Testing for minimal consciousness in complex partial and generalized tonic–clonic seizures


Address correspondence to Hal Blumenfeld, Yale Departments of Neurology, Neurobiology, Neurosurgery, 333 Cedar Street, New Haven, CT 06520-8018, U.S.A. E-mail: hal.blumenfeld@yale.edu


Impaired consciousness in epilepsy has a major negative impact on quality of life. Prior work suggests that complex partial seizures (CPS) and generalized tonic–clonic seizures (GTCS), which both cause loss of consciousness, affect similar frontoparietal networks. Milder involvement in CPS than in GTCS may spare some simple behavioral responses, resembling the minimally conscious state. However, this difference in responses has not been rigorously tested previously. During video–electroencephalography (EEG) monitoring, we administered a standardized prospective testing battery including responses to questions and commands, as well as tests for reaching/grasping a ball and visual tracking in 27 CPS (in 14 patients) and 7 GTCS (in six patients). Behavioral results were analyzed in the ictal and postictal periods based on video review. During both CPS and GTCS, patients were unable to respond to questions or commands. However, during CPS, patients often retained minimally conscious ball grasping and visual tracking responses. Patients were able to successfully grasp a ball in 60% or to visually track in 58% of CPS, and could carry out both activities in 52% of CPS. In contrast, during GTCS, preserved ball grasp (10%), visual tracking (11%), or both (7%), were all significantly less than in CPS. Postictal ball grasping and visual tracking were also somewhat better following CPS than GTCS. These findings suggest that impaired consciousness in CPS is more similar to minimally conscious state than to coma. Further work may elucidate the specific brain networks underlying relatively spared functions in CPS, ultimately leading to improved treatments aimed at preventing impaired consciousness.

Impaired consciousness drastically alters normal human quality of life. In coma, patients show no meaningful reaction to their surroundings, whereas in other disorders of consciousness, patients may exhibit some simple responses (Laureys & Tononi, 2009). For example, patients in a vegetative state can open their eyes and turn toward stimuli but do not show visual tracking or respond appropriately to questions or commands; patients in a minimally conscious state show simple responses such as ball grasping and visual tracking but do not show consistent interactive communication or object use (Giacino et al., 2002). Epileptic seizures also cause impaired consciousness, which varies in severity. In generalized tonic–clonic seizures (GTCS) there is profound impairment, transiently resembling coma (Blumenfeld et al., 2009). However, in other seizures milder deficits may be seen. In commonly used terminology (ILAE 1981), partial seizures without impaired consciousness are called simple partial, whereas those with impaired consciousness are called complex partial seizures (CPS). Determining which specific behavioral functions are impaired and spared in CPS may help elucidate the anatomic networks underlying impaired consciousness in CPS, leading ultimately to improved treatments. However, the precise nature of impairment during CPS has not been thoroughly investigated. Specifically, it is not known whether CPS more closely resembles a transient vegetative or minimally conscious state. Therefore, we prospectively evaluated the ability of patients to perform simple tasks including ball grasping and visual tracking with a standardized testing battery (Yang et al., 2012) in CPS as well as GTCS during inpatient video–electroencephalography (EEG) monitoring.

This objective method of assessing impaired consciousness is significant in that further study could allow for greater understanding of minimally conscious capacities during CPS. In addition, this could assist in functional mapping of specific neurobiologic substrates of consciousness when correlated with seizure localization based on neuroimaging and EEG findings (Blumenfeld, 2011).


All procedures were approved by the institutional review board at Yale University School of Medicine and participants provided written informed consent. We recruited inpatients at Yale-New Haven Hospital undergoing epilepsy video-EEG monitoring from June 2009 through July 2011. Inclusion criteria were the following: age 7 years or older and the ability to follow simple questions and commands at baseline. Exclusion criterion was the presence of nonepileptic seizures.

As described previously, trained examiners sat with patients during their hospitalization for an average of 27 h per patient, and immediately after seizure onset commenced a standard behavioral testing battery, the Responsiveness in Epilepsy Scale (RES) (Yang et al., 2012). The initial items of RES test ability to respond to verbal questions and commands. Patients who failed these were tested on ability to grasp a tennis ball placed on the dorsal surface of each hand while being told “take the ball,” and on visual tracking, being told “look at the mirror” as a mirror was moved slowly to the left and right in front of the face. Testing was repeated during the ictal and postictal periods until patients returned to their baseline level of behavioral performance (Yang et al., 2012).

Analysis was performed based on review of video–EEG recordings by consensus of two reviewers. Ability to grasp the ball with at least one hand or to visually track in at least one direction (left or right) were considered successful responses. As in prior work (Giacino et al., 2004), to attain a correct grasp response we required that the wrist must rotate and the fingers extend as the ball is moved along the dorsal surface of the hand, and the ball must be held without dropping it; to attain correct visual tracking, eyes must follow the mirror for 45 degrees in one direction without loss of fixation. Mean success rate per seizure was calculated across repeated administrations of each test and analyzed separately during the ictal and postictal periods. Seizures types were classified based on video-EEG review using the 1981 ILAE criteria (ILAE 1981).


Testing was initiated in a total of 84 seizures (24 patients). Patients failed to respond to questions and commands in 34 seizures, of which 27 were CPS (14 patients, five male) and 7 were GTCS (six patients, four male) (Table 1). All tested seizures had a focal onset, including the GTCS, which were secondarily generalized. Absence seizures were not observed in this patient sample. Mean (± standard error of the mean, SEM) patient ages for CPS and GTCS, respectively, were 38.7 ± 3.7 and 29.1 ± 3.3 years. Mean seizure durations respectively were 116 ± 23.2 and 209.7 ± 75.1 s. Ball grasping was tested ictally a mean of 3.6 ± 0.6 times per seizure and visual tracking 3.2 ± 0.5 times per seizure. In the postictal period, ball grasping was tested 7.7 ± 2.2 and visual tracking 6.8 ± 2.2 times per seizure.

Table 1.   Clinical information for patients tested
PtAge, sexNo. szs, typeMRIPET hypometabolismIctal SPECT hypermetabolismScalp EEG onsetaIntracranial EEG onsetaOverall localization
  1. a Based on overall EEG of all seizures, not just seizure occurring during RES.

  2. HC, hippocampal; R, right; L, left; O, occipital; T, temporal; P, parietal; F, frontal; inf, inferior; lat, lateral; mes, mesial; Bi, bilateral; H, hemisphere; GTCS, generalized tonic–clonic seizure; CPS, complex partial seizure; szs, seizures; Pt, patient.

 151M3 CPSL T atrophy and polymicrogyriaNormalL FL > R HMostly R inf F-T, but some Bi or L TUnlocalized
 227M3 CPSR HC atrophy and L T gray matter abnormalL TL T and R OL T and R TN/AR O and L T
 318F2 CPS, 1 GTCSNormalL TL TBi TR > L mes TBi T
 427M1 GTCSNormalL TN/AL HL FL F
 521M1 GTCSNormalN/AN/AL FN/AL F
 642M2 CPSPrevious L T resectionR lat TR TR > L TR mes T, and R T-OR T, O
 758F1 CPSIncreased signal Bi HC and RPN/AN/ABi TN/ABi T
 830F1 CPSR F cortical thickeningR FN/AR FR FR F
 924F2 CPSL T heterotopiaN/AN/ABi FN/AUnlocalized
1033F1 GTCSEncephalomalacia, R T-P-OR T-P-ON/AR hemisphereR F, OR H
1151M3 CPSNormalBi TN/AL HMultifocalUnlocalized
1221F2 CPSPrevious R F resectionNormalN/AMultifocalN/AUnlocalized
1340M1 GTCSNormalN/AR TR TN/AR T
1433M2 GTCSBi R > L HC atrophyNormalN/ARight temporoparietalN/AR H
1524F1 CPSNormalBi TN/AL F, TL TL T
1629F2 CPSR HC atrophy, R O indistinct sulciR TN/AR TR H diffuse onsetR H
1759M1 CPSR HC atrophy, R F, P infarctR T, and in area of R F, P infarctN/AR TN/AR T
1838F3 CPSNormalMild R TN/AL HN/AL H
1943F1 CPSPrevious L T resectionNormalN/AL TL HC, L fusiform gyrusL HC, L fusiform gyrus

We found that ball grasping and visual tracking were spared over half the time in complex partial seizures, but uncommonly in generalized tonic–clonic seizures. During CPS, the mean success rate for ball grasping was 60 ± 12%, the mean success rate for visual tracking was 58 ± 12%, and for both 52 ± 11% (Fig. 1). These were significantly better than during GTCS where mean success rate for ball grasping was only 10 ± 10%, for visual tracking 11 ± 11%, and for both 7 ± 7% (all p < 0.05, t-test). In the postictal period, ball grasping and visual tracking were also somewhat better following CPS than after GTCS (Fig. 1).

Figure 1.

 Evidence of minimal consciousness during CPS. Ictal mean success rates were >50% on the ball grasping and visual tracking tasks in CPS. This was significantly higher than during GTCS. Postictal success rates on these tasks were also somewhat higher following CPS than after GTCS. Results for success in both tasks in the same testing cycle (Both) were also significantly better in CPS than in GTCS. *p < 0.05.

In the 10 patients who had multiple seizures (Table 1) performance was relatively consistent. Therefore, within-patient standard deviation for ictal success rate was 11.7% on average, and a majority of these patients had at least two seizures with identical success rates. Of note, performance did not wax and wane, but rather reached maximal impairment and then progressively recovered in the postictal period. In only a minority of testing items (∼5%) did a patient have worse performance than previously during that same postictal period and, among these cases, all lasted only a single testing iteration.

It was recently found that blink to visual threat may be spared in some patients who otherwise meet all criteria for the vegetative state (Vanhaudenhuyse et al., 2008). We tested ictal blink to visual threat in a limited number of patients and found that it was spared in three of four CPS, yet spared in no (0 of 7) GTCS (p = 0.02, Fisher’s exact test).


The results described herein demonstrate that patients can often exhibit simple responses such as ball grasping and visual tracking during CPS. In contrast, these responses were usually absent in GTCS. For the single patient in whom these behaviors were preserved during a GTCS, the responses were made as the seizure evolved from generalized to limited unilateral involvement. These findings support prior speculation that CPS more closely resemble a transient minimally conscious state, rather than more severe disorders of consciousness such as coma or vegetative state (Blumenfeld, 2011). Patients in coma or a vegetative state do not exhibit visual tracking or reaching and grasping, nor do they have automatisms that can be seen in the minimally conscious state and CPS (Giacino et al., 2002).

Recent advances in neuroimaging have identified important anatomic networks underlying disorders of consciousness (Laureys & Tononi, 2009). Of interest, the same regions of frontoparietal association cortex and subcortical arousal circuits that are dysfunctional in coma, vegetative state, and minimally conscious state are often transiently impaired in epileptic seizures (Blumenfeld, 2011). When patients improve from vegetative to minimally conscious state, increased activity has been observed in these same networks (Laureys et al., 2008). With further work, it may be possible to determine if spared minimal conscious responses in CPS are associated with less severely impaired frontoparietal function in comparison with GTCS. Indeed, prior work suggests that although neocortical blood flow and electrical activity show significant abnormal increases in GTCS, neocortical activity is moderately decreased in CPS (Blumenfeld et al., 2009; Englot, et al., 2010; Blumenfeld, 2011). Other factors including unilateral versus bilateral or dominant versus nondominant hemispheric involvement could also participate in the degree of impairment and should be investigated further.

In previous studies, “reactive automatisms” have been described in CPS in which patients display simple automatic responses to the environment (Escueta et al., 1982). The present work adds to these observations by using standardized testing to prospectively quantify the frequency of preserved minimal consciousness in CPS. Further work is needed to identify the anatomic and physiologic basis of spared and impaired cognitive function during different seizure types. It is hoped that better understanding of these mechanisms will ultimately lead to new improved treatments for preventing impaired consciousness and its consequences.


We thank Abhijeet Gummadavelli and Asht Mishra for the useful comments on the manuscript, and all of the undergraduate volunteers who performed RES examinations. This project was supported by NIH R01NS055829, The Patrick and Catherine Weldon Donaghue Medical Research Foundation and the Betsy and Jonathan Blattmachr family. Alison McPherson was supported by the Connecticut College CELS Program, Leticia Rojas by the National Heart Lung and Blood Institute of the NIH, Joshua Motelow by NIH TG T32GM07205, and Andrew Bauerschmidt by the Doris Duke Charitable Foundation.


None of the authors have financial disclosures to make. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines. None of the authors has any conflict of interest to disclose.