Auditory aura in nocturnal frontal lobe epilepsy: a red flag to suspect an extra-frontal epileptogenic zone

Highlights

• •Eleven out of 165 nocturnal frontal lobe epilepsy (NFLE) patients reported an auditory aura as initial ictal symptom.
• •Extra-frontal origin was documented in 55% of NFLE patients with auditory aura.
• •Six patients with defined epileptogenic zone had a left temporal origin of seizures.
• •Auditory aura may be a symptom suggesting an extra-frontal epileptogenic zone.

Abstract

Objective

To describe the anatomo-electro-clinical findings of patients with nocturnal hypermotor seizures (NHS) preceded by auditory symptoms, to evaluate the localizing value of auditory aura.

Methods

Our database of 165 patients with nocturnal frontal lobe epilepsy (NFLE) diagnosis confirmed by videopolysomnography (VPSG) was reviewed, selecting those who reported an auditory aura as the initial ictal symptom in at least two NHS during their lifetime.

Results

Eleven patients were selected (seven males, four females). According to the anatomo-electro-clinical data, three groups were identified. Group 1 [defined epileptogenic zone (EZ)]: three subjects were studied with stereo-EEG. The EZ lay in the left superior temporal gyrus in two cases, whereas in the third case seizures arose from a dysplastic lesion located in the left temporal lobe. One of these three patients underwent left Heschl's gyrus resection, and is currently seizure-free. Group 2 (presumed EZ): three cases in which a presumed EZ was identified; in the left temporal lobe in two cases and in the left temporal lobe extending to the insula in one subject. Group 3 (uncertain EZ): five cases had anatomo-electro-clinical correlations discordant.

Conclusions

This work suggests that auditory aura may be a helpful anamnestic feature suggesting an extra-frontal seizure origin. This finding could guide secondary investigations to improve diagnostic definition and selection of candidates for surgical treatment.

1. Introduction

The term nocturnal frontal lobe epilepsy (NFLE) has been used to describe a rare form of focal epilepsy characterized by bizarre motor seizures occurring almost exclusively during sleep [1], in which frontal lobe involvement has been postulated since the early 1970s [2], [3]. The clinical characteristics of NFLE vary widely among different patients ranging from brief, simple motor phenomena (i.e. paroxysmal arousal, PA [4]) to major manifestations, including complex motor behavior such as epileptic nocturnal wandering (ENW) [5]. Different seizure patterns may coexist in the same patient, representing a continuum in the same epileptic condition [1], [6], [7]. In recent decades stereo-electroencephalography (sEEG) studies on patients operated for drug-resistant NFLE provided evidence of frontal lobe involvement during ictal motor phenomena with different clinical patterns depending on the specific regions involved [8], [9].

In a study of five patients with asymmetric tonic posturing, Rheims et al. have shown a relatively early activation of the supplementary motor area, with a different degree of involvement of the intermediate mesial frontal cortex and frontal cingulate gyrus [10]. Patients with hyperkinetic [11] ictal behavior showed the involvement of mesial-dorso-lateral, orbito-polar, opercular or larger lobar cortical regions [10]. All these seizure types could be included under the general term of hypermotor seizures [10] according to Lüders' classification of seizure semeiology [12].

Moreover, the complexity of motor manifestations and the identified dystonic features may reflect the involvement of subcortical structures, such as the basal ganglia and other cortical areas constituting a functional motor network responsible for ictal symptomatology [7], [9], [13], [14]. Whereas complex motor manifestations appear when the epileptic discharge involves the frontal lobe, the epileptogenic zone may lie in different cortical areas [15]. In particular it has been shown that nocturnal hypermotor seizures (NHS) may originate from the temporal and insular cortex and, recently, from the parietal and occipital lobe [9], [15], [16], [17], [18], [19]. Clinically, no semeiological markers differentiate between seizures of frontal and extra-frontal origin; the only objective symptoms suggestive of an extra-frontal focus are a lower seizure frequency and a major asymmetry in complex motor behavior prevailing ipsilaterally to the epileptic discharge [9], [15]. These findings emphasized the importance of anamnestic features, especially subjective symptoms, which are rare in frontal epilepsy and can point to an extra-frontal origin of the epileptic discharge [20]. These manifestations seem to better correlate with an extra-frontal origin of NHS. Subjective symptoms, such as laryngeal and throat discomfort, unpleasant or electrical paresthesia suggest an insular origin of NHS, whereas auditory aura may indicate a temporal origin [9], [17], [21], but these features have not been studied in depth.

Auditory auras associated with NFLE have been reported [22], [23] but these studies did not specifically address this symptom nor its correlation with NFLE. We describe the anatomo-electro-clinical data of a population of patients with NHS associated with auditory aura, diagnosed as NFLE patients.

2. Methods

The study was conducted at the IRCCS Institute of Neurological Sciences of Bologna, Department of Biomedical and NeuroMotor Sciences, University of Bologna (DiBiNeM) after approval by the local ethics committee obtained on September 16, 2010 (no. 10077). The records of all patients from our database of NFLE cases were reviewed, and those who reported an auditory aura as the initial ictal symptom in at least two NHS during their lifetime were selected. NFLE was ascertained when patients with a clinical history of NHS had a video-polysomnographic (VPSG) recording of at least one major attack (bilateral asymmetric tonic/dystonic posturing or other choreo-athetoid and ballistic movements of the limbs, i.e. hyperkinetic seizures) or at least two minor stereotyped episodes (i.e. PA) [24].

Anamnestic and clinical data were collected including: sex, age, handedness, family and personal history, age at seizure onset, seizure semeiology, anti-epileptic drugs (AEDs), therapeutic response (according to Perucca [25]), ictal and interictal EEG, neuroradiological findings, video-EEG, stereo-EEG (sEEG) findings, and surgical details when available.

The features of auditory aura, the side of origin (lateralization) and its relationship with nocturnal and diurnal seizures were described. Auditory symptoms were classified into: (1) simple auditory auras characterized by sounds such as buzzing, ringing, or whistling; (2) complex auditory auras characterized by organized sounds such as voices, music, or meaningful sounds. Three expert neurophysiologists (P.T., F.B., F.P.) reviewed all patients' VPSG data to characterize the semeiological features and classify the patterns of seizures [10], [26], into PA, hyperkinetic seizure (HK), tonic/dystonic (TD), and ENW.

The epileptogenic zone (EZ) was defined by the anatomo-electro-clinical correlation, considering ictal semeiology (including lateralization of auditory aura and language impairment), the side of interictal and ictal EEG abnormalities, VPSG data, the neuroradiological lesion when visible, sEEG findings, and surgical outcome if available. On these bases, patients were divided into three groups: (1) defined EZ, when pre-surgical and surgical information clearly identified the zone from which seizures arose; (2) presumed EZ, when surgical or pre-surgical data were not conclusive but some anatomo-electro-clinical data pointed to a specific zone; and (3) uncertain EZ, when definition of the epileptogenic zone was not possible.

Five patients, who gave their informed consent, were screened for known mutations of CHRNA4, CHRNA2, CHRNB2, KCNT1, DEPDC5, and LG1.

3. Results

3.1. Clinical and surgical data

Eleven out of 165 patients were selected based on the inclusion criteria; seven males and four females were studied (clinical data are shown in Table 1). Mean age of patients was 39.8 years (median 42 years); all but one were right-handed, and patient 4 was ambidextrous. Three patients reported a family history of epilepsy; of these, one (patient 8) had a familial form of NFLE. Two patients had had perinatal distress, another two had suffered from febrile seizures during childhood, and one patient reported head trauma. The mean age at the onset of epilepsy was 7.55 years (median 6 years). Nine out of 11 patients experienced both nocturnal and diurnal seizures: five patients had an onset of diurnal epilepsy (mean delay between nocturnal and diurnal onset of seizures was 1.8 years, range 0.5–4), two patients had an onset of nocturnal epilepsy (mean delay between nocturnal and diurnal onset of seizures was 8.75 years, range 5–12.5) whereas in two patients nocturnal and diurnal seizures began at the same time. All patients had experienced, at least once in their lifetime, very frequent NHS that occurred one or more times per night.

Table 1.

Main clinical data.

Patient	Age/sex	Family history	Personal history	Age at onset of epilepsy	Delay nocturnal–diurnal onset (years)	Anti-epileptic drugs	Drug response	Current frequency	Auras
1	36/M	−	FS	7	−5	CBZ, CNZ	SF	SF	R buzzing, R arm sensory aura
2	67/M	−	NS	12.5	−12.5	TPM	SF	SF	Complex auditory hallucination
3	45/M	−	FS, PA → quadriplegia	0	3	CBZ, CNZ	DR (IIIa)	Monthly	L leg shake sensation, siren sound
4	42/M	+	Head trauma at 20 months	6	0.5	OXC, CLB, PHT, VNS	DR (IIIb)	Monthly	R whistling sound, visual hallucination
5	47/M	−	NS	3.5	0	CBZ, TPM	DR (IIIb)	Weekly	R whistling sound
6	42/M	−	Meningo-encephalitis	7.5	0	CBZ, TPM	DR (IIIb)	Monthly	Music playing inside the head
7	25/M	−	PA	3	4	OXC, LTG, CLB	DR (IIIb)	Weekly	L whistling sound followed by animated cartoon song
8	45/F	+	NS	5.5	Only nocturnal	CBZ, PB	DR (IIIa)	Monthly	L arm sensory aura, bilateral buzzing
9	25/F	−	NS	5	1	VPA, OXC	DR (IIIb)	Daily	L whistling sound
10	37/F	+	NS	18	Only nocturnal	None	SF	SF	Tram sound inside the head
11	27/F	−	Migraine, somniloquy	15	0.5	OXC, TPM, LCS	DR (IIIb)	Daily	R arm sensory aura, bilateral buzzing sound

FS, febrile seizures; PA, perinatal asphyxia; NS, not significant; SF, seizure-free; DR, drug-resistant; L, left; R, right; LTG, lamotrigine; VPA, valproate; CBZ, carbazepine; OXC, oxcarbazepine; CLB, clobazam; TPM, topiramate; CNZ, clonazepam; PHT, phenytoin; LCS, lacosamide; PB, phenobarbital; VNS, vagus nerve stimulation.

Auditory aura was the first subjective symptom described by eight out of 11 patients; three patients reported a sensory aura that preceded the auditory symptoms. Auditory auras occurred both during the diurnal and nocturnal episodes in five patients, it was only nocturnal in five patients, whereas patient 6 reported an auditory aura only in diurnal hypermotor seizures. Usually, the auditory aura occurred while patients were falling asleep or in the early morning, and could awake the patient from sleep. Two patients could hear the same sound preceding nocturnal episodes during daytime, without motor manifestations. Five patients reported additional auras associated with auditory symptoms: four experienced sensory and one experienced visual aura. Seven patients reported simple auditory auras and only four patients reported complex auditory auras. The auditory symptoms were unilateral or prevalent in one ear in five patients, whereas they were contralateral to the presumed epileptogenic zone in three. Interictal EEG showed epileptiform anomalies in four patients (in the fronto-temporal region in two and in the temporal region in two) and non-specific anomalies in six patients. Ictal EEG recorded epileptic discharges in five patients (frontal origin in two cases, fronto-central in one subject, fronto-temporal in one and temporal in one patient) whereas two patients had non-specific modifications during seizures (flattening of background activity and bilateral desynchronization), in some cases autonomic modification (tachycardia/tachypnea) accompanied the seizure episodes. Neuroradiological examination showed abnormalities in four patients, which were highly suggestive of focal cortical dysplasia (FCD). Patients had a positive response to low doses of carbamazepine or oxcarbazepine, showing a consistent reduction of seizure frequency. Three patients are seizure-free whereas eight have drug-resistant epilepsy. According to the grade classification of drug resistance proposed by Perucca [25], six patients were classified as IIIB and two patients as IIIA.

On the basis of VPSG recording, seizure semeiology was classified as follows: six patients presented tonic/dystonic (TD) seizures, associated with PA in three cases; three patients had hyperkinetic (HK) seizures associated with PA in one; two patients presented both hypermotor manifestations (TD and HK).

The main anatomo-electro-surgical data are reported in Table 2.

Table 2.

Anatomo-electrical-surgical data.

Patient	MRI	Interictal EEG	Ictal EEG	VPSG seizure partners	sEEG/surgery
1	Infolding grey matter anterior F1 L	SC theta waves F-T R > L	NA	TD and PA	EZ L T1/Heschl's gyrus resected (FCD)
2	R temporal-polar hypotrophy	NA	Bi-frontal fast activity prevalent on R	PA and HK	NO
3	Hypoplasia of splenium and corpus callosum, cerebellar vermis dilatation	SC theta waves over L T-A and T-C regions	NA	TD and PA	NO
4	Suspicious dysplastic lesion T1/T2 L	Spike-and-wave activity over L F-T-P and L Z	Rhythmic SC theta activity over L T region	TD and HK	EZ over Wernicke's area/–
5	Dysplastic lesion in L operculum and insula	Spike-and-waves L F-T	Low voltage polyspike F3-C3, C3-T3	TD	EZ over posterior L insula/–
6	Normal	Bilateral theta burst F	NA	HK	NO
7	Normal	SC burst L F-T	Bilateral desynchronization F-T	HK	NO
8	Normal	SC burst over R hemisphere	Spike-and-wave activity over F region bilateral <R	TD and HK	NO
9	Dysplastic lesion T1 and lf peri-sylvian region	Spike-and-waves L F-T	Spike-and-wave activity L F-T region	TD	NO
10	Normal	Spike-and-wave over L Z	NA	TD and PA	NO
11	Normal	SC activity over L F-T	Abrupt flattening of background activity	TD	NO

MRI, magnetic resonance imaging; (s)EEG, (stereo-)electroencephalography; VPSG, video-polysomnography; SC, sharply contoured; T, temporal; F, frontal; P, parietal; A, anterior; C, central; Z, zygomatic electrodes; NA, not available; HK, hyperkinetic seizures; TD, asymmetrical bilateral tonic/dystonic seizures; PA, paroxysmal arousal; EZ, epileptogenic zone; NO, not operated; L, left; R, right.

3.2. Group 1 (three patients)

Pre-surgical evaluation with sEEG was performed in three patients presenting with drug-resistant epilepsy, demonstrating an extra-frontal origin of the seizures in all three. The ictal discharge started from the first temporal gyrus in two patients (patients 1 and 4), whereas the seizures arose from a lesion located between the temporal operculum and the posterior insula in one patient (patient 5). Neuroimaging and neurophysiological data (Fig. 1) were highly suggestive of FCD. Patient 1 was the only patient to undergo surgery. In this patient, nocturnal episodes started at age 7 years, characterized by a cry and motor agitation; at 12 years a tingling sensation in the right arm and a right-side buzzing preceded nocturnal episodes; 1 year later the buzzing sound preceded diurnal episodes of confusion, agitation and tachycardia; at the age of 25 years the aura disappeared. The interictal EEG showed bifrontal spike-and-wave activity more prevalent on the right; ictal EEG did not show clear abnormalities. During the night, the patient experienced tonic/dystonic seizures with left head orientation and dystonia of the right arm and leg. The patient underwent sEEG study with electrodes implanted over the left fronto-temporal regions, according to anatomo-electro-clinical localizing hypothesis. Seizures were recorded arising from Heschl's gyrus and the long insular gyrus (electrodes W, U, and L) to then involve Wernicke's area; an excision of Heschl's gyrus was performed, making the patient seizure-free (Fig. 2). The histological examination disclosed FCD type II.

Fig. 1.

Interictal electroencephalogram of patient 5. Low voltage spike-and-wave activity is visible over the left fronto-central region. Magnetic resonance imaging showed blurring of gray matter in the left temporal operculum and the posterior insula (white arrow).

Fig. 2.

Asymmetrical tonic/dystonic seizure documented with stereo-electroencephalography. Patient 1 was implanted over the left fronto-parieto-temporal area. The seizure started from electrodes placed over the auditory cortex, insula, and Wernicke's area (U, W, and L) and spread to the sensorimotor area (S) and insula (Z electrodes not shown). Excision of Heschl's gyrus was performed and the patient has been seizure-free for 1 year.

Patient 3 had perinatal anoxia and he developed spastic quadriplegia. The patient had very mild cognitive impairment but he was able to graduate at high school (middle school). We confirm that he described spontaneously a siren sound that wake up him before the seizures. In patient 3, seizures arising from Wernicke's area were recorded, but, due to the risk of language deficit, surgical treatment was contraindicated. In patient 5, seizures originated from a lesion between the left temporal operculum and the posterior insula, then the epileptic activity involved the frontal operculum, supplementary motor area and the posterior part of F1, motor area and post-central gyrus. Patient 5 died of sudden unexpected death in epilepsy (SUDEP).

All patients in group 1 reported simple auditory auras, which were always contralateral to the epileptogenic zone.

3.3. Group 2 (three patients)

In all cases, the anatomo-electro-clinical data pointed to an extra-frontal origin of seizures: patient 7 presented automatism in the left arm and contralateral dystonic posture during seizures. In addition, slow left head deviation and post-ictal aphasia were suggestive of left temporal lobe epilepsy. Patient 9 exhibited right dystonic posture associated with left automatisms, MRI disclosed a temporo-insular dysplastic lesion. Patient 10 presented ictal aphasia, and spike-and-wave activity on the left zygomatic electrode was recorded during interictal EEG. Patients 7 and 9 experienced simple auditory auras, in both cases ipsilateral to the presumed epileptic zone.

3.4. Group 3 (five patients)

Definition of EZ was not possible. Patient 8 felt a tingling sensation in the left arm that preceded NHS; even if sensory aura, interictal and ictal EEG were in agreement and lateralized the EZ to the right hemisphere, we were unable to conclude whether the EZ was in the right insula, right temporal, or right parietal regions. Sometimes the sensory aura was followed by a bilateral buzzing sound that preceded major motor manifestations characterized by rocking trunk movements, wide right hemisoma motion, left dystonic arm posture, and jerking of the left limbs. This patient was classified in group 3, though the association of sensory and auditory aura may reflect the parietal–temporal origin of the discharge. We also lateralized the EZ to the left side in patient 11, but the ictal semeiology, sensory auras, and EEG did not allow the localization of a certain EZ. Patient 6 had interictal EEG showing frontal theta bursts and no ictal, stereo or surgical findings to suggest extratemporal origin. The patient followed all inclusion criteria provided in the methodological section and so we decided to include him, regardless of ictal\inter-ictal EEG findings that are often negative in NFLE cases [1]. The remaining patients had discordant anatomo-electro-clinical data or they did not have information with localizing value. Three patients in group 3 had simple bilateral auditory symptoms.

3.5. Genetic analysis

Five out of 11 patients were screened for CHRNA4, CHRNA2, CHRNB2 mutations and had negative results. Four of these 5 patients were also screened for KCNT1, DEPDC5 and 3 for LG1 with negative findings.

4. Discussion

To our knowledge, this is the first study to explore the relationship between NHS and auditory aura. We found a possible role of auditory aura in localizing an extra-frontal origin of the discharge responsible for hypermotor seizures. The extra-frontal origin of the discharge was documented or highly probable in six out of 11 of our NFLE patients experiencing auditory aura. The EZ could not be localized exactly in the remaining five cases, but an extra-frontal EZ was likely, at least in some cases.

Auditory aura is a rare subjective symptom affecting 1.9% of focal epilepsy patients and is considered a symptom of neocortical temporal lobe epilepsy. The auditory aura occurs when ictal discharges involve the first temporal gyrus (T1): more specifically, the auditory illusions originate from the anterior–middle part of T1 and auditory hallucinations from Heschl's gyrus (middle–posterior part of T1) [27], [28].

Our series of NFLE cases had twice the frequency of auditory aura (6.7%) than that observed in the general population of epilepsy patients [27], [29], thus confirming that the prevalence of auditory aura is probably underestimated [30]. Simple auditory auras were found to be prevalent in groups 1 and 2 where the temporal lobe origin of epileptic discharges was demonstrated or at least suggestive. This characteristic of auditory auras seemed to be more specific to focal discharges arising from the temporal lobe compared with complex auditory symptoms that reflected the involvement of widespread peri-sylvian regions [28]. These data are in agreement with other studies correlating temporal lobe epilepsy to simple auditory aura [27], [31]. It is still debated whether unilateral auditory aura has a lateralizing value. Our sample did not show any predominance in contralateral or ipsilateral aura representation (three versus two) and it was too small to be significant. Florindo et al. studied 154 patients with auditory symptoms, showing that auditory aura had no lateralizing value, but they emphasized that patients with an EZ defined with pre-surgical work-up always had contralateral auditory aura representations. This study also reported a prevalence of dominant hemisphere epilepsy in patients with defined EZ: verbal hallucinations and negative auditory illusions seemed to correlate better with a dominant hemisphere EZ [27]. Interestingly in all our patients in groups 1 and 2 (all right-handed) seizures arose from left/dominant temporal lobe; since most of these patients experienced simple auditory symptoms, the previous association could not be confirmed. Moreover, due to the sample size and the discrepancy between our findings and other studies, a pathophysiological hypothesis could not be formulated to explain the association of auditory aura with dominant lobe epilepsy.

NHS may appear when ictal discharges involve the frontal lobe. In particular, the seizure onset zone in patients with asymmetric dystonic posturing is generally localized in the posterior mesial frontal cortex with a primary involvement of the supplementary motor area, whereas complex motor manifestations (hyperkinetic seizures) appeared when ictal discharges spread to the dorso-lateral frontal cortex, orbito-polar, opercular, and fronto-antero-mesial regions [9], [16].

Our sEEG data documented an extra-frontal onset of the ictal discharge, with a subsequent fast propagation to the cortical area next to the epileptogenic zone and to frontal structures. It is well recognized that the superior temporal gyrus and auditory cortex are highly connected with the frontal lobe. Studies of monkeys and humans have shown anatomical and functional pathways between the superior temporal gyrus and the ventro-lateral-posterior frontal cortex [32], [33], [34]. Studying patients with temporal lobe epilepsy (TLE) by EEG/functional MRI, Fahoum et al. recently observed a preferential propagation of TLE to the frontal lobe. These results may explain the clinical difficulty in differentiating seizures arising from frontal and temporal lobes [35]. By means of diffusor tensor imaging and advanced MRI, Catani and Thiebaut de Schotten indicated the two main pathways connecting the superior temporal gyrus to the ventro-lateral cortex and to the dorso-lateral frontal regions: the arcuate fasciculum and the uncinate fasciculum [36]. Subcortical fasciculi activity could not be investigated during epileptic discharges in our cases, but it may be speculated that the arcuate and uncinate fasciculi play a role in preferential propagation of temporal lobe epilepsy to frontal structures.

FCDs are thought to be the etiological factor responsible for epilepsy in our patients undergoing pre-surgical evaluation. This is in agreement with different series of operated patients reporting a high prevalence of FCD in NHS patients [10], [16], [23], [37]. FCD has also been associated with multiple structurally normal epileptogenic areas connected in an epileptic network [38]. Functional imaging studies have also shown secondary areas remote from the main FCD focus that increase metabolic activity during interictal spiking. A preferential connection of these areas with the EZ was hypothesized, demonstrating that resection of secondary areas influenced the surgical outcome [39]. Studying the relationship between sleep-related epilepsy (SRE) and Taylor's FCD, Nobili et al. suggested that slow-wave sleep could facilitate the spread of epileptic discharges arising from the FCD foci [37], [40]. They found a high prevalence of frontal EZ in SRE, but bivariate analysis proved that the correlation between FCD and SRE was independent from FCD localization [37]. It is therefore possible that the high cortical connectivity of FCD and the interaction between slow-wave sleep synchronization and its peculiar pattern of activity could have a role in the pathophysiology of NHS with frontal and extra-frontal onset.

It is difficult a priori to identify the generator of epileptic discharges responsible for NHS. Besides the frontal lobe, other extra-frontal cortical regions [9], [15], [17] connected preferentially with the frontal and subcortical structures may create an epileptic network. Few clinical symptoms indicate an extra-frontal generator of NHS. Our work further supports the importance of analysing subjective manifestations and suggests that auditory auras may reflect a temporal origin of NHS. These data are clinically relevant, especially when considering patients with drug-resistant NHS for whom surgical treatment could be resolutive.

Funding

This work was supported by the Telethon Foundation (grant GGP 13200 to P.T.).

Conflict of interest

The ICMJE Uniform Disclosure Form for Potential Conflicts of Interest associated with this article can be viewed by clicking on the following link: http://dx.doi.org/10.1016/j.sleep.2014.06.019.

Conflict of interest

ICMJE Form for Disclosure of Potential Conflicts of Interest form.