Apparent False Lateralization of Seizure Onset by Scalp EEG in Temporal Lobe Epilepsy Associated with Cerebral Cavernous Malformation: A Case Report and Overview

False lateralization of ictal onset by scalp electroencephalogram (EEG) is an infrequent entity that has been reported in patients with mesial temporal lobe epilepsy associated with hippocampal sclerosis (HS). In these cases, a tendency for rapid seizures that spread through the frontal-limbic system and hippocampal commissural pathways to the contralateral hemisphere has been proposed. Cerebral cavernous malformations (CCMs), which constitute a collection of abnormally configured small blood vessels with irregular structures, is a well-defined epilepsy-associated pathology. Their primary association with seizures might be explained either as a result of physiological changes affecting the cerebral cortex immediately surrounding the CCM (an epileptogenic mechanism that is relevant for both, temporal and extratemporal lesions) or as a result of promoting epileptogenicity in remote but anatomo-functionally connected brain regions, a mechanism that is particularly relevant for temporal lobe lesions. To date, there have been only two publications on falsely lateralizing ictal onsets by EEG in temporal cavernoma, but not in other regions. Here, we report a rare case of apparent false lateralization of ictal onset by scalp EEG in a patient with a left medial frontal gyrus cavernoma (supplementary motor area), and discuss some relevant pathophysiological mechanisms of false lateralization.


Introduction
Infrequently, seizures may be falsely lateralized to the contralateral hemisphere based on ictal scalp electroencephalogram (EEG) recordings. False lateralization of ictal onset by scalp EEG has been reported in patients with severe hippocampal sclerosis (HS) and cortical atrophy [1,2] as well as lesions in other supratentorial regions like the superior temporal gyrus, hippocampus, and neocortical temporal lobe [3] (Table 1). Mintzer et al. found that 4.7% of their cohort with severe HS had ictal onset in the normal temporal lobe on scalp ictal EEGs, but depth electrode recordings then showed ictal onset in the mesial temporal lobe ipsilateral to the imaging abnormality [1,4].

Normal
No area of hypo metabolism Interictal: sharp waves from both temporal head regions. Ictal: rhythmic 5-6 ictal activity from the left temporal head region.
Ictal seizure onset in the right temporal lobe. 1.5 s after seizure onset on the right: gamma activity in the left mesial temporal and the left lateral temporal electrodes.
Temporal neocortex epilepsy: rapidly involve the contralateral mesolimbic structures; spread through the hippocampal commissures. Initial low voltage fast activity in the right temporal was undetectable on scalp EEG.
[ Interictal. sphenoidal EEG: bilateral temporal, independent sharp waves with left sided predominance Ictal: discharge arise from the right frontotemporal regions, predominance over the right temporal region, 12 s after the seizure onset, ictal discharge spread to the contralateral hemisphere.
Onset of the ictal discharge in the left temporal lobe Cerebral damage reduces the voltage of the ictal discharge on the side of the lesion and higher amplitude in the normal hemisphere.
[ Reduced epileptiform activity in the atrophic hemisphere. Scalp EEG and MEG propagated electricals signals to the right side are easier to detect than signs to the left side. [3] Brain Sci. 2020, 10,584   Decreased glucose uptake in the left temporal lobe as compared with the right temporal lobe in both lateral and mesial cortical regions.
Onset in the right amygdala, although one subclinical seizure began in the left temporal lobe.
EEG from the right amygdala showed continuous low-amplitude fast activity (>13 Hz) for extended times; at other times, this pattern was entirely absent.
The fast activity was believed to be epileptic, although it was asymptomatic. Several discrete subclinical seizures with a different electrographic pattern originated in the right amygdala, later spreading to the right hippocampus. Abnormal increase in temporal lobe metabolism.

Case Presentation
The patient was a 53-year-old right-handed female who presented an unknown onset seizure with motor activity (tonic-clonic) at the age of 17 years. Personal history for other neurological conditions was negative. Epilepsy risk factors, including febrile seizures, head trauma, central nervous system (CNS) tumor, and family history, were also negative. The patient denied smoking, drinking alcohol, or using illicit drugs. The neurological exam was normal, and she scored 26/30 on the Montreal Cognitive Assessment (MOCA) test. Interestingly, she did not have more seizures until she was 26 years old. This last seizure had a focal onset, with impaired awareness, and a motor component that involved manual automatism, blinking, cephalic version to the right, and asymmetric tonic limb posturing with progression to bilateral tonic-clonic seizures.
Her seizures worsened progressively due to increased frequency and severity, with an ictal frequency of five per month. She was on carbamazepine 400 mg daily, levetiracetam 3000 mg daily, lacosamide 250 mg daily, and lamotrogine 100 mg daily with poor response. Then, she was referred to our hospital (Hospital Universitario Mayor-Méderi) to characterize her seizures. A long-term (72 h) video EEG with scalp electrodes was performed with a 50% reduction of the antiepileptic drug (AED) doses. The video EEG recorded 14 seizures in total. She presented two different types of seizures with different semiology: the first type was a seizure with a focal onset, impaired awareness, and non-motor components such as behavior arrest, cognitive impairment due to disorientation and emotional seizures (psychomotor agitation, fear, and anxiety) and eye blinking. The second one was a focal onset seizure with impaired awareness, motor components such as eye blinking, cephalic version to the right, dystonic extension posture in the right arm (Figure 4 sign), and finally a progression to bilateral tonic-clonic seizure. After video EEG was completed, she was restarted on the full dose of AEDs, achieving seizure control within the next 24 h.
The scalp EEG showed interictal rhythmic spike activity in the left mesial temporal region maximal at the T3 electrode. A Left F7-T3 spike with intermittent fronto-central delta slowing was found. A right interictal temporal spike was also seen at the end of the recording (Figure 1). Ictal scalp EEG showed 4-5 Hz rhythmic activity that involved the right frontotemporal region. It spread to the contralateral hemisphere, where it was associated with epileptiform activity (Figures 2 and 3).

Case Presentation
The patient was a 53-year-old right-handed female who presented an unknown onset seizure with motor activity (tonic-clonic) at the age of 17 years. Personal history for other neurological conditions was negative. Epilepsy risk factors, including febrile seizures, head trauma, central nervous system (CNS) tumor, and family history, were also negative. The patient denied smoking, drinking alcohol, or using illicit drugs. The neurological exam was normal, and she scored 26/30 on the Montreal Cognitive Assessment (MOCA) test. Interestingly, she did not have more seizures until she was 26 years old. This last seizure had a focal onset, with impaired awareness, and a motor component that involved manual automatism, blinking, cephalic version to the right, and asymmetric tonic limb posturing with progression to bilateral tonic-clonic seizures.
Her seizures worsened progressively due to increased frequency and severity, with an ictal frequency of five per month. She was on carbamazepine 400 mg daily, levetiracetam 3000 mg daily, lacosamide 250 mg daily, and lamotrogine 100 mg daily with poor response. Then, she was referred to our hospital (Hospital Universitario Mayor-Méderi) to characterize her seizures. A long-term (72 h) video EEG with scalp electrodes was performed with a 50% reduction of the antiepileptic drug (AED) doses. The video EEG recorded 14 seizures in total. She presented two different types of seizures with different semiology: the first type was a seizure with a focal onset, impaired awareness, and non-motor components such as behavior arrest, cognitive impairment due to disorientation and emotional seizures (psychomotor agitation, fear, and anxiety) and eye blinking. The second one was a focal onset seizure with impaired awareness, motor components such as eye blinking, cephalic version to the right, dystonic extension posture in the right arm (Figure 4 sign), and finally a progression to bilateral tonic-clonic seizure. After video EEG was completed, she was restarted on the full dose of AEDs, achieving seizure control within the next 24 h.
The scalp EEG showed interictal rhythmic spike activity in the left mesial temporal region maximal at the T3 electrode. A Left F7-T3 spike with intermittent fronto-central delta slowing was found. A right interictal temporal spike was also seen at the end of the recording (Figure 1). Ictal scalp EEG showed 4-5 Hz rhythmic activity that involved the right frontotemporal region. It spread to the contralateral hemisphere, where it was associated with epileptiform activity (Figures 2 and 3).    A computed tomography (CT) scan showed a left frontal parasagittal lesion with calcifications without mass effect. Then, a 1.5 Tesla brain magnetic resonance image (MRI) was obtained and confirmed the presence of a 9 × 6 mm left frontal parasagittal lesion with the so-called 'popcorn' appearance. The nodule showed smooth, thin, and hypointense borders, with a homogeneously hyperintense core on the T2W (weighted) images. The T2 gradient echo (GRE) images defined the hemosiderin component of the lesion, also known as 'the blooming effect'. It did not enhance on T1W post-gadolinium images (Figure 4). Positron emission tomography (PET) brain imaging using [18]   A computed tomography (CT) scan showed a left frontal parasagittal lesion with calcifications without mass effect. Then, a 1.5 Tesla brain magnetic resonance image (MRI) was obtained and confirmed the presence of a 9 × 6 mm left frontal parasagittal lesion with the so-called 'popcorn' appearance. The nodule showed smooth, thin, and hypointense borders, with a homogeneously hyperintense core on the T2W (weighted) images. The T2 gradient echo (GRE) images defined the hemosiderin component of the lesion, also known as 'the blooming effect'. It did not enhance on T1W post-gadolinium images (Figure 4). Positron emission tomography (PET) brain imaging using [18]  A computed tomography (CT) scan showed a left frontal parasagittal lesion with calcifications without mass effect. Then, a 1.5 Tesla brain magnetic resonance image (MRI) was obtained and confirmed the presence of a 9 × 6 mm left frontal parasagittal lesion with the so-called 'popcorn' appearance. The nodule showed smooth, thin, and hypointense borders, with a homogeneously hyperintense core on the T2W (weighted) images. The T2 gradient echo (GRE) images defined the hemosiderin component of the lesion, also known as 'the blooming effect'. It did not enhance on T1W post-gadolinium images (Figure 4). Positron emission tomography (PET) brain imaging using [18] FDG fluorodeoxyglucose showed a pseudonodular lesion with internal calcifications of 9 × 6 mm approximately without radiopharmaceutical uptake (ametabolic). No abnormal perfusion in the surrounding area next to the cavernoma was found. An acceptable distribution and radiofrequency uptake by other brain structures, including the temporal lobes, was shown ( Figure 5).
FDG fluorodeoxyglucose showed a pseudonodular lesion with internal calcifications of 9 × 6 mm approximately without radiopharmaceutical uptake (ametabolic). No abnormal perfusion in the surrounding area next to the cavernoma was found. An acceptable distribution and radiofrequency uptake by other brain structures, including the temporal lobes, was shown ( Figure 5).
In summary, the present case characterized by ictal semiology and onset in the left hemisphere matched to the side of the cavernoma localization. However, the scalp EEG showed a right onset and fast spread to the contralateral hemisphere, meaning there was conflicting evidence regarding lateralization.

Outcome and Follow Up
According to several guidelines for cavernoma-related epilepsy management [9,15], surgical resection of the cavernoma was offered to the patient by our epilepsy group. However, the patient declined the procedure for personal reasons, and AED therapy was restarted. Medication dosage was adjusted (levetiracetam 2000 mg daily and lamotrigine 400 mg daily) due to the development of somnolence and drowsiness, resulting in partial control of the seizures.

Discussion
This case illustrates a patient with different types of seizures, some of them with lateralizing and localizing features. The second type seizure of our patient was characterized by head version to the right, followed by asymmetric tonic limb posturing and bilateral progression to tonic-clonic seizures, which localized the symptomatogenic zone to the left side. When the head version is the first and predominant symptom of a seizure, it usually helps to localize the epileptogenic area to the contralateral side as this is true in around 90% of cases [16,17]. If the seizure progresses to a bilateral tonic-clonic seizure, the forced head deviation would be caused by the activation of the frontal eye field and motor areas, anterior to the precentral gyrus and close to the left cerebral cavernous malformation (CCM) [17]. The right elbow assumes an extended position, and the left elbow flexes over the chest, assuming an asymmetric tonic limb posturing. This sign provides correct lateralization in most cases, in which the extended elbow is contralateral to the epileptogenic hemisphere [16]. According to the ictal semiology in our patient, the epileptogenic zone was identified in the left side, matching the CCM location. However, neither a scalp EEG nor functional tests were concordant.
Functional connections between the frontal lobe and hippocampus, and later rapid spread from commissural pathways between both hippocampi may underlie false lateralization. Indeed, rapid contralateral propagation of seizures by activation of temporal and frontal limbic pathways has been proposed in several studies [12,18,19]. Another hypothesis suggests that false lateralization by scalp EEG might be caused by severe tissue damage induced by inflammation and reactive gliosis, which leads to a reduction in the voltage on the side of the lesion. Likewise, in patients with temporal lobe epilepsy and focal brain lesion, the epileptiform discharges could be reduced and difficult to detect by extracranial EEG [1,2]. However, in the present case, no radiologic evidence of encephalomalacia surrounding the CCM was found. Finally, electrical interference caused by anatomical components such as meninges, skull, or scalp must also be considered in cases of false lateralization as it can affect the detection of the epileptic focus [20] (a summary of false lateralization hypothesis are included in Table 1).
Another possible explanation of false lateralization may be the presence of two epileptogenic loci. Correspondingly, there are some reports of coexistence of cavernoma and cortical dysplasia in patients with refractory epilepsy [21][22][23]. Some authors consider cavernomas as congenital abnormalities, similar to neurodevelopmental brain lesions (NDLs) such as cortical dysplasia and glioneuronal tumors [24]. Therefore, the exclusion of dual pathology, as well as the study of other In summary, the present case characterized by ictal semiology and onset in the left hemisphere matched to the side of the cavernoma localization. However, the scalp EEG showed a right onset and fast spread to the contralateral hemisphere, meaning there was conflicting evidence regarding lateralization.

Outcome and Follow Up
According to several guidelines for cavernoma-related epilepsy management [9,15], surgical resection of the cavernoma was offered to the patient by our epilepsy group. However, the patient declined the procedure for personal reasons, and AED therapy was restarted. Medication dosage was adjusted (levetiracetam 2000 mg daily and lamotrigine 400 mg daily) due to the development of somnolence and drowsiness, resulting in partial control of the seizures.

Discussion
This case illustrates a patient with different types of seizures, some of them with lateralizing and localizing features. The second type seizure of our patient was characterized by head version to the right, followed by asymmetric tonic limb posturing and bilateral progression to tonic-clonic seizures, which localized the symptomatogenic zone to the left side. When the head version is the first and predominant symptom of a seizure, it usually helps to localize the epileptogenic area to the contralateral side as this is true in around 90% of cases [16,17]. If the seizure progresses to a bilateral tonic-clonic seizure, the forced head deviation would be caused by the activation of the frontal eye field and motor areas, anterior to the precentral gyrus and close to the left cerebral cavernous malformation (CCM) [17]. The right elbow assumes an extended position, and the left elbow flexes over the chest, assuming an asymmetric tonic limb posturing. This sign provides correct lateralization in most cases, in which the extended elbow is contralateral to the epileptogenic hemisphere [16]. According to the ictal semiology in our patient, the epileptogenic zone was identified in the left side, matching the CCM location. However, neither a scalp EEG nor functional tests were concordant.
Functional connections between the frontal lobe and hippocampus, and later rapid spread from commissural pathways between both hippocampi may underlie false lateralization. Indeed, rapid contralateral propagation of seizures by activation of temporal and frontal limbic pathways has been proposed in several studies [12,18,19]. Another hypothesis suggests that false lateralization by scalp EEG might be caused by severe tissue damage induced by inflammation and reactive gliosis, which leads to a reduction in the voltage on the side of the lesion. Likewise, in patients with temporal lobe epilepsy and focal brain lesion, the epileptiform discharges could be reduced and difficult to detect by extracranial EEG [1,2]. However, in the present case, no radiologic evidence of encephalomalacia surrounding the CCM was found. Finally, electrical interference caused by anatomical components such as meninges, skull, or scalp must also be considered in cases of false lateralization as it can affect the detection of the epileptic focus [20] (a summary of false lateralization hypothesis are included in Table 1).
Another possible explanation of false lateralization may be the presence of two epileptogenic loci. Correspondingly, there are some reports of coexistence of cavernoma and cortical dysplasia in patients with refractory epilepsy [21][22][23]. Some authors consider cavernomas as congenital abnormalities, similar to neurodevelopmental brain lesions (NDLs) such as cortical dysplasia and glioneuronal tumors [24]. Therefore, the exclusion of dual pathology, as well as the study of other vascular malformations are mandatory [25,26]. In a retrospective study, patients with cavernomas and NDL exhibit similar frequencies of sporadic spikes, as well as continuous spikes, burst, and recruiting spikes in the neocortical region [24]. In the same study, it is reported that the absence of coincident bursts was associated with high microglia density in patients with cavernomas, aiding to inhibit epileptogenic discharges in this area. In our case, continuous spikes were recorded in most of the EEG studies associated with occasional burst and recruiting spikes. This suggests a moderate microglia density. However, we did not have neuropathological studies for this patient.
Complementary to ictal semiology, invasive methods may help to reduce the probability of false lateralization. In fact, a retrospective study in patients with refractory focal epilepsy who were implanted with bilateral depth electrodes found that in 60% of cases, the scalp EEGs were discordant with intracranial EEG recordings: 27% of the scalp EEGs falsely lateralized the onset, and 33% suggested bilateral independent or single ictal onset [20]. However, advances in anatomic and functional neuroimaging (i.e., 18F-FDG PET and single photon emission computed tomography SPECT) have reduced the need for invasive monitoring in certain patients with focal epilepsy [27,28]. Although functional neuroimaging has shown several metabolic and perfusion changes in patients with refractory and focal epilepsy [13,27], in our clinical case the metabolism surrounding the lesion was unaffected according to the PET scan. Usually, it is expected to find increases in brain metabolism and cerebral blood flow in the epileptogenic focus. However, functional neuroimaging, such as SPECT, has a low sensitivity and high false-positive rate depending on the time the tracer injection is administered [29]. Although intracranial EEG is the gold standard to guide an epileptogenic lesion resection, our patient did not accept any invasive procedure.
Conservative management with AED and regular follow-up with an epileptologist are considered in patients unable to be compliant with the surgical treatment, as well as patients with cavernomas that are difficult to get to (i.e., lesion in the brainstem) or contiguous to eloquent brain areas. In concordance, several reports have shown adequate symptomatic response in patients with cavernoma-related epilepsy treated only with AED [30,31]. Conversely, early surgery is considered in patients with a high risk of bleeding with poor adherence to the medication or drug-resistant epilepsy [9]. For instance, Rammo et al. reported a 12-year-old male with refractory epilepsy caused by a right superior temporal cavernoma associated with false lateralization by scalp EEG. In that case, a cavernoma lesionectomy was performed. In addition, FDG-PET images showed bilateral hypometabolism with two irritative zones (frequent spikes) detected by magnetoencephalography (MEG). One of them was surrounded by the cavernoma in the left anterior temporal lobe and the second one in the right intrasylvian region, which was difficult to detected by EEG [10]. Even when findings on MRI and EEG are discordant, surgical intervention may still be offered if ictal onset is definitely present on the intracranial EEG or intraoperative electrocorticography (ECoG), in order to improve prognosis and treatment outcomes in drug-resistant epileptic patients [11,32]. In a study of Jin et al. using MEG and EEG, it was found that patients with a single cavernoma have different spike foci distributions: perilesional, at a remote site in the ipsilateral hemisphere and in a contralateral homologous location (called as "mirror"). Moreover, they reported a higher detectability of mirror and remote spikes by MEG than scalp EEG in patients with single cavernoma and, as in our case, the mirror spike suggests a rapid propagation of the interictal activity from the perilesional epileptogenic focus [33]. In the case of residual epileptic foci after extended lesionectomy and intraoperative ECoG monitoring, additional thermocoagulation could be required. Patients with lesions in surgically deep areas or located in regions of high morbimortality would benefit from stereotactic radiosurgery [15], together with new technologies in surgical neuronavigation, tractography, and awake mapping [34][35][36]. New reports derived from in vitro and in vivo studies suggest that in-depth knowledge in the molecular and cellular mechanism underlying the cavernoma formation may open the possibility for the development of new target inhibitors such as antineoplastic and antiangiogenic drugs as an alternative treatment to surgical and AED management in the case of intractable epilepsy induced by cavernoma malformations [37].

Conclusions
Most reports of the false lateralization of seizures have been described in patients with temporal epilepsy. Indeed, the few cases of false lateralization of cavernomas reported in the literature correspond to the temporal lobe, including the hippocampus. This is the first case of apparent false lateralization in the temporal region reported in a patient with a cavernoma located in the frontal lobe. Although the exact mechanism of false lateralization is not known, this case suggests a predilection for rapid electrical spread through the frontal system and the hippocampal commissure to the contralateral hemisphere that is difficult to detect by scalp EEG. Considering that cortical lesions in brain MRI do not necessarily indicate the exact region of ictal onset, individuals with discordance between ictal semiology and scalp EEG would benefit from intracerebral recordings and functional neuroimaging, such as MEG or PET. Adequate neurophysiological monitoring and a histopathological study, as well as greater knowledge of the molecular and genetic background, could be useful for developing new therapeutic measures in patients with cavernoma-related epilepsy that does not respond to AED or who cannot have surgery. Although our patient declined surgical workup, including invasive monitoring (a procedure that is considered the gold standard to support the false lateralization diagnosis), we consider that ictal symptomatology provides important information in cases of apparent false lateralization.