John Libbey Eurotext

Postictal signs of lateralizing and localizing significance [published with videosequences]. Volume 4, numéro 1, March 2002


  • Postictal signs of lateralizing and localizing significance

Although most seizure-related events (except those documented by video-EEG-monitoring) that are seen by physicians belong to the postictal state, this critical period of time between seizure end and complete normalization of cognitive and motor function, has received relatively little attention from epileptologists. After the first description of postictal symptoms by Bravais in 1827 [1], and later by Todd [2], Gowers [3] and Jackson [4], for more than a century, no systematic study of postictal phenomena has been performed. This may be in part due to (1) the wide range of postictal phenomena, including deficits in cognitive functions, as well as in motor, sensory and autonomic systems, (2) difficulties in defining seizure end clinically and electrographically, (3) infrequent observations of epileptic seizures by physicians until video-EEG-monitoring has been established, and (4) difficulties in distinguishing postictal confusion or lethargy from specific cognitive disorders such as aphasia. However, since the introduction of presurgical epilepsy monitoring, clinical seizure characteristics constitute a fundamental aspect for defining the localization and lateralization of the seizure onset zone. In recent years, it has become evident that in addition to a large number of ictal signs, some clinical features that appear during the postictal state may be of lateralizing or localizing significance and therefore could add some information to the presurgical evaluation process during prolonged video-EEG-monitoring as well as to the differential diagnosis (epileptic versus non-epileptic) of paroxysmal ("seizure-like") events, even in an ambulatory/emergency room setting.

Lateralizing postictal signs

Postictal hemiparesis (Todd's paralysis)

The first description of postictal paresis dates back to 1827, when Bravais reported hemiplegic symptoms after epileptic seizures [1]. However, it was Robert Todd in 1854 who studied postictal hemiplegia in more detail [2]: "A patient has a fit, distinctly of the epileptic kind; he comes out of it paralysed in one side of the body; generally that side is paralysed which had alone been convulsed, but the paralysis may occur where both sides had been convulsed equally. The paralytic state remains for a longer or shorter time, varying perhaps from a few minutes or a few hours to three or four days or even much longer. " Todd suspected some kind of neuronal exhaustion to be the underlying mechanism of postictal hemiparesis, a view shared by John Hughlings Jackson [4] who introduced the term "Todd's paralysis".

More than a century later, Rolak et al. [5] studied the incidence and clinical features of postictal hemiparesis in a large cohort of 229 patients with tonic clonic seizures. This heterogenous patient group was not classified according to the epilepsy syndrome, in many cases the aetiology of the seizure disorder was not clarified, the remaining patients were mostly classified as having stroke-related seizures or alcohol-withdrawal seizures. Postictal hemiparesis occurred in 6.1% of all patients, affecting the face, arm or leg, and persisted from 30 min to 36 hours. The only significant finding was that postictal hemiparesis occurred more often in patients with a structural lesion (i.e. pre-existing ischemic stroke) compared to those with normal findings on CT or MRI scans.

In our own study [6], consisting of 160 consecutive patients with focal epilepsy who underwent prolonged video-EEG-monitoring, we found postictal hemiparesis restricted to one arm in 13.8% of all patients, or in 10.4% of 135 patients with temporal lobe epilepsy, 33.3% of 15 patients with frontal lobe epilepsy, 28.6% of 7 patients with occipital lobe epilepsy, and in 33.3% of 3 patients with parietal lobe epilepsy. The hemiparesis was restricted to the arm contralateral to the hemisphere of seizure onset in every case, regardless of whether the seizure activity spread to the contralateral hemisphere on surface EEG or not.

In contrast to the study of Rolak et al. postictal hemiparesis in our patient group was of short duration and never exceeded 10 min. This may in part reflect the different study population, since we did not include patients with underlying ischemic strokes or alcohol-withdrawal as a possible explanation for the seizure disorder.

Similar findings were obtained by Adam et al. [7] who found postictal hemiparesis in 40% of thirty patients undergoing prolonged video-EEG-monitoring. The higher incidence may be in part explained by the fact that hemiparesis was not restricted to the arms, but also included the face and legs. The duration of these motor deficits was short, although no definite time gate was presented.

Concerning the pathophysiological mechanisms of postictal hemiparesis, the two major theories propounded by Jackson and Gowers are still under discussion [8]. Jackson accepted Todd's original hypothesis of neuronal exhaustion as the mechanism of postictal hemiparesis, relying on his clinical observation that it was always "in the parts first and most convulsed" [4]. He even thought postictal cognitive deficits to be due to "universal or widespread paralysis, a sort of paralysis of higher functions". Gowers in 1881 [3], rejected the neuronal exhaustion theory in favour of his own hypothesis that some kind of inhibitory mechanisms, maybe the same that stopped the seizure, produced postictal hemiparesis. He argued, among others, with a case reported by Jackson himself in 1868: a man with focal seizures was able to stop the attack sometimes by vigorously rubbing his thigh. If this manoeuvre succeeded he had a paresis of his right leg; if it did not, he had no postictal weakness despite a longer, more severe seizure. Furthermore, Gowers cited cases in which relatively mild seizures resulted in a severe and long-lasting postictal hemiparesis, whereas severe seizures in the same patient did not. Gowers' theory has been lent some credence by studies of Kaibara et al. [9] who reported a marked dissociation between the length and extent of a seizure and the subsequent EEG depression or slowing, as well as by the studies of Matsumoto et al. [10] who suspected an ictal accumulation of extracellular potassium as one possible mechanism of postictal depression. Modern speculations on the pathophysiology of postictal depression have focused on adenosine [11] and endogenous opioids [12]. Experiments in kindled rats suggested postictal motor deficits to be mediated by the substantia nigra [13].

Postictal nose wiping

Several studies in the last three years have identified postictal nose wiping as a frequent and easy to assess symptom, of good lateralizing and potential localizing value in patients with focal seizures. In our own group of patients with focal epilepsy (n = 101), we found postictal nose wiping significantly more often in patients with temporal as compared to those with extratemporal lobe epilepsy (51.3% versus 12%; P = 0.001). The hand used for nose wiping was ipsilateral to the hemisphere of seizure onset in 86.5% of all seizures originating in the temporal lobe. Conversely, patients with extratemporal lobe epilepsy performed nose wiping with the ipsilateral hand in only 54.5% of all seizures [14]. Similar results have been reported by other authors [15-17].

The aetiology of postictal nose wiping remains somewhat speculative. A special kind of (ipsilateral) automatism has also been proposed as a specific and purposeful reaction to increased upper airway secretion during temporal lobe seizures, which furthermore, results in hypersalivation and coughing. In accordance to the finding that autonomic phenomena such as vomiting, lacrimation, yawning and coughing were more common in right temporal lobe epilepsy, we also found nose wiping more often in patients with right temporal lobe epilepsy, further supporting the theory of nose wiping to be a reaction to some kind of autonomic hyperexcitation. A depth electrode study has pointed out that involvement of the amygdala is crucial for the induction of postictal nose wiping, while seizures restricted to the hippocampus did not result in nose wiping [17]. Concern-ing the lateralizing value of postictal nose wiping, we suggest that the hand ipsilateral to the hemisphere of seizure onset is used because of a discrete contralateral paresis/weakness or neglect. The few patients with falsely lateralizing postictal nose wiping may be due to a spread of seizure activity to the contralateral hemisphere, a hypothesis that is strongly supported by the study of Wennberg who pointed out that a wrong lateralization (i.e. contralateral nose wiping) was in every case combined with a secondary seizure spread to the contralateral hemisphere [17]. This is in accordance with our own findings, suggesting nose wiping to be of less lateralizing value in patients with extratemporal lobe epilepsy, since those seizures tend to propagate more rapidly to the contralateral hemisphere as compared to temporal lobe seizures.

Postictal language dysfunction

Although speech disturbances are common during and after complex partial seizures, the literature is relatively scarce. There may be several reasons for this: (1) impaired consciousness, which according to the ILAE is by definition a key criterion of complex partial seizures, may interfere with speech. (2) Speech disturbances during and after epileptic seizures are often inconsistent in different seizures in the same patient, moreover they are often complex and difficult to analyse. Unresponsiveness of a patient during the postictal state for example can be (i) due to an inability to perform motor tasks which are requested by the investigator, (ii) due to a peculiar motivational state (i.e. the patient may be so absorbed in an experiental hallucination that he becomes oblivious to all stimuli not related to the prevailing focus of his attention), (iii) due to amnesia, and last but not least (iiii) due to aphasia. Aphasia itself can be due to a blocked verbal output (expressive or Broca-aphasia), due to an inability to comprehend verbal tasks (perceptive or Wernicke-aphasia) or both. Moreover, aphasia can present with more subtle symptoms such as paraphasias or difficulties in expressing certain words. (3) Before the development of video EEG monitoring, seizure localization was uncertain, which did not allow accurate evaluation of the importance of speech disturbances in seizure lateralization.

Gabr et al. [18] performed one of the first studies on speech manifestations in surgically-treated patients with temporal lobe epilepsy whose language dominance was verified by the intracarotid sodium amobarbital (WADA) test. Among all of the speech manifestations reported by the authors (ictal vocalization, identifiable ictal speech, abnormal ictal speech (not specified further) and postictal dysphasia), only two symptoms turned out to have reliable lateralizing value: identifiable ictal speech was significantly correlated with seizure onset in the non-dominant hemisphere, while seizures in patients with postictal dysphasia originated from the dominant hemisphere in 92%.

To avoid the problem of defining postictal aphasia and its possible overlap with postictal confusion or altered consciousness, Privitera et al. [19] performed a simple test where patients were asked to read a test phrase aloud as soon as seizure onset was detected. By quantifying the time delay in reading the test phrase, the authors were able to correctly lateralize seizure onset in all 26 patients. Patients with left temporal lobe seizure onset took more than 68 seconds to read the test phrase correctly, while patients with right temporal lobe seizures took less than 54 seconds. In a larger series by the same authors, this test was able to correctly lateralize the seizure onset zone in 72% of patients with left language dominance, but in only 55% of those with non-left hemisphere language dominance [20].

In our own series of patients (data not published), we tested 62 consecutive patients with temporal lobe epilepsy for postictal speech manifestations in at least one seizure. Patients were asked to raise their hands, name an object (i.e. pencil), to show how to use it and if the speech showed no obvious disturbances, spontaneous speech was carefully investigated for paraphasias, neologisms and difficulties in word-fluency. These tests allowed us to subcategorize speech disturbances into global, expressive and perceptive aphasias. All patients underwent the intracarotid sodium amobarbital test (WADA test) to determine the speech dominant hemisphere. Secondarily generalized seizures were excluded, as well as seizures with no well-defined unilateral EEG seizure onset. Postictal speech disturbances were found in thirty two patients (51.6%), consisting of complete aphasia lasting from 25 to 123 seconds in twenty one, expressive aphasia (i.e. ability to correctly perform motor commands and apparently normal comprehension but inability to speak) in five and paraphasias or difficulties in finding certain words in six patients. Overall, postictal speech disturbances correctly lateralized the seizure onset zone in the dominant hemisphere in 81.3% of all patients, regardless of which kind of dysfunction was documented (i.e. no statistically significant difference, although one must recognize the small groups of expressive aphasisa and paraphasias). Normal postictal speech was associated with a non-dominant hemisphere of seizure onset in 81.8%. In a smaller group of twenty nine patients, we repeatedly presented a test phrase immediately after seizure end. In patients with dominant hemisphere seizure onset, the time delay to read the test phrase correctly was more than 68 s in 72% (mean 323 s), while it was less than 54 s in three out of four patients with non-dominant EEG seizure onset. Overall, postictal speech disturbances seem to be a reliable lateralizing sign. Our data do not indicate the need for further differentiating the kind of speech disturbances with more subtle linguistic tests, since the different types of language disorders did not add any further information on the lateralization of the seizure onset zone. Nevertheless, one must keep in mind that a false lateralization by testing speech manifestations alone can occur in up to 20% of patients. This could be due either to (1) bilateral speech representation, (2) propagation of seizure activity to the contralateral hemisphere not evident on surface recordings (since no invasive EEG recordings in patients tested for postictal speech disturbances were performed), or (3) due to postictal confusion or other cognitive deficits that coincide with language testing.

Postictal headache

Postictal headache was recently reported in 52% of one hundred patients with focal epilepsy undergoing presurgical video EEG monitoring [21]. While there was no statistically significant association between the side of headache and the hemisphere of seizure onset in patients with extratemporal lobe epilepsy, postictal headaches occurred ipsilateral to the side of seizure onset in 89% of patients with temporal lobe epilepsy. Thus, lateralized postictal headache seems to be a reliable predictor of the side of seizure onset. Concerning the pathophysiological mechanisms of postictal headache, which is often migrainous in character, the authors favour a neurovascular theory where vasodilatation of large cranial vessels and subsequent stimulation of the trigenminal sensory nervous pain pathways may play a key role. This is in accordance with studies of Penfield and Jaspers who observed a widespread vasodilatation and prolonged postictal hyperemia in large pial vessels over the epileptic focus [22].

Postictal hemianopsia

We found only a few case reports of patients with transient postictal hemianopsia [23, 24]. All patients had more or less well defined focal epilepsy and the hemianopsia was contralateral to the suspected epileptogenic region in each case. The real incidence of postictal hemianopsia remains obscure since investigations of the visual fields after a seizure do not seem to be clinically practical, and the patient himself often does not recognize his deficit. The relatively complex investigation makes postictal hemianopsia a lateralizing sign of questionable value in routine investigation, even during video-EEG-monitoring.

Postictal cognitive dysfunction

Interictal cognitve deficits in patients with focal epilepsy, especially of temporal lobe origin, have been reported or at least hypothesized in several studies. The most accepted association is that of left sided-hippocampal sclerosis and verbal memory deficits, while patients with left-sided temporal lobe epilepsy without hippocampal pathology as well as patients with right-sided temporal lobe epilepsy, did not show significant deficits [25-27]. In contrast, the association of right-sided temporal lobe epilepsy and visual memory deficits remains somewhat speculative since patients with right hippocampal pathology performed less well in certain tests of delayed recall of visual material, while they did not in other tests of immediate or delayed recall [25]. With respect to the hypothesis that these deficits are caused by the "epileptogenic" lesion, frequent seizures or interictal epileptiform discharges, Helmstaedter et al. studied the hypothesis that these deficits are more pronounced immediately after a seizure than in the interictal period [28]. With a computerized test-program especially designed for the postictal period, they found a statistically significant correlation between left hemispheric temporal lobe epilepsy and deficits in verbal memory tasks on the one hand, and between right hemispheric temporal lobe epilepsy and visual memory deficits on the other hand. Interictal test results did not differ significantly between patient groups. Similar results were obtained by Pegna et al. [29] who concluded that interictal neuropsychology only helps occasionally in determining the localization of seizure onset, while postictal testing may be significantly more efficient even in differentiating temporal lobe epilepsy from frontal lobe epilepsy.

Localizing signs during the postictal state

Besides the lateralization of the hemisphere of seizure onset, the localization of the epileptogenic focus within one hemisphere is the most important issue of clinical seizure semiology. This is especially true for the differentiation of (a) temporal versus extratemporal lobe epilepsy and (b) mesial versus neocortical temporal lobe epilepsy. However, while some ictal symptoms may be helpful in the localization of the epileptogenic focus, postictal symptoms of true localizing value are extremely scarce. Postictal dysphasia is strongly associated with temporal lobe epilepsy [30, 31], although in our own study it was also seen in 15% of patients with occipital lobe epilepsy [6]. Prolonged postictal confusion was found more often in patients with amygdalar atrophy as compared to those with hippocampal atrophy [32]. Moreover, in differentiating syncopes from epileptic seizures, prolonged postictal confusion point to the latter. Postictal cough or sigh was seen more often in mesial as compared to lesional neocortical temporal lobe epilepsy [33], while it was not helpful in distinguishing hippocampal from extrahippocampal temporal lobe epilepsy in another study [34].

As regards postictal automatisms, we classified them in accordance to Penry et al. [35] and Szabo et al. [36] as "perseverative" automatisms which begin during the seizure and persist in the postictal state, and "de novo" automatisms which are only observed after EEG seizure end. In our group of 51 consecutive patients with postictal automatisms who were seizure-free after surgical treatment of their epilepsies, "perseverative" as compared to "de novo" automatisms were significantly more frequent in temporal lobe epilepsy (67.5% versus 32.5%, P < 0.02) while patients with frontal lobe epilepsy exclusively showed "de novo" automatisms [6].

Videosequences captions

Case 1: Postictal paresis

A 37 year-old male patient with unilateral left-sided mesial temporal lobe epilepsy who has been seizure-free since amygdalo-hippocampectomy was performed in 1997.

Clinical seizure semiology: 00:03:50 clinical seizure onset: oroalimentary automatisms, with the patient awakening, elevating his right upper extremity and head 2 s. later; 00:04:07 turns his head to the left; 00:04:11 beginning of automatisms with the left upper extremity followed by automatisms of the right upper extremity at 00:04:33; 00:05:12 clinical seizure end with patient wiping his face; 00:05:30 patient is asked if he can understand and nods; 00:06:02 marked postictal paresis of the right upper extremity, which has nearly complete disappeared 25 s later; 00:07:05 is not able to name various objects although he is repeatedly asked to do so and nods his head when he is asked if he can understand the question; 00:08:31 normal testing of upper extremity paresis.

EEG: 00:03:48 EEG seizure onset with non-lateralized rhythmic theta activity; 00:04:07 left temporal rhythmic theta activity; 00:05:07 EEG seizure end.

Case 2: Postictal nose wiping

A 29 year-old female patient with unilateral mesial temporal lobe epilepsy who became seizure-free after selective amygdalo-hippocampectomy in 1995.

Clinical seizure semiology: patient is reading a book; 18:05:58 clinical seizure onset: the patient suddenly stops reading, turns her head to the left and performs discrete oroalimentary automatisms; 18:06:00 automatisms of both legs followed by more prominent oroalimentary automatisms; 18:07:06 turns her head to the left; 18:07:30 ictal smile; 18:07:33 clinical seizure end; 18:07:36 postictal nose wiping with the right hand followed by clearing her throat.

EEG: 18:05:58 EEG seizure onset with unilateral right temporal rhythmic theta activity; 18:07:33 EEG seizure end.

Case 3: Postictal ("expressive") aphasia

A 37 year-old male patient with lesional temporal lobe epilepsy. After surgical removal of a dysembryoblastic tumor in the anterior parts of the left temporal lobe in 1996, he remains seizure-free except for several complex partial seizures within the first four months after surgery. The sodium amobarbital (WADA-) test revealed left hemispheric dominance for speech and memory.

Clinical seizure semiology: 19:56:26 clinical seizure onset: patient has a gastrointestinal aura, lays back and rings the bell to call the nurse; 19:56:43 ictal smile and automatisms of the left upper extremity; 19:56:50 dystonic posturing of the right upper extremity; 19:57:05 early head turning to the left, automatisms left upper extremity; no verbal response when the nurse calls the patient by name; 19:57:33 clinical seizure end with postictal nose wiping left arm; 19:57:37 nods when asked if everything is O.K.; 19:57:48 postictal automatisms left hand; 19:58:13 beginning of postictal language testing which reveals no perceptive aphasia but marked word finding difficulties and impaired word fluency.

EEG: 19:56:30 EEG seizure onset with left temporal rhythmic delta activity; 19:56:42 contralateral spread to the right temporal lobe; 19:57:21 EEG seizure end.

Case 4: Postictal aphasia after non-dominant hemispheric temporal lobe seizure

A 40 year-old female patient with unilateral right-sided mesial temporal epilepsy and left hemispheric dominance for speech and memory on sodium amobarbital (WADA-) test. The patient became almost seizure-free after selective amygdalo-hippocampectomy in 1996.

Clinical seizure semiology: 12:43:50 possible clinical seizure onset with staring; 12:43:55 oroalimentary automatisms; 12:44:30 automatisms right upper extremity; 12:44:56 turns head to the nurse who calls her name; 12:45:00 no verbal reaction if she is asked her name or to name an object; 12:47:05 clinical seizure end with postictal nose wiping with the right hand (several times); 12:47:27 postictal ("perseverative") automatisms right hand; no verbal reaction if she is asked to name objects; 12:49:15 no postictal paresis; 12:49:31 is not able to read a test phrase; 12:50:00 first verbal reaction if she is asked to name objects although there is some difficulty in finding the name of some objects; 12:51:35 is able to read the test phrase correctly.

However, in this patient it remains speculative if the language disturbance is due to a postictal confusional state or due to real postictal aphasia as the result of seizure propagation to the contralateral temporal lobe that was not detectable on surface electrodes. Nevertheless, this example illustrates that one has to be careful in interpreting the results of postictal language dysfunction.

EEG: 12:43:46 EEG seizure onset with right temporal theta activity; without (contralateral) propagation the seizure ends electrographically at 12:47:03.

Received June 1, 2001 ; Accepted October 23, 2001


Although postictal symptoms have received relatively little attention as regards their possible localizing or lateralizing significance in defining the seizure onset zone, we could show that this critical period of time in which the patient regains full consciousness after an epileptic attack offers valuable information concerning the seizure onset zone. Some postictal symptoms, especially postictal hemiparesis, speech disturbances and postictal nose wiping can be used to predict the hemisphere of seizure onset as well or even with a greater reliability as most accepted ictal symptoms. We therefore think that the postictal period should be evaluated with the same accuracy as the seizure itself, to further improve the localization accuracy of the seizure onset zone which is a prerequisite for the successful surgical treatment of epileptic seizures. Moreover, postictal symptoms may help to differentiate epileptic seizures from other paroxysmal events and they may offer some insight into the functional topography of the human brain.

Acknowledgements: We gratefully acknowlegde Dr Wolfgang Serles, Dr Achim Olbreich, Dr Ekaterina Pataraia and Dr G. Gröppel for medical patient care, Stefanie Lurger, Elfriede Antoni, Michaela Demel, and Birgit Seidl for technical assistance during prolonged video-EEG-monitoring and Dr Gerald Lindinger for technical support of the EEG-monitoring system. This research was supported by the "Jubiläumsfonds der Österreichischen Nationalbank" (project 8135).