OUP user menu

The localizing value of ictal consciousness and its constituent functions
A video‐EEG study in patients with focal epilepsy

S. Lux , M. Kurthen , C. Helmstaedter , W. Hartje , M. Reuber , C. E. Elger
DOI: http://dx.doi.org/10.1093/brain/awf276 2691-2698 First published online: 1 December 2002


Using ictal neuropsychological testing in pre‐surgical patients with focal epilepsies, we examined the localizing value of the constituent functions of consciousness as opposed to ‘conscious behaviour’ as a unitary variable. ‘Conscious behaviour’ was defined in terms of awareness and responsiveness. The constituent functions of consciousness examined included the orientation to the examiner, intentional behaviour demonstrated by expressive or receptive speech, and postictal memory. Frequency and patterns of impairment of constituent functions and ‘conscious behaviour’ were assessed. To achieve this, pre‐surgical video‐EEG (n = 40) or video‐electrocorticography recordings (n = 76) of ictal neuropsychological assessments were reviewed retrospectively. Patients were divided into groups with frontal (n = 29), right temporal (n = 21), left temporal (n = 38) and bitemporal (n = 28) seizure activity. Consciousness was most commonly impaired in patients with bitemporal and left temporal seizure activity. There were different patterns of impairment of the assessed constituent functions in the four groups: patients with frontal seizure activity showed loss of orientation behaviour and expressive speech whereas patients with left temporal seizure activity had impairments of memory, expressive and receptive speech. Patients with seizure activity limited to the right temporal lobe rarely exhibited ictal impairment of any of the assessed functions. In contrast, patients with bitemporal seizure activity showed impairment of all examined functions. Hence, normal functioning of the left temporal lobe or both temporal lobes is necessary for the preservation of all constituent aspects of consciousness. The localizing value of patterns of impairment of constituent functions is superior to that of ‘consciousness’ as a whole.

  • Keywords: consciousness; video analysis; EEG; ictal neuropsychological testing; epilepsy
  • Abbreviations: ECoG = electrocorticography; MS‐CFA = multiple‐sample configuration‐frequency analysis; NCC = neural correlate of consciousness


The term ‘consciousness’ is difficult to define because it has several different meanings ranging from philosophical and psychological to biological concepts. In neuroscience, it is generally accepted that consciousness is dependent on brain functions. Hence, neuroscientific research has attempted to identify the neural correlate of consciousness (NCC) (Crick and Koch, 1990). Neuropsychological and physiological studies suggest that the NCC is neither an emergent property of the brain as a whole nor a function of a single consciousness ‘centre’ (Delacour, 1997). Rather, the NCC consists of different interacting neural subsystems. It is conceivable that some of these subsystems are indispensable for the maintenance of consciousness while others are not. In humans, this assumption can be assessed by examining the NCC in patients with topologically different focal cerebral dysfunctions. In particular, focal epileptic seizures are known to manifest themselves with more or less localization‐specific behavioural signs. Seizures can serve as a model of short‐term, circumscribed functional impairment of certain brain regions. Using epileptic activity as a model means that the time course and topographical distribution of the neurophysiological correlates of the impairments can be easily monitored by EEG or electrocorticography (ECoG). Ictal neuropsychological testing in video‐EEG‐documented focal seizures of different origins may therefore help to determine the NCC in humans (Gloor, 1986). But what are the behavioural signs of consciousness? Frith and colleagues suggested that a patient’s verbal or behavioural report of perceptions, intentions and memories is the most direct way to assess consciousness (Frith et al., 1999). One way of assessing patients’ perception is to examine their orientation behaviour—their immediate, directed motor and vegetative response to sensory stimulation (visual, auditory or tactile). Patients can demonstrate intentional behaviour by reacting adequately to verbal requests (‘receptive speech’). A further function, namely the ability to adequately utter words or sentences (‘expressive speech’), provides evidence of ongoing perception and intention. These three variables offer proof of a patients’ perception and intention, but the most important and most direct sign of ictal awareness is the patients’ ability postictally to recall what happened around them during a seizure. Maintenance of consciousness can be assumed if postictal memory is intact and/or orientation behaviour, expressive and receptive speech are preserved. This definition of consciousness is in keeping with that given by the International League Against Epilepsy (ILAE) (Commission on Classification and Terminology of the International League Against Epilepsy, 1981). In the current ILAE seizure classification, impaired consciousness is used as a behavioural variable to distinguish between complex focal (with impaired consciousness) and simple focal (with maintained consciousness) seizures. Operationally, the ILAE links consciousness to the degree of awareness and/or responsiveness. While responsiveness is understood as the ability to carry out simple commands or willed movements (intention and perception), awareness is defined as the patients’ contact with the event and its recall (memory).

In the present study we considered ‘consciousness’ to be unimpaired if memory and/or the three functions of orientation behaviour, receptive speech and expressive speech were unimpaired (see Table 1). We examined patients with focal epilepsy undergoing pre‐surgical evaluation to assess the localizing value of ‘consciousness’ compared with its four constituent functions (as analysed separately). As most focal epileptic seizures originate from the temporal (70%) or frontal (20%) lobes, we focused on patients with bitemporal, left temporal, right temporal or frontal seizure activity as determined by surface EEG or ECoG. Patients underwent interictal neuropsychological testing of the constituent functions of consciousness under video‐EEG observation. Video‐EEG tapes were then analysed retrospectively.

View this table:
Table 1

Definition of the examined functions

ConsciousnessMemory and/or adequate verbal and non‐verbal behaviour
Orientation behaviourPatient is able to turn towards the examiner on sensory stimulation.
Expressive speechPatient can speak words or sentences without dysphasic errors.
Receptive speechAdequate verbal or non‐verbal reaction on verbal requests
MemoryPatient can remember details of the seizure or elements of the examination postictally.

Before analysing the videotapes, we generated hypotheses about ictal loss of consciousness on the basis of previously reported evidence (see Methods). We found that ictal impairment of consciousness has rarely been studied in detail. Those investigators who did focus on this issue have suggested that consciousness is most frequently impaired in patients with bitemporal seizure activity (Gloor et al., 1980; Munari et al., 1980; Bancaud et al., 1994; Inoue and Mihara, 1998). Expressive and receptive speech have been found to be impaired in patients with left temporal seizure activity as long as left hemisphere speech dominance could be assumed (Serafetinides and Falconer, 1963; McKeever et al., 1983; Gabr et al., 1989). Ictal impairment of memory has been described in association with bitemporal seizure activity (Schulz et al., 1995). Impairment of orientation behaviour during focal epileptic seizures has not been studied; we therefore did not know what to expect with respect to this variable. Furthermore, no study published to date has examined ictal performance in two or more of these aspects of consciousness in terms of a performance profile.



Video‐EEG recordings of 469 seizures of consecutive patients with medically intractable focal epilepsies were considered for inclusion. All recordings were made whilst patients were undergoing evaluation for epilepsy surgery at the Department of Epileptology, University of Bonn, Germany, between August 1990 and January 1998. All patients gave written informed consent to the examinations that were retrospectively reviewed for the purpose of the present study. Only patients in whom strictly ictal, video‐documented neuropsychological bedside testing had been completed during at least one focal seizure were included in the study. Focal seizures were defined as episodes with focal ictal epileptiform changes in the EEG. ‘Subclinical’ seizures, i.e. seizures without subjective experiential phenomena, were included. Neuropsychological bedside testing was performed by trained nurses, electroencephalographers, physicians or neuropsychologists. Testing started with an attempt to elicit orientation behaviour. If found to be preserved, this was followed by testing of receptive and expressive speech. Finally, memory for the seizure and testing procedure was assessed. For further details of the testing procedure, see Table 2.

View this table:
Table 2

Overview of ictal bedside neuropsychological test procedure

Orientation behaviour (to acoustic, visual or tactile stimulation)‘What is your name?’ Wave hand in front of patient’s eyes. Touch or pinch patient. ‘Raise your right hand.’
Receptive speech (with non‐verbal or verbal response)‘Tell me your name.’
Expressive speech‘What is this?’ (pen, pictures of common objects)
Postictal memory‘What happened just then?’
‘Can you remember what I just asked you?’

*The questions and tasks listed are typical examples; the procedure was not standardized throughout the study period.

Focal seizures with secondary tonic clonic generalization were excluded from the study. Furthermore, seizures were excluded if EEG or ECoG recordings were uninterpretable during the seizure or if the localization of ictal discharges (frontal, right temporal/left temporal/bitemporal) could not be determined. Seizures were also excluded if it was not clear that ictal epileptic activity had continued throughout the testing procedure. Finally, we excluded patients with evidence of right hemispheric speech dominance during intracarotid amobarbital testing (Wada test). If several seizures had been recorded in one patient, only the seizure with the most comprehensive bedside testing was assessed.

Seizures of 116 patients met all these criteria. All patients were >12 years old. EEG was sampled from the scalp (n = 40) or from chronically implanted intracranial electrodes (n = 76). In all patients except one, a detailed interictal neuropsychological examination was performed. Five patients showed low average performances in all examined subfunctions. In all except four patients, MRI was performed during presurgical evaluation. MRI showed left temporal lesions in 35 patients, right temporal lesions in 29 patients, bitemporal lesions in four patients and frontal lesions in 19 patients. Twenty patients had no lesions and five had non‐specific changes. Histopathological diagnoses in the 96 patients who underwent surgery were hippocampal sclerosis or hippocampal atrophy (43 patients), cavernoma (7), ganglioglioma (6), postsurgical scarring (5), vascular lesion (4), tumour (3), heterotopia (2), hamartia (1) and non‐specific findings (25).

Assessment of ictal test performance and EEG

An observation inventory containing written instructions for all five examined variables (Table 1) was developed for this study. Prior to the videotape analysis, four psychologists watched a tape showing 11 typical examples of seizures. All functions except memory were examined in all 11 seizures; memory was only tested six times. Memory was disrupted in one, orientation behaviour in five, expressive speech and consciousness in six and receptive speech in seven seizures. Ictal loss of aspects of consciousness was initially rated by one of the authors (S.L.). The other raters were blinded to the initial assessment or the results of other raters. Inter‐observer variation (agreement between different observers) was measured using kappa statistics. One of the authors (S.L.) analysed the video‐EEG documents on this tape for a second time after an interval of six weeks to obtain a measure of intra‐observer variation (agreement between two observations made by the same observer). This was also assessed by calculating a kappa value. This index specifies the likelihood that inter‐ and intra‐observer agreement is different from chance. A kappa value >0.75 denotes excellent agreement beyond chance, a value between 0.4 and 0.75 represents fair to good agreement beyond chance, and a value <0.4 indicates poor agreement.

Another one of the authors (M.K.) reviewed all EEG recordings to determine the spatial distribution of ictal epileptiform discharges at the time of neuropsychological bedside testing. According to this evaluation, the patients were divided into four groups: patients with frontal, left temporal, right temporal and bitemporal seizure activity. The lateralization of frontal seizures was not taken into account because a considerable proportion of frontal seizures lacked clear lateralization (particularly in surface recordings).

The frequency of impairment of the assessed functions in the four patient groups was compared using χ2‐tests or Fisher exact tests. In addition, a multiple‐sample configuration‐frequency analysis (MS‐CFA) (Krauth and Lienert, 1995) with adjustment of alpha (Holm, 1979) was performed on the four patient groups for the variables orientation behaviour, expressive speech and receptive speech. Combinations of characteristic features with a frequency beyond chance were determined for each sample via MS‐CFA.


Objectivity of the observer inventory

Kappa values of inter‐ and intra‐observer agreement for the five examined variables were as follows: 0.4 and 0.5 for orientation behaviour; 0.7 and 1 for expressive speech; 0.7 and 0.7 for receptive speech; 0.7 and 0.8 for postictal memory; and 1 and 0.9 for ‘consciousness’. Thus, inter‐ and intra‐observer agreement was excellent or far beyond chance for all variables, and the observation inventory demonstrated a sufficient degree of objectivity.

Localization of seizure activity

After all the EEG recordings had been reviewed, patients were divided into four groups with right, left and bitemporal or frontal seizure activity. The number of patients in each group, sex distribution, mean and standard deviation of age, and the number of patients investigated with implanted electrodes are given in Table 3.

View this table:
Table 3

Distribution of sex, age and kind of electrodes

Seizure activity
FrontalRight temporalLeft temporalBitemporal
n 29213828
Sex (M/F)9/2023/1513/814/14
Age (years)*26.7 ± 8.936.8 ± 10.333.4 ± 9.130.3 ± 8.4
Subdural and/or depth electrodes (n)9182524

*Given as mean ± SD.

Localizing values

The following sections contain the results of the comparison of the frequencies of impairment in the patient subgroups and the MS‐CFA. The MS‐CFA was performed on 97 patients. The results of this analysis are summarized in Table 4.

View this table:
Table 4

Results of the multiple‐sample configuration‐frequency analysis

FunctionsLocation of seizure activity
Orientation behaviourExpressive speechReceptive speechFrontal (n = 25)Left temporal (n = 33)Right temporal (n = 18)Bitemporal (n = 21)

– = impaired; + = unimpaired; *P < 0.05; ** P < 0.01.

Impairment of ‘consciousness’ within the four patient groups

‘Consciousness’—defined as preservation of memory and/or orientation behaviour, expressive and receptive speech—was assessed in a total of 104 patients. As shown in Fig. 1, ‘consciousness’ was more frequently impaired in patients with bitemporal seizure activity than in all other patient groups [right temporal versus bitemporal: F(n = 21), P < 0.01; left temporal versus bitemporal: F(n = 27), P < 0.05; frontal versus bitemporal: F(n = 25), P < 0.05]). In addition, nine out of 23 patients (39%) with left temporal seizures showed impaired ‘consciousness’. There was a tendency for this group to be more frequently impaired than patients with right temporal seizure activity [F(n = 40), P = 0.06].

Fig. 1 Frequency of impairment of ‘consciousness’.

Patients with right temporal seizure activity

These patients rarely showed impairment of any of the examined functions. Results of the MS‐CFA showed that absence of impairment of all three functions was more frequent in this than in any other patient group.

Patients with left temporal seizure activity

Expressive speech, receptive speech and memory were more often impaired than in patients with right temporal seizure activity [χ2 (1, n = 55) = 7.41, P < 0.01; χ2 (1, n = 54) = 9.9, P < 0.01; F(n = 18, P = 0.09]. These patients were not more frequently impaired with respect to orientation behaviour than any other group. According to the results of the MS‐CFA, co‐occurrence of unimpaired orientation behaviour and impaired speech functions was more frequently found here than in any other patient group.

Patients with bitemporal seizure activity

This patient group was more frequently impaired in all examined functions than patients with right seizure activity [orientation behaviour: χ2 (1, n = 49) = 11.7, P < 0.01; expressive speech: χ2 (1, n = 40) = 17.4, P < 0.01; receptive speech: χ2 (1, n = 48) = 29.6, P < 0.01; memory: (F(n = 9, P < 0.01)] or left [orientation behaviour: χ2 (1, n = 66) = 10.6, P < 0.01; expressive speech: χ2 (1, n = 57) = 4.6, P < 0.05; receptive speech: χ2 (1, n = 62) = 9.28, P < 0.01; memory: F(n = 17), P = 0.09)]. Furthermore, receptive speech and memory were more frequently impaired than in patients with frontal seizure activity [χ2 (1, n = 55) = 13.3, P < 0.01); F(n = 20, P < 0.05)].

The MS‐CFA showed that impairment of all examined functions was found more commonly in this group than in any other patient groups.

Patients with frontal seizure activity

Orientation behaviour and expressive speech were more frequently impaired in patients with frontal than in those with right or left temporal seizure activity [orientation behaviour: χ2 (1, n = 49) = 11.7, P < 0.01; χ2 (1, n = 66) = 10.6, P < 0.01; expressive speech: χ2 (1, n = 40) = 17.4, P < 0.01; χ2 (1, n = 57) = 4.6, P < 0.05]. Receptive speech was more frequently impaired in this group than in patients with right temporal seizure activity [χ2 (1, n = 62) = 5.5, P < 0.05]. There was no significant increase of memory impairment compared with the other patient groups.

According to the results of the MS‐CFA, the combination of impaired orientation behaviour, impaired expressive speech and unimpaired receptive speech was found more frequently in this group than in any other.


There are controversial theories about the role different brain regions play in the maintenance of consciousness. Inspired by split‐brain research, Eccles (1965) speculated that everything that is truly human emanates from the left hemisphere. Gazzaniga (1998) appeared to concur. He held that the right hemisphere’s level of awareness is limited, while the left hemisphere houses most higher mental capacities. He believed that the ‘left brain interpreter’ has the capacity to unify conscious experience by providing post hoc rationalization for ongoing behaviour. In contrast, Sperry (1966) thought that split‐brain patients have two independent conscious units. The specificity of the contribution of each hemisphere to consciousness is still unclear (Delacour, 1997).

We approached the question of the cerebral localization of consciousness by a different route: consciousness and its constituent functions were assessed with respect to their localizing value in well characterized epileptic seizures of focal origin. We expected most frequent impairments in patients with bitemporal seizure activity. The few previous studies of consciousness during temporal epileptic seizures found consciousness to be impaired in up to 81% of patients with bilateral seizure activity (Gloor et al., 1980; Munari et al., 1980; Bancaud et al., 1994; Inoue and Mihara, 1998). Our expectations in this respect were confirmed, and we were able to replicate previous results by showing that all of our patients with bitemporal seizure activity displayed impairment of consciousness.

However, bilateral seizure spread has not been thought to be obligatory for an impairment of consciousness. Previous investigators also described impaired consciousness in up to 46% of patients with unilateral temporal seizure activity. Inoue and Mihara (1998) thought that this occurred more frequently with left‐sided rather than right‐sided seizure activity. Our results are in keeping with this observation. While behaviour in 39% of our patients with left temporal seizure activity suggested impairment of consciousness, this was found in only 12% of our patients with right temporal seizure activity. Inoue and Mihara (1998) described loss of consciousness in 19% of their patients with frontal seizure activity. In the present study, 32% of these patients were impaired and, in this respect, we found no significant difference between patients with frontal seizures and those with right temporal seizure activity.

Taken together, our results suggest that the functioning of the left or both temporal lobes together is very important for the maintenance of consciousness, whereas the functioning of the right temporal lobe itself seems to be of minor relevance. At first glance, our results seem to support the assumption of Eccles (1965) and Gazzaniga (1998) that the right hemisphere contribution to consciousness is small. However, we can only really say that the contribution of the right temporal lobe to ‘consciousness’ as defined for the purpose of this study is weak. One may wonder whether typically right temporal functions were included in our definition of consciousness at all. Remember that some early critics of the left‐hemisphere hypothesis of consciousness have argued that ‘consciousness’ is usually operationalized in a way that clearly favours left‐hemispheric behavioural markers like verbal reports (Benson and Zaidel, 1985). Although this criticism may also apply to some of the behavioural markers or ‘constituent functions’ examined in the present study (namely receptive and expressive speech, and perhaps even postictal memory), orientation behaviour should not depend on the ictal integrity of verbal functions.

In terms of the localizing value of the four hypothesized constituent functions as opposed to global ‘conscious behaviour’, we expected patients with left temporal seizure activity to be frequently impaired in expressive and receptive speech. Many studies have looked at speech functions in patients with unitemporal seizure activity (Hécaen and Piercy, 1956; Bingley, 1958; Serafetinides and Falconer, 1963; Currie et al., 1971; Lecours and Joanette, 1980; McKeever et al., 1983; Theodore et al., 1983; Koerner and Laxer, 1988; Gabr et al., 1989; Kanemoto and Janz, 1989; Morrell et al., 1991; Chee et al., 1993; Fakhoury et al., 1994; Yen et al., 1996; Dantas et al., 1998; Marks and Laxer, 1998; Serles et al., 1998; Steinhoff et al., 1998; Williamson et al., 1998). Only three of these investigations distinguished between receptive and expressive speech (Serafetinides and Falconer, 1963; McKeever et al., 1983; Gabr et al., 1989). The studies found that all or most of the patients with left temporal seizure activity had impairment of both speech functions. Our findings concur with these results. We found patients with left temporal seizure activity to be more frequently impaired in receptive and expressive speech than patients with right temporal seizure activity. But in addition, we found that postictal memory was also more frequently impaired in patients with left temporal seizure activity than in patients with right temporal seizure activity. This might be explained by our predominantly verbal memory assessment. Since the seminal work of Milner (1954), there has been an increasing body of evidence suggesting that verbal memory functions are represented in the left hemisphere, while figural memory functions are preferentially supported by the right hemisphere or bilaterally.

We expected memory as a postictal marker of awareness to be more frequently impaired in patients with bitemporal seizure activity. Schulz and colleagues found that memory and encoding of the ictal events was impaired during bitemporal seizure activity (Schulz et al., 1995). Further indication of the need of bitemporal integrity for memory is provided by the syndrome of ictal amnesia. This is a purely amnesic syndrome lasting for some minutes and combining impairment of encoding with retrograde amnesia. It has been described most frequently in patients with bitemporal spread of ictal epileptic activity (Lou, 1968; Deisenhammer, 1981; Dugan et al., 1981; Meador et al., 1985; Pritchard et al., 1985; Galassi et al., 1986, 1988; Miller et al., 1987; Kapur et al., 1989; Stracciari et al., 1990; Palmini et al., 1992; Kapur, 1993; Zeman et al., 1998). Our results in this respect matched our expectation. We also found memory to be impaired more frequently in patients with bitemporal seizure activity than in all other patient groups. In addition, however, more patients with bitemporal seizure activity were impaired in terms of orientation behaviour, expressive and receptive speech than patients with unilateral temporal seizure activity. Typically, the bitemporal group therefore had impairment of all investigated functions.

In terms of expressive speech and orientation behaviour, bitemporal seizures caused the same impairment as frontal seizure activity. The impairment of expressive speech in patients with frontal seizure activity may be a little surprising at first sight, but it is in keeping with the results of studies of ictal speech arrest. This phenomenon (which may be comparable to the impairment of expressive speech described in this study) has been described most commonly in patients with frontal seizure activity (Tharp, 1972; Ludwig et al., 1975; Shield et al., 1977; Cascino et al., 1991; Chee et al., 1997; Sakai et al., 1997). In our investigation, patients with frontal seizure activity were set apart from the other groups by the preservation of receptive speech and memory in the presence of impairments of orientation behaviour and expressive speech.

As mentioned before, patients with right temporal seizure activity were distinguishable from the other patient groups because they typically showed preservation of all examined aspects of consciousness. This, however, is only a negative sign of right temporal seizure activity. We assessed no function that we would have expected to be impaired during right seizure activity. To aid the localization of seizure origin, such a function should be included in the examination programme. The assessment of memory for faces or figures would be a suitable test for this.

In summary, we have found that patients with frontal lobe, right, left or bitemporal lobe epilepsy can be distinguished by the pattern of impairment of four constituent functions of consciousness. The good differentiation of the four patient groups by constituent functions contrasts with the poor localizing value of global ‘conscious behaviour’. This function only allowed the distinction between patients with left or bitemporal seizure activity (impaired ‘conscious behaviour’) and patients with right or frontal activity (unimpaired ‘conscious behaviour’). This raises the question whether ‘conscious behaviour’ should be used as a variable in the context of seizure categorization. Our results illustrate the limitations of a subclassification of focal epileptic seizures into ‘complex’ and ‘simple’, and support the view of some epileptologists that the rather arbitrary criterion of consciousness should be abandoned (Lüders et al., 1999).

In conclusion, normal function of left or both temporal regions together seems to be important for the preservation of consciousness. Ictal involvement of the right temporal lobe alone has no adverse effects on the maintenance of consciousness as defined for the purpose of this study. The localizing value of global impairment of ‘conscious behaviour’ as an ictal sign is poor, while the patterns of impairment of the constituent functions of consciousness are useful in the localization of focal seizure activity. These results are in accordance with recent advances in neurobiological consciousness research (Crick and Koch, 1998)—namely the subdivision of the fuzzy concept of ‘consciousness’ into well‐defined functions, which can be assessed separately and subjected to empirical investigation. It should be pointed out that, like our study, this body of work depends on operational definitions of aspects of consciousness and has little to do with the more colloquial, subjective, understanding of consciousness as the ‘occurrence of experience’.


View Abstract