|Home | About | Journals | Submit | Contact Us | Français|
Epileptic seizures cause dynamic, reversible changes in brain function and are often associated with loss of consciousness. Of all seizure types, absence seizures lead to the most selective deficits in consciousness, with relatively little motor or other manifestations. Impaired consciousness in absence seizures is not monolithic, but varies in severity between patients and even between episodes in the same patient. In addition, some aspects of consciousness may be more severely involved than other aspects. The mechanisms for this variability are not known. Here we review the literature on human absence seizures and discuss a hypothesis for why effects on consciousness may be variable. Based on behavioral studies, electrophysiology, and recent neuroimaging and molecular investigations, we propose absence seizures impair focal, not generalized brain functions. Imapired consciousness in absence seizures may be caused by focal disruption of information processing in specific corticothalamic networks, while other networks are spared. Deficits in selective and varying cognitive functions may lead to impairment in different aspects of consciousness. Further investigations of the relationship between behavior and altered network function in absence seizures may improve our understanding of both normal and impaired consciousness.
What is the relationship between brain activity and conscious thought? In science, difficult questions of this kind often yield to investigation, provided a good model system is available. One approach is to determine relationships between brain function and consciousness in situations where both vary. Examples include the investigation of sleep, anesthesia, brain lesions, evolution, development, and epilepsy. Epilepsy is an attractive model system for studying consciousness because epileptic seizures can cause selective, dynamic, and rapidly reversible changes in consciousness associated with altered function in specific brain networks. In this review, we will first introduce the major seizure types that cause impaired consciousness, including absence seizures. We next discuss electrophysiology and behavioral studies of impaired consciousness in absence seizures. Finally, we will briefly discuss recent neuroimaging and molecular studies suggesting that selective corticothalamic network involvement in absence seizures may, ultimately, explain the specific cognitive impairments that cause loss of consciousness.
Epileptic seizures are usually classified as either partial, involving focal brain regions, or generalized, involving widespread regions of the brain bilaterally (ILAE, 1981). Partial seizures associated with impaired consciousness are referred to as complex partial seizures, while those that spare consciousness are called simple partial seizures. Complex partial seizures arise most commonly from the temporal lobe. They typically begin with an unusual abdominal sensation or premonition, followed by staring, unresponsiveness, and simple repetitive movements of the mouth and limbs lasting 1–2 min, followed by amnesia for the episode. We and others have recently discussed a possible mechanism for impaired consciousness in temporal lobe complex partial seizures, in which network interactions with the upper brainstem and medial thalamus inhibit function of the bilateral fronto-parietal association cortex (Menzel et al., 1998; Lee et al., 2002; Norden and Blumenfeld, 2002; Blumenfeld and Taylor, 2003; Van Paesschen et al., 2003; Blumenfeld et al., 2004a, b).
Generalized seizures should invariably cause loss of consciousness if they truly involve the whole brain. Interestingly, however, this is not the case. Consciousness is spared in several types of generalized seizures (Gokygit and Caliskan, 1995; Bell et al., 1997; Vuilleumier et al., 2000). Moreover, recent work suggests that the so-called “generalized” seizures preferentially involve certain cortical–subcortical networks, while sparing others (Blumenfeld, 2003; McNally and Blumenfeld, 2004; McNally et al., 2004a, b; Nersesyan et al., 2004a, b). An important and common form of generalized seizure is the grand mal, or generalized tonic-clonic seizure. These typically consist of a dramatic rigid stiffening of the limbs for approximately 10 s, followed by synchronous contractions of the extremities for 1–2 min, with unresponsiveness throughout, and profound lethargy and confusion for a variable period after the episode ends. We recently proposed that the severe impairment in consciousness during and following generalized tonic-clonic seizures is caused by intense, abnormally increased neuronal activity in focal regions of the fronto-parietal association cortex and their related subcortical networks (Fig. 1) (Blumenfeld et al., 2003a, b; Blumenfeld and Taylor, 2003; McNally and Blumenfeld, 2004). Thus, abnormal decreased activity in bilateral fron-to-parietal networks may impair consciousness in complex partial seizures, while abnormal increased activity in these same networks may impair consciousness in generalized tonic-clonic seizures.
The most “pure” example of impaired consciousness in epilepsy is absence seizures. Like generalized tonic-clonic seizures, absence seizures (also called petit mal) are classified as generalized. However, in absence seizures the impaired consciousness is accompanied by relatively few additional motor or other complicating phenomena (like those seen in tonic-clonic and complex partial seizures).
Typical absence seizures consist of brief episodes of staring and unresponsiveness, often accompanied by mild eyelid fluttering or myoclonic jerks. Duration is usually less than 10 s, and no obvious deficits are present after the seizure concludes, apart from the “missed time” during the seizure. Absence seizures are most common in childhood, and can occur up to hundreds of times per day, leading to impaired school performance. Electroencephalogram (EEG) recordings during typical absence seizures reveal large amplitude bilateral 3–4 Hz spike-wave discharges (Fig. 2). In atypical forms of absence seizures, and in absence status epilepticus, consciousness may be preserved, and the EEG usually reveals a slower or more irregular spike-wave pattern. In this review we will focus on typical absence seizures, and on the impaired consciousness accompanying these episodes. It should be noted, however, that typical and atypical absence are not always clearly distinct categories, and in fact lie along the same diagnostic spectrum (Holmes et al., 1987).
Numerous human and animal studies have suggested that absence seizures are generated through abnormal network oscillations involving both the cortex and thalamus (Williams, 1953; Avoli et al., 1990; Blumenfeld and McCormick, 2000; Kostopoulos, 2001; McCormick and Contreras, 2001; Blumenfeld, 2002; Crunelli and Leresche, 2002; Blumenfeld 2003). Despite much work, it is still not known with certainty whether absence seizures are generated through an overall increase or decrease in neuronal activity in cortical and subcortical circuits (Engel et al., 1985; Salek-Haddadi et al., 2003; Aghakhani et al., 2003; Nersesyan et al., 2004a, b). In addition, even in typical absence seizures there is substantial variability in the degree of impaired consciousness from one patient to another, and even from one episode to the next in the same patient. The cause of this variability is not known.
To understand the fundamental mechanisms for impaired consciousness in absence seizures it will be necessary to obtain a more detailed understanding of both the physiological changes in the brain and the specific cognitive deficits that occur during these events. As we will discuss, prior studies have begun to investigate the anatomy and physiology of absence seizures, and altered behavior during these events, but little work has been done to directly relate impaired brain function to behavior. For example, can the variability in impaired consciousness during absence seizures be explained by variability in brain dysfunction during these episodes? Are there specific anatomical regions of the brain that are involved or spared when patients do or do not show impaired responsiveness during absence seizures? Is the impaired brain function caused by abnormally increased activity, abnormally decreased activity, or both? What are the molecular and circuit mechanisms causing specific regions but not others to be involved during seizures? We will now review prior work on the physiology and behavior of typical human absence seizures, and highlight areas that need to be addressed through further investigations. We mention atypical absence, and animal models only briefly, as these are reviewed in detail elsewhere (Holmes et al., 1987; Avoli et al., 1990; Yaqub, 1993; Crunelli and Leresche, 2002; Coenen and Van Luijtelaar, 2003).
The characteristic 3–4 Hz spike-wave discharge of absence seizures was first described by Gibbs and colleagues in 1935 (Gibbs et al., 1935), just 6 years after Berger (1929) developed human EEG. Since then, countless recordings have been made of these discharges, which are usually bilaterally symmetric, but with an occasional left or right amplitude preponderance (Ebersole and Pedley, 2003). Although considered a form of generalized epilepsy, it has long been appreciated that the amplitude of spike-wave discharges on scalp EEG is not uniform in all electrodes. In most cases, there is a clear bifrontal amplitude maximum, greatest in the midline (Figs. 2A and and3)3) (Weir, 1965; Rodin and Ancheta, 1987; Coppola, 1988; Holmes et al., 2004). The focal electrographic distribution of spike-wave discharge is also supported by recent studies in rodent models of absence seizures (Vergnes et al., 1990; Meeren et al., 2002; Nersesyan et al., 2004a, b). In agreement with human scalp EEG recordings, invasive recordings from rodent models show bilateral spike-wave discharges of maximal amplitude in the anterior regions, sparing the occipital cortex (Fig. 2B). Thalamic nuclei corresponding to the anterior cortical regions are intensely involved, while the thalamic lateral geniculate nuclei (with connections to the occipital cortex), are spared (Nersesyan et al., 2004b). Kostopoulos and others (Kostopoulos, 2001; Blumenfeld and Taylor, 2003) have proposed that focal involvement of specific thalamocortical sectors (e.g. in the frontal cortex), during absence seizures may cause selective deficits during absence seizures, affecting some cognitive modalities, and sparing others. To more fully investigate this hypothesis in humans, noninvasive methods with higher spatial resolution than scalp EEG, such as neuroimaging, will be needed. However, it is first necessary to understand in as much detail as possible the specific cognitive deficits caused by absence seizures.
There has been an interesting historical trend in behavioral studies of impaired consciousness during absence seizures (Fig. 4). These studies, initiated by Schwab and Jung in the late 1930s (Jung, 1939; Schwab, 1939), reached a peak in the 1960s, and then dropped off precipitously, with hardly any published investigations in the last two decades. The rise in behavioral studies of absence seizures can easily be explained by the development of EEG (Berger, 1929), improved quantitative testing methods, and increasing interest in experimental psychology. However, the cause for the decline is more elusive, especially given the explosion of interest in cognitive neuroscience and consciousness in recent years. One possible explanation is the development of effective anti-absence medications such as ethosuximide in 1960 and valproate in 1978, leading to a smaller patient pool with uncontrolled episodes to study. However, studies should still be possible using patients with newly diagnosed absence, or patients who are unable to tolerate therapeutic doses of these medications. Another, more intriguing possibility is that studies on impaired consciousness in absence stopped because most questions were answered within the technical limitations of the times. In other words, once the major behavioral features of absence seizures were described, and their relationship to EEG discharges measured, there was little novel work left to be done until better methods for studying brain function were developed. As we discuss in the next section, it is our hope that this “forgotten science” will be reinvigorated now that modern functional neuroimaging and other methods are available.
Meanwhile, there remains much to be learned from the first wave of behavioral studies of absence seizures. These investigations utilized a variety of techniques to assess patients’ cognitive function during absence seizures. Because the literature in this field is finite, an effort has been made to include all or nearly all pertinent studies in this review. Some of the studies have significant methodological limitations, including heterogeneous patient types, limited sample size, and descriptive rather than quantitative analysis. Nevertheless, the majority provide useful data, allowing some preliminary conclusions to be drawn, which may spark future research. We will now review three common themes in these studies: (1) Which cognitive functions are involved and which preserved during absence seizures? (2) What is the timing of deficits relative to spike-wave discharges on EEG? (3) Is the variability in impaired consciousness from one episode to the next related to any properties of the EEG?
A variety of different testing paradigms have been used to study impaired consciousness during absence seizures (Table 1). In fact, the different tests used illustrate that several cognitive functions may contribute to consciousness. There has been considerable dispute in the absence literature, whether impairment of some cognitive functions with sparing of others, qualifies as impaired consciousness. Does a patient who appears normal and is able to talk, but has transient slowing of reaction times during spike-wave episodes, have impaired consciousness? On the other extreme, can a patient who is completely unresponsive and has no recall afterwards of episodes be considered unconscious, or is the patient simply immobile, mute, and amnestic due to selective motor, verbal, and memory impairment (Gloor, 1986)? Ultimately, to answer questions of this kind, a more specific, mechanistic definition of consciousness is needed, along with a better understanding of the underlying physiological processes. Until this is achieved, and perhaps as a means of achieving such a definition, it is useful to study each of the specific cognitive processes that may contribute to loss of consciousness.
Testing paradigms during absence seizures have varied from simple repetitive motor tasks to higher order decision making (Table 1). Some functions are more severely impaired than others during absence seizures. The tests in Table 1 are listed in approximate order of impairment by absence seizures. For example, several investigators reported that simple repetitive motor tasks such as finger tapping are more often conserved than more demanding tasks, such as counting aloud, and impairment is even more common with tasks requiring a response to verbal questions (Schwab, 1939; Cornil et al., 1951; Courtois et al., 1953; Gastaut, 1954; Goldie and Green, 1961; Davidoff and Johnson, 1964). Decision making based on sensory input, and complex motor performance were more severely affected than simpler tests (Guey et al., 1965; Mirsky and Buren, 1965). Furthermore, impaired memory of items presented during seizures depended on task difficulty (Jus and Jus, 1962; Mirsky and Buren, 1965; Geller and Geller, 1970). Interestingly, while verbal responses were very often interrupted during seizures, some patients recalled questions or commands given during a seizure, and responded appropriately after the seizure ended (Courtois et al., 1953; Kooi and Hovey, 1957; Boudin et al., 1958). In addition, while some patients were unable to respond verbally, they were able to turn their gaze toward the examiner (Goldie and Green, 1961). Some investigators also found that absence seizures could be terminated by external stimulation, with loud or painful stimuli being most effective (Jung, 1939, 1954; Boudin et al., 1958), and seizure frequency was diminished by tasks requiring active mental attention (Lennox et al., 1936; Gastaut, 1954; Kooi and Hovey, 1957; Davidoff and Johnson, 1964; Bureau et al., 1968). Another interesting observation was that even “subclinical” (also called “larval” or “phantom”) spike-wave discharges, with no obvious clinical impairment, were often associated with more subtle deficits in reaction time or continuous motor performance (Schwab, 1939; Kooi and Hovey, 1957; Yeager and Guerrant, 1957; Tuvo, 1958; Grisell et al., 1964; Guey et al., 1965).
In summary, these findings suggest that absence seizures can independently affect some cognitive functions that may be important for consciousness, while sparing certain others. The mechanism for this remains unknown, but it can be hypothesized that selective cognitive deficits may be related to selective impairment of specific brain networks during seizures, while others are spared.
On the basis of verbal response to questions and motor tests, it was observed that patients tend to have little impairment in the initial 1–3 s and final 3–5 s of absence seizures, but enter a “trough” of impaired consciousness in between (Shimazono et al., 1953; Goldie and Green, 1961; Mirsky and Buren, 1965; Goode et al., 1970). On the other hand, some investigators reported on the basis of vigilance or reaction time tasks that impairment coincides with electrographic seizure onset (Browne et al., 1974), or may even occur a few seconds before the onsets (Mirsky and Buren, 1965). In addition, although impairment of vigilance (continuous performance task) was reported to end about 5 s before the end of seizures, impaired simple tapping and sometimes responses to verbal questions have been reported to be impaired for several seconds after the end of the electrographic seizure (Goldie and Green, 1961; Mirsky and Buren, 1965). The different timing effects in these studies may again reflect the different tasks and methods used. Thus, some cognitive functions may be more severely affected during certain phases of seizures, while others are left relatively unharmed. This, once again, supports the notion that absence seizures dynamically disrupt specific functional brain networks, while sparing others.
There has also been some variability in reports of the timing of amnesia during absence seizures. Thus, some investigators reported no retrograde amnesia for stimuli presented prior to seizure onset (Courtois et al., 1953), while others reported retrograde amnesia extending up to 14 s before seizure onset, with gradual shrinking of the retrograde amnesia after the seizure ended (Jus and Jus, 1962; Mirsky and Buren, 1965; Geller and Geller, 1970).
One important common theme in all these studies is that impaired responsiveness is not a monolithic feature of absence seizures, but rather highly variable between patients, and even from one episode to the next within a single patient. The mechanism for this variability is of great potential importance for understanding the fundamental mechanisms of impaired consciousness in absence seizures. Efforts have been made to relate certain EEG features to the degree of impairment. This subject has also been a matter of some controversy among investigators, most likely because of varying testing methods and cognitive tasks that were employed.
Longer seizure durations, especially those lasting more than 3–4 s are reported to cause more severe impairments in simple reaction time, continuous performance task, memory testing, and continuous motor performance (Schwab, 1939; Guey et al., 1965; Mirsky and Buren, 1965; Goode et al., 1970). Some impairment can be seen on careful reaction time testing even with brief spike-wave discharges (Browne et al., 1974). However, the impairment in this case is more obvious and lasts longer with longer duration discharges. Amplitude (Courtois et al., 1953) and “generalization” of discharges were associated with worse performance on reaction time and other tasks (Sellden, 1971; Porter and Penry, 1973; Browne et al., 1974). Mirsky and Van Buren conducted a detailed investigation of EEG characteristics, and found that spike-wave duration, amplitude, rhythmicity, and fronto-central distribution were all associated with more severe impaired performance on vigilance and memory testing (Mirsky and Buren, 1965). A recent small case series suggested different EEG spike-wave distributions during impaired or preserved consciousness (Vuilleumier et al., 2000). Other investigators, however, found no association between EEG duration (Browne et al., 1974), or any other EEG features with impaired consciousness (Gastaut, 1954).
The majority of these studies suggest that an association exists between certain EEG features, and impaired performance on cognitive tasks contributing to impaired consciousness in absence. However, the precise anatomical and physiological source of this variability in EEG and behavioral impairment will require further investigation with more advanced methods.
Human imaging studies during absence seizures have been highly controversial. Some studies have shown global increases in cerebral metabolism or blood flow (Engel et al., 1982, 1985; Theodore et al., 1985; Prevett et al., 1995; Diehl et al., 1998; Yeni et al., 2000), while others have shown no change, focal or generalized increases or decreases (Theodore et al., 1985; Ochs et al., 1987; Park et al., 1994; Ferrie et al., 1996; Diehl et al., 1998; Salek-Haddadi et al., 2003; Aghakhani et al., 2003; Archer et al., 2003). Some of these variabilities may reflect technical limitations, such as limited temporal resolution of fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET) or Tc99 m single photon emission computed tomography (SPECT) imaging relative to absence seizure duration. In addition, the variability may result from heterogeneous patient populations, and the inclusion in these studies of pediatric and adult patients along with variable numbers of patients with atypical absence seizures.
Because absence episodes are relatively brief, it is crucial to simultaneously record EEG with neuroimaging so that images acquired during seizures and baseline can be identified and analyzed separately. In addition, an imaging modality with sufficient temporal resolution to capture individual absence seizures should be used. Functional magnetic resonance imaging (fMRI) fulfils both of these criteria, and several recent studies have begun to show interesting results with this method (Aghakhani et al., 2003; Archer et al., 2003; Salek-Haddadi et al., 2003). Interestingly, fMRI changes during absence seizures include both increases and decreases in signal, and these changes are heterogeneous, not involving the entire brain in a uniform fashion (Fig. 5). These findings agree with behavioral and EEG studies suggesting that absence seizures may selectively involve certain networks, while sparing others. Additional neuroimaging studies with improved spatial and temporal resolution will be needed to investigate the possible role of focal brain dysfunction in loss of consciousness during absence seizures. In addition, fMRI studies are limited because the signals measured are only indirectly related to neuronal activity. Further work will be needed to fully interpret the relationship of fMRI signal increases and decreases during absence seizures with the underlying changes in neuronal activity (Smith et al., 2002; Nersesyan et al., 2004b).
Studies in animal models may also shed light on mechanisms of spatial heterogeneity in spike-wave seizures (Blumenfeld, 2003). Recent studies on rodent models on absence seizures have demonstrated that, as in humans, seizure discharges often have an anterior predominance, with the posterior brain regions relatively spared (Vergnes et al., 1990; Meeren et al., 2002; Klein et al., 2004; Nersesyan et al., 2004a, b) (Fig. 2B). These investigations have included electrophysiologic, fMRI, and Doppler flow measurements, all implying that spike-wave seizures that appear fairly generalized on scalp EEG recordings, in fact, may intensely involve some corticothalamic networks, while sparing others.
Ultimately, it will be crucial to identify the underlying mechanisms that cause certain corticothalamic networks to be intensely involved in spike-wave seizures, while others remain functional. Many factors can contribute to altered excitability and synchronous firing of corticothalamic networks, and only a few candidate molecules and mechanisms have begun to be explored (Crunelli and Leresche, 2002). It was recently found in a rodent model of absence seizures that enhanced excitability in focal cortical regions may be related to altered expression of voltage gated sodium channels (Klein et al., 2004) or to reduced function of hyperpolarization-activated cation channels (Ih) (Strauss et al., 2004). If the specific genes and proteins can be identified, which generate absence seizures in selective brain regions, but not others, it may ultimately be possible to map networks predisposed to seizure discharges using molecular techniques. Furthermore, identification of these molecular changes will allow targeted therapies aimed at reducing abnormal excitability in these networks, and to prevent absence seizures and the accompanying loss of consciousness.
Loss of consciousness during absence seizures may serve as a good model system for studying normal mechanisms of consciousness. Impairments in consciousness during absence seizures are dynamic and rapidly reversible, and are relatively selective, without many of the motor or other manifestations seen in other seizure types. Several converging lines of evidence suggest that impaired consciousness during absence seizures is not a global phenomenon, but rather can be decomposed into specific deficits in a number of different cognitive functions, associated with impaired function in selective neuroanatomical networks. On the basis of the studies reviewed here, we hypothesize that loss of consciousness in absence seizures is not an “all-or-none” phenomenon resulting from involvement of the entire brain in the seizure discharge. Rather, focal involvement of the bilateral association cortex, perhaps most prominently in frontal neocortex and related subcortical structures, disrupts normal information processing in these brain regions leading to impairment of specific cognitive functions necessary for normal consciousness. The detailed spatial distribution of these changes, the physiological effects of spike-wave discharges on normal neuronal signaling, the molecular and cellular mechanisms underlying selective network involvement, and the relationship of these changes to behavior will need additional investigation.
I gratefully thank Dr. Alain Hyman, Dr. David L. Blumenfeld, and Dr. Ulrich Schridde for translating of papers from French and German literature. I also thank George Varghese for assistance in preparing the figures. This work was supported by NIH R01 NS049307 and by the Betsy and Jonathan Blattmachr family.