the automatism occurred after a fixed period. The sequence of the clinical features of the automatisms in our patients was consistent with the temporal theory of Stefan et al15 in 61% of seizures with automatisms. Although we observed an increased incidence of automatisms in longer seizures, there was no increase in the likelihood of automatisms after 13 seconds of seizure duration. Thus, it would appear that automatisms typically occur at a set time, and, unless artificially induced by external stimuli, oral automatisms occur 3 seconds before manual automatisms.
Our data synthesize the views of Penry et al and Stefan et al, with evidence that there is an inherent progression to a reactive phenomenon. Perhaps this stereotypical pattern of automatisms, with oral automatisms beginning 2 to 3 seconds after seizure onset, followed 3 seconds later by manual automatisms, reflects decreasing inhibition of afferent inputs to the cortex as the seizure progresses. This theory would fit with the concept of reactive automatisms, which surmises some degree of awareness when the child is reacting to an internal or external stimulus. This theory is also supported by the finding that awareness, as assessed by response testing, increases toward the end of the seizure.16,17 The predominance of automatisms during provoked activities rather than spontaneous seizures further supports their reactive nature.
Correspondence: Ingrid E. Scheffer, MBBS, PhD, Austin Health, Level 1, Neurosciences Building, Banksia Street, Heidelberg, Victoria 3081, Australia ().
Accepted for Publication: December 12, 2008.
Author Contributions:Study concept and design: Sadleir, Scheffer, Connolly, and Farrell. Acquisition of data: Sadleir, Smith, Connolly, and Farrell. Analysis and interpretation of data: Sadleir, Scheffer, Connolly, and Farrell. Drafting of the manuscript: Sadleir, Scheffer, and Farrell. Critical revision of the manuscript for important intellectual content: Sadleir, Scheffer, Smith, Connolly, and Farrell. Statistical analysis: Sadleir. Obtained funding: Connolly and Farrell. Administrative, technical, and material support: Sadleir, Smith, and Farrell. Study supervision: Scheffer, Connolly, and Farrell.
Financial Disclosure: None reported.
Additional Contributions: We acknowledge and thank Bendix Carstensen, MSc, Clinical Epidemiology & Biostatistics Unit, Royal Children's Hospital, Melbourne, Victoria, Australia, presently at Steno Diabetes Centre, Gentofte, Denmark, for performing the statistical analysis in this study. We also acknowledge the additional statistical assistance received from John Carlin, PhD, Clinical Epidemiology & Biostatistics Unit, Royal Children's Hospital, Melbourne, Victoria, Australia.
1.
Blume WTLüders HOMizrahi ETassinari Cvan Emde Boas WEngel J Jr Glossary of descriptive terminology for ictal semiology: report of the ILAE task force on classification and terminology. Epilepsia 2001;42 (9) 1212- 1218PubMedGoogle ScholarCrossref
2.
Panayiotopoulos CPObeid TWaheed G Differentiation of typical absence seizures in epileptic syndromes: a video EEG study of 224 seizures in 20 patients. Brain 1989;112 (pt 4) 1039- 1056PubMedGoogle ScholarCrossref
3.
Penry JKPorter RJDreifuss FE Simultaneous recording of absence seizures with video tape and electroencephalography: a study of 374 seizures in 48 patients. Brain 1975;98 (3) 427- 440PubMedGoogle ScholarCrossref
4.
Hirsch EBlanc-Platier AMrescauz C What are the relevant criteria for a better classification of epileptic syndromes with typical absences? Malafosse AGenton PHirsch E et al. Idiopathic Generalized Epilepsies. Clinical, Experimental and Genetic Aspects. London, England John Libby & Co Ltd1994;87- 93Google Scholar
5.
Stefan HBurr WHildenbrand KPenin H Basic temporal structure of absence symptoms. Akimoto AKazamatsuri HSeino MWard A Advances in Epileptology.XIIIth Epilepsy International Symposium. New York, NY Raven Press1982;Google Scholar
6.
Yagi KMorikawa TMiyakoshi M et al. The Duration of Epileptic Absence Seizures and the Accompanying Automatisms. New York, NY Raven Press1982;
7.
Stefan HBurr WHildenbrand KPenin H Computer-supported documentation in the video analysis of absences: preictal-ictal phenomena: polygraphic findings. Dunn MGram LPenry J Advances in Epileptology: XIIth Epilepsy International Symposium. New York, NY Raven Press1981;Google Scholar
8.
Stefan HSnead OC III Absence seizures. Engel J Epilepsy :A Comprehensive Textbook. Philadelphia, PA Lippincott-Raven Publishers1997;579- 590Google Scholar
9.
Commission on Classification and Terminology of the International League Against Epilepsy, Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia 1989;30 (4) 389- 399PubMedGoogle ScholarCrossref
10.
Sadleir LGScheffer IESmith S et al. Factors influencing clinical features of absence seizures. Epilepsia 2008;49 (12) 2100- 2107Google ScholarCrossref
12.
Holmes GL McKeever MAdamson M Absence seizures in children: clinical and electroencephalographic features. Ann Neurol 1987;21 (3) 268- 273PubMedGoogle ScholarCrossref
13.
Shylaja NNegoro TWatanabe K et al. Simultaneous EEG-VTR and transverse topographical analyses of absence seizures in children: some prognostic implications. Brain Dev 1993;15 (2) 91- 96PubMedGoogle ScholarCrossref
15.
Stefan HBurr WPenin H Time structure analysis of motor phenomena in absence epilepsies. Speckman EElger C Epilepsy and Motor System. Munich, Germany Urban und Schwarzenberg1983;310- 324Google Scholar
16.
Browne TRPenry JKProter RJDreifuss FE Responsiveness before, during, and after spike-wave paroxysms. Neurology 1974;24 (7) 659- 665PubMedGoogle ScholarCrossref
17.
Goode DJPenry JKDreifuss FE Effects of paroxysmal spike-wave on continuous visual-motor performance. Epilepsia 1970;11 (3) 241- 254PubMedGoogle ScholarCrossref