The neural bases of ictal tachycardia in temporal lobe seizures

Accepted Manuscript The neural bases of ictal tachycardia in temporal lobe seizures Florian Chouchou, Romain Bouet, Vincent Pichot, Hélène Catenoix, François Mauguière, Julien Jung PII: DOI: Reference:

S1388-2457(17)30249-3 http://dx.doi.org/10.1016/j.clinph.2017.06.033 CLINPH 2008173

To appear in:

Clinical Neurophysiology

Received Date: Revised Date: Accepted Date:

28 February 2017 3 May 2017 2 June 2017

Please cite this article as: Chouchou, F., Bouet, R., Pichot, V., Catenoix, H., Mauguière, F., Jung, J., The neural bases of ictal tachycardia in temporal lobe seizures, Clinical Neurophysiology (2017), doi: http://dx.doi.org/10.1016/ j.clinph.2017.06.033

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

The neural bases of ictal tachycardia in temporal lobe seizures Florian Chouchou,1 Romain Bouet,2 Vincent Pichot,3 Hélène Catenoix,4 François Mauguière,1,4,5 Julien Jung2,4

1) NeuroPain Lab, Lyon Neuroscience Research Center – Inserm U 1028/CNRS UMR 5292, University of Lyon, France 2) Dycog Lab, Lyon Neuroscience Research Center – Inserm U 1028/CNRS UMR 5292, University of Lyon, France 3) Clinical physiology Department, CHU Nord, Saint-Etienne, France ; EA 4607 SNA-EPIS Lab, University of Jean Monnet, Saint-Etienne, University of Lyon, France 4) Epilepsy and Fonctional Neurology Department, Neurological Hospital Pierre Wertheimer, Hospices Civils de Lyon, Bron, France 5) Claude Bernard Lyon 1 University, Lyon, France

Correspondence:

Florian Chouchou NeuroPain - Central Integration of Pain in Humans Lyon Neuroscience Research Center Hôpital Neurologique, Unité hypnologie, Rdj 59 bvd Pinel, 69 677 Bron cedex France Tel.: +33 4 72 35 78 88

Keywords: Epilepsy, Tachycardia, Autonomic nervous system.

Highlights 

The neural bases of tachycardia that accompanied temporal lobe seizures remain elusive.



The study suggests that hippocampal and amygdalar ictal activity play a pivotal role in tachycardia.



The present study also suggests that ictal tachycardia is independent of ictal insular activity.

2

Abstract

OBJECTIVE: Due to limited information from scalp electroencephalographic (EEG) recordings, brain areas driving changes in cardiac rhythm during Temporal lobe (TL) seizures are not clearly identified. Using stereotactic EEG (SEEG) recordings, we aimed at identifying which of the brain regions involved in autonomic control trigger ictal tachycardia. METHODS: The neural activity of several mesial temporal lobe structures including amygdala, hippocampus, insula, and lateral temporal lobe recorded with SEEG were collected during 37 TL seizures in 9 patients, using indices based on High Frequency Activity (HFA). R-R intervals (RR) monitoring and time-frequency spectral analysis were performed to assess parasympathetic (High frequency power (HF)) and sympathetic (Low frequency/High frequency (LF/HF) ratio) reactivities. RESULTS: Tachycardia was associated with a significant increase in LF/HF ratio and decrease in HF. Autonomic cardiac changes were accompanied by simultaneous SEEG signal changes with an increase in seizure-related HFA in anterior hippocampal formation and amygdala, but not in insula. CONCLUSION: In our sample, TL seizures are thus accompanied by an early decrease in parasympathetic control of cardiac rhythm and by an increase of sympathetic tone, concomitant to seizure activity in anterior hippocampus and amygdala. SIGNIFICANCE: These results support a pivotal role of hippocampus and amygdala in tachycardia occurring during TL seizures.

3

Introduction

Temporal lobe epilepsy (TLE) can lead to intense changes in cardiac autonomic functions during both interictal and ictal periods (Lotufo et al., 2012; Eggleston et al., 2014). Tachycardia, defined as a decrease in the ECG R-R intervals (RR) (increase in heart rate), is the most common cardiac change occurring during epileptic seizures (Opherk and Hirsch, 2002; O’Regan and Brown, 2005; Toth et al., 2010) whereas bradycardia occurs in only 2% of them (Moseley et al., 2010). Tachycardia was proposed as a biomarker providing a somatic indicator of temporal lobe (TL) seizure onset (Osorio, 2014; Schiecke et al., 2014; Osorio and Manly, 2014; Jeppesen et al., 2015; Osorio and Manly, 2015; Behbahani et al., 2016; Van de Vel et al., 2016), opening the way to automatized and noninvasive seizure detection and treatment delivery prior to, or at the onset of, a TL seizure. However, the time course and the neural bases of tachycardia in TL seizures is still incompletely understood. Firstly, the timing of the change in cardiac rhythm relatively to seizure onset is still debated. Previous results are mainly based on scalp EEG, which implies some uncertainty regarding time relationship between brain structures involved in seizure and tachycardia (Weil et al., 2005; Kato et al., 2014). Secondly, the central autonomic network includes several brain regions such as amygdala, hippocampus and insula, which are commonly involved in temporal lobe seizures, (Critchley and Harrison, 2013; Beissner et al., 2013) but the respective role of those three structures in ictal tachycardia is not yet established. Many studies in animals support a major role of the insula in the cardiac autonomic control (Zhang et al., 1998; Oppenheimer and Cechetto, 2016; Marins et al., 2016). In Humans, Oppenheimer and colleagues showed in five epileptic patients that insular stimulations induce tachycardia (Oppenheimer et al., 1992), suggesting a potential role of 4

insula in ictal tachycardia, which remains to be demonstrated by coupling ECG with intracerebral recordings during spontaneous seizures. Thirdly, if some studies suggested that right seizure lateralization is associated with an earlier and higher cardiac response this finding is not consistent across all studies (Massetani et al., 1997; Leutmezer et al., 2003; Di Gennaro et al., 2004; Mayer et al., 2004; Adjei et al., 2009; Kato et al., 2014). Lastly, some animal experiments and human studies suggest that autonomic control is lateralized in the brain, the right hemisphere being involved in sympathetic control and the left one in parasympathetic control (Oppenheimer et al., 1992; Wittling et al., 1998; Zhang et al., 1998); thus an exhaustive evaluation of cardiac changes during temporal lobe seizures should not only assess heart rate, but also autonomic cardiac control, which has been evaluated in very few studies in epilepsy (Sevcencu and Struijk, 2010; Jaychandran et al., 2016). Evaluating RR variability (heart rate variability) is a common noninvasive method for assessing sympathetic and parasympathetic activities well suited to probe the influence of the autonomic nervous system on heart activity during seizures. The power of high-frequency (HF) variations of the RR interval is considered as a marker of parasympathetic control of cardiac rhythm, while that of the low frequency (LF) RR variations and the ratio of low-to-high-frequency powers of RR variations (LF/HF ratio) are markers of the sympathetic tone (Malliani et al., 1991). The main objective of the present study was to assess which cerebral regions, engaged in the central autonomic control including insula, amygdala, and hippocampus (Beissner et al., 2013), are involved in ictal tachycardia. To this end, we analyzed simultaneously intracranial EEG and ECG recordings in 9 patients. The temporal relationships of quantitative stereotactic EEG (SEEG) and RR changes were evaluated. Secondly, we also evaluated autonomic cardiac changes using a time-frequency analysis of heart rate

5

variability (HRV) (Pichot et al., 1999, 2016) to study the influence of sympathetic and parasympathetic systems on ictal tachycardia.

Methods

Patients Nine patients (6 women, mean age 30.4 ± 9.4 (mean ± standard deviation (SD)) years, range [21-50]) who suffered from long lasting partial drug refractory epileptic seizures (range [426] years) were enrolled during presurgical evaluation of their epilepsy. Clinical features of patients are reported in Table 1. Criteria of inclusion were: i) a diagnosis of TLE based on the data from non-invasive investigations with atypical features requiring intracranial SEEG investigation targeting both mesial and lateral temporal lobe structures and insula; ii) the occurrence of tachycardia during spontaneous seizures and iii) a seizure-onset zone in the temporal lobe evidenced by SEEG recordings. All patients underwent a pre-surgical noninvasive investigation, including continuous video-EEG recordings, high resolution Magnetic Resonance Imaging (MRI),

18

Fluorodeoxyglucose Positron Emission Tomography (FDG-PET)

scan, and dipole modeling of epileptic spikes recorded with magnetoencephalography. SEEG was performed because electroclinical data or MRI suggested bitemporal lobe epilepsy (Di Vito et al., 2016) or temporal lobe epilepsy with atypical features such as normal MRI, bilateral hippocampal sclerosis, extratemporal lesion or signs of “temporal plus” epilepsy (Ryvlin et al., 2014). None of the patients had known arrhythmia, cardiac pathology or systemic disease. This study was approved by the local Ethics Committee (CCP Léon BérardLyon). In agreement with French regulations relative to invasive investigations with a direct individual benefit, patients were fully informed about electrode implantation, SEEG and 6

cortical stimulation procedures used to localise the epileptogenic cortical areas. All patients gave their informed consent to the experiments, which were conducted according to the Declaration of Helsinki.

Depth stereotactic EEG recordings (video-SEEG) In order to delineate the extent of the epileptogenic area and to plan a tailored surgical treatment, depth EEG recording electrodes (diameter 0.8 mm; 5-15 recording contacts 2 mm long, inter-contact interval 1.5 mm) were implanted perpendicular to the mid-sagittal plane, according to Talairach’s stereotactic technique (Bancaud and Talairach, 1973). The decision to explore specific areas resulted from the observation during scalp video-EEG recordings of ictal manifestations suggesting the possibility of seizures propagating to, or originating from these regions (Guenot et al., 2001). The number of electrodes implanted per patient varied between 7 and 18 (mean: 13 per patient), with a total number of recording contacts between 90 and 128 per patient.

Recording Procedures SEEG recordings: Data acquisition was performed in the Functional Neurology and Epileptology Department (Lyon Neurological Hospital). Seven to 10 days of continuous SEEG monitoring was performed in the patient’s room. Antiepileptic drugs had been tapered down so that all patients were under mono- or bi-therapy (carbamazepine, valproate, lamotrigine, levetiracetam, or pregabalin). Online recordings (Micromed BrainQuick®, Lyon, France) were obtained using a 128-channel amplified device at a sampling frequency of 256 Hz. For clinical interpretation of SEEG recordings, SEEG data were band-pass filtered at 0.03– 100 Hz. For quantitative SEEG signal analysis, SEEG data were not filtered. The reference 7

electrode was chosen for each patient on an implanted contact located in the skull. SEEG was recorded continuously and was stored for offline analysis. All SEEG signals were rereferenced using bipolar montages between neighbouring contacts of the same electrode. ECG recordings: Two to three electrodes placed on thorax were dedicated to the ECG acquisition. For each patient, we selected the electrode with the highest signal-to-noise amplitude for offline analysis. The acquisition of the ECG signal has been done with the same parameters than SEEG, at a sampling frequency of 256 Hz. ECG was recorded continuously and data were stored for offline analysis.

Data analyses

Anatomical localisation of the recording sites Coordinates of relevant targets were determined on the patient’s brain magnetic resonance images (MRIs) according to previously described procedures (Ostrowsky et al. 2002). The implanted electrodes could be directly visualized on the post-operative 3D-MRIs (3-Tesla Siemens Avanto). An example of depth electrodes localization on T1 MRI sagittal slices is illustrated in Figure 1. In one patient recorded with electrodes that were not MRI compatible, the post-operative 3D-MRI was not available and the exact position of each electrode was verified by fusing post-implantation coronal radiography at scale 1, which precisely delineated the position of each electrode contacts, with the corresponding preimplantation coronal T1 weighted MRI slice. Our analysis was focused on the key structures of cerebral control of cardiac autonomic activity (Beissner et al., 2013): we considered hippocampal (anterior and posterior), amygdalar and insular (anterior, antero-inferior, posterior) contacts. Eight 8

patients underwent a bitemporal SEEG recording with depth electrodes targeting several sites in hippocampus and lateral temporal cortex ipsilateral to the epileptogenic zone and several sites in the contralateral temporal lobe. Implantations in hippocampus were bilateral in 8 patients and unilateral in 1 patient. Implantations in insular cortex were bilateral in 2 patients and lateralized on the epileptogenic side in the 7 others. All amygdalar implantations were unilateral and ipsilateral to the epileptogenic focus. For one patient (patient n°8), all electrodes were implanted in the left hemisphere only, including insula (see Table 2 for details). When several contacts reached the same anatomical region of interest, the contact considered as representative of the region was the one showing the highest fast ictal activity during the seizures (see below).

Autonomic cardiac reactivity analysis

Pre-processing For each patient, we extracted 10 minutes of ECG signal recorded before, during and after the SEEG seizure activity. In order to perform a beat-to-beat analysis of the RR intervals, ECG signals were subjected to peak-to-peak analysis to detect QRS complex (R waves) using a dedicated Matlab® routine (MathWorks, Naticks, MA, USA). Initial automatized extraction of ECG data was subsequently checked by visual inspection, so that undetected QRS, ectopic beats or artifacts were manually corrected. If correction was not possible, a cubic spline interpolation was used to correct for isolated artifacts and ectopic beats (Pichot et al., 2016). If several consecutive beats could not be corrected, the recording was not used for subsequent analysis.

9

Evaluating the time course of tachycardia using automatic detection of autonomic cardiac changes during seizures For each seizure, after visual inspection of our data and according to previous studies (Weil et al., 2005; Calandra-Buonaura et al., 2012; Kato et al., 2014), we divided ictal tachycardia in 3 periods: 1- the attack, marked by a continuous decline in RR (cardiac acceleration), 2the plateau, marked by a RR stabilization at a lower level than pre-tachycardia values, and 3the return, marked by a gradual return to pre-tachycardia RR (see Figure 2). The time boundaries between those three periods were objectively defined using an algorithm inspired from that previously developed by van Elmpt et al. (2006) (see Figure 2A). The onset of ictal tachycardia (attack) was calculated from a basal period (at least 120 s) considered as stable within the 180 s preceding the seizure. The basal period allowed to calculate the standard deviation (SD1) of the basal RR. When RR decreased to a value of at least 2 ˣ SD1 below the mean, a first approximate onset time was defined. To improve the accuracy of the detection of onset time, the RR mean was calculated over the 20 seconds immediately preceding this approximate onset time (pre-tachycardia RR mean), and thus, the (1) ictal tachycardia onset was the time of the first RR below the pre-tachycardia RR mean (Figure 2A). (2) The start of the plateau (and the end of the attack) was defined as the time when the slope of five successive RR reached 0, corresponding to the end of cardiac acceleration (see Figure 2B). (3) The end of the plateau was defined as the time when the slope of RR intervals became positive, considered as the onset of return to the basal level. (4) The end of the tachycardia was the time when RR reached 2 ˣ SD1 below the pre-tachycardia RR mean. Lastly, a recovery period was also defined and consisted of a 30 s time period at least 60 s away from the end of the ictal tachycardia. Thus, at the end of the procedure six periods

10

were defined relatively to RR changes for each studied seizure: a baseline period, a pretachycardia period, the attack, the plateau, the return to baseline, and a recovery period.

Assessing the time course of sympathetic and parasympathetic reactivities using timefrequency analysis of heart rate variability In order to estimate autonomic cardiac modifications over time, a wavelet analysis was applied. Wavelet transform was used to decompose the RR signal in time and frequency domains (Pichot et al., 1999) with MATLAB® 2015 (MathWorks, Natick, MA, USA), using the mother function Daubechies 4. Wavelet analysis was applied on 2.4 Hz re-sampled RR signal (Pichot et al., 2016). Fast frequency in RR signal were gathered in High Frequency power (HF, in ms², 0.15 to 0.4 Hz) to assess parasympathetic reactivity, and in Low Frequency power (LF, in ms², 0.04 to 0.15 Hz) to assess sympathetic reactivity (Pichot et al., 1999, 2016). Because LF is controlled by both the sympathetic and parasympathetic systems, while HF is only controlled by the parasympathetic system, LF/HF ratio was used to assess relative sympathetic activity (Malliani et al., 1991). An example is shown in Figure 2C-D.

Evaluating the brain structures involved during tachycardia and the influence of seizures lateralization using quantitative EEG indices It is well known that epileptogenic brain areas generate high-frequency EEG activity (HFA, typically above 12 Hz, (Weiss et al., 2016)) during seizures. Several studies have shown that it is possible to investigate the epileptogenicity of different brain structures by quantifying the increase of high frequency EEG activity in the seconds following seizure-onset for each intracranial EEG contact. In the present study, we used a quantitative index derived from a method previously described (Bartolomei et al., 2008) to assess the involvement of each 11

brain structure during the course of seizures. For each patient, we analyzed the same 10 minutes of SEEG signals as those used for ECG analysis. For each channel, based on a Fourier transform, we estimated the energy spectral density of the signal over window periods of 5 s sliding by 1 s steps. For each 1s step we extracted the high-frequency signal energy between 12 Hz and 100 Hz (HFA) to which we subtracted the mean pre-tachycardia signal energy in the same frequency range. Thus HFA was a time-varying index estimated before, during and after the seizure representing the increase of high frequency content relatively to a baseline period. Lastly, we averaged the HFA during the 6 periods of the RR response: the basal period, the pre-tachycardia period, the attack, the plateau, the return, and the recovery (see Figure 2). Please note that for each period, we averaged only 13 seconds of HFA because it was the minimal duration common to all of the six periods of interest.

Statistical analyses For all statistical analyses, we used the Statview® software (SAS Institute, Inc., Cary, NC, USA) and the R® environment (R core team, https://www.R-project.org).

Statistical computing of cardiac changes: RR, HF, LF and LF/HF ratio values were normalized in percentage of the basal period values to control for inter-individual variability. For each of the 37 seizures, the values of RR, HF, LF and LF/HF were averaged for each of the 6 periods and the hemispheric lateralization of the seizure was determined (based on visual inspection by two experts, JJ and FM). Then those values were submitted to a repeated-measures analysis of variance (RM-ANOVA) with two within factors: the first factor was Time (6 periods) while the second factor was Seizure

12

lateralization. Newman–Keuls post-hoc test was performed when appropriate. P-values were considered as significant at p<0.05.

Statistical computing of HFA: For statistical computing of HFA, we used linear mixed-effects models (lme4 package, Linear Mixed Effects version 4, (Bates et al., 2015)) to test the influence of different factors on HFA values. This choice was motivated by the fact that there was large inter-individual variability in HFA, number of seizures, and sites of SEEG electrodes across patients (Table 2). To optimize our model, we checked the normality of residual model. Gamma distribution was used to describe data. Data with residual error upper than 2.5 standard deviations were considered as artifacts and removed from analysis. For post-hoc tests we used the Lsmean package (Lsmean version 2.20-23, (Searle et al., 1980)) where p-values were considered as significant at p<0.05 and was adjusted (Tukey method) regarding the number of comparisons performed. Analyses focused on four possible fixed effects: 1) Structure (7 levels: Amygdala, Anterior hippocampus, Posterior hippocampus, Anterior Insula, Inferior Insula, Posterior Insula and Lateral Temporal cortex); 2) Time (6 levels, according to autonomic cardiac changes during seizures); 3) Structure lateralization (2 levels: right and left) and; 4) Seizure lateralization (2 levels: right and left). We also considered the heterogeneity of HFA values across patients with one random effect: Patient. Two types of analyses were performed on HFA. A first analysis was driven mostly to determine the neural structures ipsilateral to the seizure-onset zone involved during tachycardia. In that first analysis, we used only HFA

13

values corresponding to structures ipsilateral to the seizure-onset zone. Two fixed effects were considered: Structure and Time. The second analysis was driven to determine the influence of the lateralization of the seizure and of recorded structures on HFA values. We tested the hypothesis that one hemisphere was more involved in generating ictal tachycardia (Kato et al., 2014). In that second analysis, we used HFA values corresponding both to structures ipsilateral and contralateral to the seizure focus. Four fixed effects were considered: Structure, Time, Structure lateralization and Seizure lateralization. Only structures that reached a statistically significant level in the first analysis were considered in the second analysis.

Results

The time course of ictal tachycardia in Temporal Lobe Epilepsy All of the 37 video-SEEG seizures recorded in the nine patients were analysed (4.1 ± 3.2 seizure by patient). For all recorded seizures, the time course of the tachycardia had a similar profile, with a monophasic course characterized by a progressive decrease of RR intervals (the attack), followed by a sustained a stable RR decline (the plateau) and subsequently a progressive return to baseline values. However, the relative duration of each period varied largely across patients. Using automatic detection of the tachycardia and of its time course, we found that tachycardia lasted 110 ± 65 s (from ictal tachycardia onset to the end of tachycardia). The attack period lasted 18 ± 3 s, the plateau 42 ± 27 s, and the return 51 ± 48 s. (see Figure 2).

Dynamics of RR intervals and RR variability during temporal lobe seizures 14

In order to assess the time course of sympathetic and parasympathetic reactivities during TL seizures, RR, LF, HF and LF/HF ratio were studied during each period of ictal tachycardia and according to seizure lateralization. The time course of RR, LF, HF and LF/HF ratio according to periods of ictal tachycardia and lateralization are presented in Figure 3 and statistical analysis data in Table 3. RM-ANOVA showed that RR, LF, HF and LF/HF ratio varied according to the time but not to the side of TL seizure focus. Post-hoc analysis revealed significant changes in RR from pre-tachycardia to recovery periods (p<0.05): HF power from the attack to the return (p<0.05), LF/HF ratio from plateau to return periods (p<0.05), and at last, LF during return and recovery periods (p<0.05) (see Figure 3). Thus whatever the side of the TL seizure, ictal tachycardia was associated with intense changes in RR variability indexes, with a marked decrease in RR and HF, a continuous increase in LF/HF ratio, and a LF stability during the seizure, whereas during recovery LF remained high and RR low.

Evaluating the brain structures and the side of the epileptic focus involved in tachycardia using quantitative EEG indices To evaluate the brain structures involved during different periods of tachycardia, we studied HFA in several regions of interest, presented in Figure 4. Statistical analysis showed that HFA varied during TL seizure according to the region of interests (see Table 4). Post-hoc analysis showed a significant higher HFA for pre-tachycardia, attack and plateau periods in amygdala (p < 0.0001, p < 0.0001 and p = 0.0001 respectively) and anterior hippocampal structures (p < 0.0001, p < 0.0001 and p < 0.0001 respectively). Thus, only anterior hippocampus and amygdala showed HFA increase during pre-tachycardia, attack and plateau periods. It is noteworthy that inferior, anterior and posterior insula did not show any HFA increase time related to ictal tachycardia. 15

To study if one hemisphere was more involved in generating ictal tachycardia, HFA was computed in the anterior hippocampus that was bilaterally explored in 7/9 patients (see Figure 4B and Table 4). Changes in HFA within anterior hippocampus were independent of the hemisphere and of the seizure lateralization. Moreover, the ictal activity was not different according to the side of epileptic focus, and showed no spread to the other side concomitant to ECG changes. Figure 4B illustrates this effect.

16

Discussion

Temporal lobe seizures are often accompanied by intense tachycardia (Opherk and Hirsch, 2002; O’Regan and Brown, 2005; Toth et al., 2010; Lotufo et al., 2012; Eggleston et al., 2014; Kato et al., 2014). However, the influence of the sympathetic and parasympathetic systems, the time course and the neural bases of tachycardia remain elusive. In the present work, based on direct intracranial recordings of pivotal regions belonging to the central autonomic network such as amygdala, insula and hippocampus (Critchley and Harrison, 2013; Beissner et al., 2013), we observed that, in this sample, ictal tachycardia during TL seizures is: 1) concomitant to an increase in unilateral ictal HF epileptic activity in anterior hippocampus and amygdala; 2) independent of ictal insular activity; 3) driven by a sympathetic reactivity and a decrease in parasympathetic tone; 4) without right-left hemispheric predominance and; 5) independent of the lateralization of seizures. These results suggest that focal seizure activity in anterior hippocampus and amygdala plays a pivotal role in the ictal tachycardia during TL seizures.

Time course of cardiac parasympathetic and sympathetic reactivity during TLE Although several studies quantifying RR variability to detect ictal onset have been carried out (Schiecke et al., 2014; Jeppesen et al., 2015; Behbahani et al., 2016; Van de Vel et al., 2016), our knowledge of sympathetic and parasympathetic dynamics during TL seizures remains poor. Novak and colleagues (Novak et al., 1999) using time-frequency analysis based on Wigner distribution, to study parasympathetic and sympathetic cardiac activity during TL seizures, showed that autonomic reactivity during TL seizures is characterized by parasympathetic decrease and sympathetic predominance, corroborating other studies 17

(Jeppesen et al., 2010; Kolsal et al., 2014). Using a similar time frequency method, the present study confirmed that sympathetic modulation dominates during TL seizures (maintained LF power with increase in LF/HF ratio), with a decrease in parasympathetic activity (decrease in HF power). Furthermore, the present study was able to characterize the temporal course of ictal autonomic activation. The decrease of parasympathetic tone begins at the tachycardia onset within the attack period, reaches its maximum during the plateau period, and returns to basal level after the end of seizures. Conversely, sympathetic reactivity is noticeable during the plateau and return periods. Thus, TL seizures induce a cardiac reflex activation, implying a time-ordered parasympathetic withdrawal followed by a sympathetic drive. The key structures leading to each of these autonomic changes remain to be determined.

The neural bases of ictal tachycardia in temporal lobe seizures and central autonomic network From animal experiments and human studies (Oppenheimer et al., 1992; Benarroch, 1993; Saper, 2002), a central autonomic network, consisting in brain structures generating modulations of autonomic response, has been proposed (Saper, 2002; Critchley and Harrison, 2013). In a human neuroimaging meta-analysis of different somatosensory, cognitive and affective tasks, Beissner and colleagues (Beissner et al., 2013) identified several brain regions belonging to this network including insula, amygdala, hippocampus, hypothalamus, thalamus, anterior and mid-cingulate and ventromedial prefrontal cortices. It may be speculated that ictal functional dysregulation of one or several of these structures mediates ictal tachycardia.

18

On the other hand, several human studies, mostly based on non-invasive EEG recordings, investigated the main features of ictal tachycardia and the lobar localization of seizures with tachycardia. The main findings of these studies are as follows : 1) the frequency of occurrence of seizures associated with tachycardia is similar for generalized and partial onset seizures (Eggleston et al., 2014); 2) temporal seizures are more likely to be associated to tachycardia compared to extra-temporal seizures (Opherk and Hirsch, 2002; O’Regan and Brown, 2005; Toth et al., 2010); 3) right seizure lateralization seems to be associated with an earlier and higher tachycardia (Massetani et al., 1997; Leutmezer et al., 2003; Di Gennaro et al., 2004; Mayer et al., 2004; Adjei et al., 2009; Kato et al., 2014); and 4) the onset of tachycardia can precede by 30s that of the electro- clinical seizure as assessed by scalp EEG recordings (Weil et al., 2005; Eggleston et al., 2014; Kato et al., 2014). These studies support the idea that medial temporal regions play a pivotal role in ictal tachycardia, with a rightsided predominance. However, the spatial resolution of scalp EEG studies is not sufficient to investigate directly the relative roles of the key structures including amygdala, insula and hippocampus. In this context, we observed in our sample that ictal tachycardia is concomitant to an increase in epileptic activity in anterior hippocampus and amygdala, but surprisingly, is independent of ictal insular activity. This suggests that focal epileptic discharges in hippocampus and amygdala play a pivotal role in ictal tachycardia while the insular discharges do not. Although the insula is a cerebral region often involved in the development of TL seizures (Isnard et al., 2000, 2004), our study does not support the view that insula is the trigger of ictal tachycardia associated with TL seizures in spite of our exploration of three sub-regions, inferior, anterior, and posterior, which covered the main cyto-architectonic subdivisions of this structure (Evrard et al., 2014). Conversely, our results pointed out that only seizure activity in anterior hippocampus and amygdala is time-locked 19

to pre-tachycardia, attack and plateau periods of tachycardia during TL seizures. These regions are known to be involved in central autonomic control (Critchley and Harrison, 2013; Beissner et al., 2013) and have physiological and anatomical connections with brain stem structures involved in cardiovascular autonomic regulation, such as the nucleus tractus solitarius (Ruit and Neafsey, 1990; Rowe et al., 1999; Pedemonte et al., 2005; Reyes and Van Bockstaele, 2006). Moreover, seizure activity in hippocampus and amygdala were shown to precede ictal tachycardia in our SEEG exploration, unlike scalp studies, which showed that cardiac changes may precede clinical and EEG manifestations of TL seizures (Weil et al., 2005; Eggleston et al., 2014; Kato et al., 2014). This difference is likely to be related; 1) to the depth of these brain regions where seizure activity is most often missed by scalp recordings and; 2) to the use in our study of objective markers of seizure activity, based on quantified analysis of high frequency activity of SEEG signals. We applied time-frequency analysis of RR intervals to better understand the influence of sympathetic and parasympathetic activities on heart activity during seizures. For that, we quantified ictal parasympathetic reactivity using HF power, based on pharmacological (Akselrod et al., 1981; Pomeranz et al., 1985) and physiological (Malliani et al., 1991) studies showing that only parasympathetic activity influences this spectral power. In our study, changes in HF power occur in the early period of tachycardia and changes in parasympathetic activity can therefore be considered as driving the onset of tachycardia. The study of the sympathetic system is more complex (Rajendra Acharya et al., 2006). Indeed, according to pharmacological studies, the LF power band is both influenced by sympathetic and parasympathetic activities and only the LF/HF ratio is considered as a reliable index to approximate the sympathetic activity (Saul et al., 1990; Malliani et al., 1991; Nakata et al., 1998). In our study, if the HF power decreased, the LF power remained stable 20

and then increased after the TL seizure leading to an increase of the LF/HF ratio reflecting an increase of sympathetic activity during TL seizures. These results show that TL seizures profoundly modify the sympathetic-vagal balance.

The absence of hemispheric lateralization in the genesis of ictal tachycardia The issue of lateralization of the central autonomic network is still debated. Lateralization of autonomic cardiovascular control has been suggested by human studies based on electrical stimulations of the insular cortex (Oppenheimer et al., 1992), functional magnetic resonance imaging (Shoemaker et al., 2012), and clinical evaluation of patients with stroke (Colivicchi et al., 2004). These studies suggested that the cortex of the right hemisphere was critical in sympathetic control but this asymmetry, which could reflect an asymmetrical organization of afferent and efferent of autonomic nervous system (Craig, 2005), was not consistently reported, particularly during epileptic seizures (Opherk and Hirsch, 2002; Moseley et al., 2011; Surges et al., 2013). The rapid spread of TL seizure activity to contralateral hemisphere, missed by surface recordings, could be the reason why a lateralization of discharges producing ictal tachycardia could not be evidenced in scalp EEG studies. Unlike previous studies, our SEEG exploration in our sample showed that ictal tachycardia can occur in conjunction with a right or left seizure activity in hippocampus and amygdala in the absence of seizure propagation to these structures in the contralateral hemisphere.

Perspectives and limitations The main limitation of this work is related to the incomplete and inhomogeneous brain SEEG exploration in a small number of patients. Indeed, we focused on certain areas of the brain whose involvement in the autonomic control is commonly accepted (Critchley and Harrison, 21

2013; Beissner et al., 2013). Thus we cannot discard the possibility that regions unexplored in our study could play a role in ictal tachycardia during TL seizures. Moreover there is a possibility that our incomplete exploration of the insula, with a maximum of three electrodes implanted in this structure in the same patient, could have led us to miss a critical activity in this large region where direct stimulations (Isnard et al., 2004; Mazzola et al., 2017) and functional imaging studies (Kurth et al., 2010) demonstrated multiple functional representations. However, the advantage of the SEEG approach is to permit a precise timing analysis of tachycardia and related cortical dynamics response recorded in situ. Lastly HRV analysis has been demonstrated reliable in assessing autonomic activity (Malliani et al., 1991) but the reliability of this method is discussed regarding the assessment of sympathetic activity because LF is related to both sympathetic and parasympathetic activity. However, sympathetic activity can be determined by examining relative changes in LF and HF as well as normalized indices such as the LF/HF ratio (Malliani et al., 1991; Chouchou and Desseilles, 2014). Several studies have shown that the LF/HF ratio allows approaching relative sympathetic activity in response to tilt testing under atropine (Taylor et al., 2008) and experimental pain (Burton et al., 2009; Chouchou et al., 2011). These findings are consistent with our interpretation of the present results. Whatever the methodological limitations of this study, our results contribute to better understand the neural base of ictal tachycardia in TL seizures. However, it remains that other brain regions should be explored by coupling SEEG with heart rate recordings to have a complete picture of the brain network involved in ictal tachycardia. At last, this work contributes to several clinical perspectives. Firstly, from a surgical point of view, those results suggest that the occurrence of tachycardia during TL seizures should prompt an exploration of hippocampus and amygdala with depth electrodes when 22

invasive recordings are needed to localize the epileptogenic area. On the contrary, tachycardia in isolation (without typical insular semiology) is not a direct biomarker of seizure spread to insula and should not directly misguide exploration with insular depth electrodes implantation. Secondly, detection of seizures using RR variability analysis has potential clinical utilities. Tachycardia was recently proposed as a biomarker providing a somatic indicator of TL seizure onset (Osorio, 2014; Schiecke et al., 2014; Osorio and Manly, 2014; Jeppesen et al., 2015; Osorio and Manly, 2015; Behbahani et al., 2016; Van de Vel et al., 2016). Although scalp studies showed that cardiac changes may precede clinical and EEG manifestations of TL seizures, our work showed that seizure activity in hippocampus and amygdala is concomitant to ictal tachycardia in our sample. Thus if ECG analysis, given its ease of acquisition and processing, could be viewed as a valuable and cost-effectiveness approach, alternative to EEG, for monitoring TL seizures, the possibility that tachycardia could be used as a warning signal of incoming TL seizures is not supported by our study.

Conclusions TL epileptic seizures are often accompanied by intense tachycardia. The present results suggest that focal seizure discharges in hippocampus and amygdala in either of the two hemispheres play a pivotal role in the genesis of ictal tachycardia associated with TL seizures. However, further studies with intracranial EEG and ECG recordings exploring systematically widespread cortical areas are needed to bring a complete view of ictal networks leading to tachycardia in TL and extratemporal seizures.

23

Conflicts of interest

None.

Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

References

Adjei P, Surges R, Scott CA, Kallis C, Shorvon S and Walker MC (2009). Do subclinical electrographic seizure patterns affect heart rate and its variability? Epilepsy Res 87, 281–285. Akselrod S, Gordon D, Ubel FA, Shannon DC, Berger AC and Cohen RJ (1981). Power spectrum analysis of heart rate fluctuation: a quantitative probe of beat-to-beat cardiovascular control. Science 213, 220–222. Bancaud J and Talairach J (1973). [Methodology of stereo EEG exploration and surgical intervention in epilepsy]. Rev Otoneuroophtalmol 45, 315–328. Bartolomei F, Chauvel P and Wendling F (2008). Epileptogenicity of brain structures in human temporal lobe epilepsy: a quantified study from intracerebral EEG. Brain 131, 1818–1830. Bates D, Mächler M, Bolker B and Walker S (2015). Fitting Linear Mixed-Effects Models Usin lme4. J Stat Softw; DOI: 10.18637/jss.v067.i01. Behbahani S, Dabanloo NJ, Nasrabadi AM and Dourado A (2016). Prediction of epileptic seizures based on heart rate variability. Technol Health Care Off J Eur Soc Eng Med; DOI: 10.3233/THC-161225. Beissner F, Meissner K, Bär K-J and Napadow V (2013). The autonomic brain: an activation likelihood estimation meta-analysis for central processing of autonomic function. J Neurosci Off J Soc Neurosci 33, 10503–10511.

24

Benarroch EE (1993). The central autonomic network: functional organization, dysfunction, and perspective. Mayo Clin Proc 68, 988–1001. Burton AR, Birznieks I, Bolton PS, Henderson LA and Macefield VG (2009). Effects of deep and superficial experimentally induced acute pain on muscle sympathetic nerve activity in human subjects. J Physiol 587, 183–193. Calandra-Buonaura G, Toschi N, Provini F, Corazza I, Bisulli F, Barletta G, Vandi S, Montagna P, Guerrisi M, Tinuper P and Cortelli P (2012). Physiologic autonomic arousal heralds motor manifestations of seizures in nocturnal frontal lobe epilepsy: implications for pathophysiology. Sleep Med 13, 252–262. Chouchou F and Desseilles M (2014). Heart rate variability: a tool to explore the sleeping brain? Front Neurosci 8, 402. Chouchou F, Pichot V, Perchet C, Legrain V, Garcia-Larrea L, Roche F and Bastuji H (2011). Autonomic pain responses during sleep: a study of heart rate variability. Eur J Pain Lond Engl 15, 554–560. Colivicchi F, Bassi A, Santini M and Caltagirone C (2004). Cardiac autonomic derangement and arrhythmias in right-sided stroke with insular involvement. Stroke J Cereb Circ 35, 2094–2098. Craig ADB (2005). Forebrain emotional asymmetry: a neuroanatomical basis? Trends Cogn Sci 9, 566–571. Critchley HD and Harrison NA (2013). Visceral Influences on Brain and Behavior. Neuron 77, 624–638. Di Gennaro G, Quarato PP, Sebastiano F, Esposito V, Onorati P, Grammaldo LG, Meldolesi GN, Mascia A, Falco C, Scoppetta C, Eusebi F, Manfredi M and Cantore G (2004). Ictal heart rate increase precedes EEG discharge in drug-resistant mesial temporal lobe seizures. Clin Neurophysiol Off J Int Fed Clin Neurophysiol 115, 1169–1177. Di Vito L, Mauguière F, Catenoix H, Rheims S, Bourdillon P, Montavont A and Isnard J (2016). Epileptic networks in patients with bitemporal epilepsy: the role of SEEG for the selection of good surgical candidates. Epilepsy Res 128, 73–82. Eggleston KS, Olin BD and Fisher RS (2014). Ictal tachycardia: the head-heart connection. Seizure 23, 496–505. Evrard HC, Logothetis NK and Craig ADB (2014). Modular architectonic organization of the insula in the macaque monkey. J Comp Neurol 522, 64–97. Frot M, Faillenot I and Mauguière F (2014). Processing of nociceptive input from posterior to anterior insula in humans. Hum Brain Mapp 35, 5486–5499. Guenot M, Isnard J, Ryvlin P, Fischer C, Ostrowsky K, Mauguiere F and Sindou M (2001). Neurophysiological monitoring for epilepsy surgery: the Talairach SEEG method. StereoElectroEncephaloGraphy. Indications, results, complications and therapeutic 25

applications in a series of 100 consecutive cases. Stereotact Funct Neurosurg 77, 29– 32. Harden C, Tomson T, Gloss D, Buchhalter J, Cross JH, Donner E, French JA, Gil-Nagel A, Hesdorffer DC, Smithson WH, Spitz MC, Walczak TS, Sander JW and Ryvlin P (2017). Practice guideline summary: Sudden unexpected death in epilepsy incidence rates and risk factors Report of the Guideline Development, Dissemination, and Implementation Subcommittee of the American Academy of Neurology and the American Epilepsy Society. Neurology 88, 1674–1680. Isnard J, Guénot M, Ostrowsky K, Sindou M and Mauguière F (2000). The role of the insular cortex in temporal lobe epilepsy. Ann Neurol 48, 614–623. Isnard J, Guénot M, Sindou M and Mauguière F (2004). Clinical manifestations of insular lobe seizures: a stereo-electroencephalographic study. Epilepsia 45, 1079–1090. Jaychandran R, Chaitanya G, Satishchandra P, Bharath RD, Thennarasu K and Sinha S (2016). Monitoring peri-ictal changes in heart rate variability, oxygen saturation and blood pressure in epilepsy monitoring unit. Epilepsy Res 125, 10–18. Jeppesen J, Beniczky S, Fuglsang-Frederiksen A, Sidenius P and Jasemian Y (2010). Detection of epileptic-seizures by means of power spectrum analysis of heart rate variability: a pilot study. Technol Health Care Off J Eur Soc Eng Med 18, 417–426. Jeppesen J, Beniczky S, Johansen P, Sidenius P and Fuglsang-Frederiksen A (2015). Detection of epileptic seizures with a modified heart rate variability algorithm based on Lorenz plot. Seizure 24, 1–7. Kato K, Jin K, Itabashi H, Iwasaki M, Kakisaka Y, Aoki M and Nakasato N (2014). Earlier tachycardia onset in right than left mesial temporal lobe seizures. Neurology 83, 1332–1336. Kolsal E, Serdaroğlu A, Cilsal E, Kula S, Soysal AŞ, Kurt ANÇ and Arhan E (2014). Can heart rate variability in children with epilepsy be used to predict seizures? Seizure 23, 357–362. Kurth F, Zilles K, Fox PT, Laird AR and Eickhoff SB (2010). A link between the systems: functional differentiation and integration within the human insula revealed by metaanalysis. Brain Struct Funct 214, 519–534. Leutmezer F, Schernthaner C, Lurger S, Pötzelberger K and Baumgartner C (2003). Electrocardiographic changes at the onset of epileptic seizures. Epilepsia 44, 348– 354. Lotufo PA, Valiengo L, Benseñor IM and Brunoni AR (2012). A systematic review and metaanalysis of heart rate variability in epilepsy and antiepileptic drugs. Epilepsia 53, 272– 282. Malliani A, Pagani M, Lombardi F and Cerutti S (1991). Cardiovascular neural regulation explored in the frequency domain. Circulation 84, 482–492. 26

Marins FR, Limborço-Filho M, Xavier CH, Biancardi VC, Vaz GC, Stern JE, Oppenheimer SM and Fontes MAP (2016). Functional topography of cardiovascular regulation along the rostrocaudal axis of the rat posterior insular cortex. Clin Exp Pharmacol Physiol 43, 484–493. Massetani R, Strata G, Galli R, Gori S, Gneri C, Limbruno U, Di Santo D, Mariani M and Murri L (1997). Alteration of cardiac function in patients with temporal lobe epilepsy: different roles of EEG-ECG monitoring and spectral analysis of RR variability. Epilepsia 38, 363–369. Mayer H, Benninger F, Urak L, Plattner B, Geldner J and Feucht M (2004). EKG abnormalities in children and adolescents with symptomatic temporal lobe epilepsy. Neurology 63, 324–328. Mazzola L, Mauguière F, and Isnard J (2017). Topical Issue Insular Cortex Epilepsy: from diagnosis to surgery Guest Editor: Pr Dang Nguyen. Electrical stimulations of the human insula: their contribution to the ictal semiology of insular seizures. J Clin Neurophysiol. In Press. Moseley BD, Nickels K, Britton J and Wirrell E (2010). How common is ictal hypoxemia and bradycardia in children with partial complex and generalized convulsive seizures? Epilepsia 51, 1219–1224. Moseley BD, Wirrell EC, Nickels K, Johnson JN, Ackerman MJ and Britton J (2011). Electrocardiographic and oximetric changes during partial complex and generalized seizures. Epilepsy Res 95, 237–245. Nakata A, Takata S, Yuasa T, Shimakura A, Maruyama M, Nagai H, Sakagami S and Kobayashi K (1998). Spectral analysis of heart rate, arterial pressure, and muscle sympathetic nerve activity in normal humans. Am J Physiol 274, H1211-1217. Novak V, Reeves L, Novak P, Low A and Sharbrough W (1999). Time-frequency mapping of RR interval during complex partial seizures of temporal lobe origin. J Auton Nerv Syst 77, 195–202. Opherk C and Hirsch LJ (2002). Ictal heart rate differentiates epileptic from non-epileptic seizures. Neurology 58, 636–638. Oppenheimer S and Cechetto D (2016). The Insular Cortex and the Regulation of Cardiac Function. Compr Physiol 6, 1081–1133. Oppenheimer SM, Gelb A, Girvin JP and Hachinski VC (1992). Cardiovascular effects of human insular cortex stimulation. Neurology 42, 1727–1732. O’Regan ME and Brown JK (2005). Abnormalities in cardiac and respiratory function observed during seizures in childhood. Dev Med Child Neurol 47, 4–9. Osorio I (2014). Automated seizure detection using EKG. Int J Neural Syst 24, 1450001.

27

Osorio I and Manly BFJ (2014). Is seizure detection based on EKG clinically relevant? Clin Neurophysiol Off J Int Fed Clin Neurophysiol 125, 1946–1951. Osorio I and Manly BFJ (2015). Probability of detection of clinical seizures using heart rate changes. Seizure 30, 120–123. Ostrowsky K, Magnin M, Ryvlin P, Isnard J, Guenot M and Mauguière F (2002). Representation of pain and somatic sensation in the human insula: a study of responses to direct electrical cortical stimulation. Cereb Cortex N Y N 1991 12, 376– 385. Pedemonte M, Rodríguez-Alvez A and Velluti RA (2005). Electroencephalographic frequencies associated with heart changes in RR interval variability during paradoxical sleep. Auton Neurosci Basic Clin 123, 82–86. Pichot V, Gaspoz JM, Molliex S, Antoniadis A, Busso T, Roche F, Costes F, Quintin L, Lacour JR and Barthélémy JC (1999). Wavelet transform to quantify heart rate variability and to assess its instantaneous changes. J Appl Physiol Bethesda Md 1985 86, 1081–1091. Pichot V, Roche F, Celle S, Barthélémy J-C and Chouchou F (2016). HRVanalysis: A Free Software for Analyzing Cardiac Autonomic Activity. Front Physiol; DOI: 10.3389/fphys.2016.00557. Pomeranz B, Macaulay RJ, Caudill MA, Kutz I, Adam D, Gordon D, Kilborn KM, Barger AC, Shannon DC and Cohen RJ (1985). Assessment of autonomic function in humans by heart rate spectral analysis. Am J Physiol 248, H151-153. Rajendra Acharya U, Paul Joseph K, Kannathal N, Lim CM and Suri JS (2006). Heart rate variability: a review. Med Biol Eng Comput 44, 1031–1051. Reyes BAS and Van Bockstaele EJ (2006). Divergent projections of catecholaminergic neurons in the nucleus of the solitary tract to limbic forebrain and medullary autonomic brain regions. Brain Res 1117, 69–79. Rowe K, Moreno R, Lau TR, Wallooppillai U, Nearing BD, Kocsis B, Quattrochi J, Hobson JA and Verrier RL (1999). Heart rate surges during REM sleep are associated with theta rhythm and PGO activity in cats. Am J Physiol 277, R843-849. Ruit KG and Neafsey EJ (1990). Hippocampal input to a “visceral motor” corticobulbar pathway: an anatomical and electrophysiological study in the rat. Exp Brain Res 82, 606–616. Ryvlin P, Cross JH and Rheims S (2014). Epilepsy surgery in children and adults. Lancet Neurol 13, 1114–1126. Saper CB (2002). The central autonomic nervous system: Conscious visceral perception and autonomic pattern generation. Annu Rev Neurosci 25, 433–469.

28

Saul JP, Rea RF, Eckberg DL, Berger RD and Cohen RJ (1990). Heart rate and muscle sympathetic nerve variability during reflex changes of autonomic activity. Am J Physiol 258, H713-721. Schiecke K, Wacker M, Piper D, Benninger F, Feucht M and Witte H (2014). Time-variant, frequency-selective, linear and nonlinear analysis of heart rate variability in children with temporal lobe epilepsy. IEEE Trans Biomed Eng 61, 1798–1808. Searle SR, Speed FM and Milliken GA (1980). Population Marginal Means in the Linear Model: An Alternative to Least Squares Means. Am Stat 34, 216–221. Sevcencu C and Struijk JJ (2010). Autonomic alterations and cardiac changes in epilepsy. Epilepsia 51, 725–737. Shoemaker JK, Wong SW and Cechetto DF (2012). Cortical Circuitry Associated With Reflex Cardiovascular Control in Humans: Does the Cortical Autonomic Network “Speak” or “Listen” During Cardiovascular Arousal. Anat Rec Adv Integr Anat Evol Biol 295, 1375– 1384. Surges R, Jordan A and Elger CE (2013). Ictal modulation of cardiac repolarization, but not of heart rate, is lateralized in mesial temporal lobe epilepsy. PloS One 8, e64765. Taylor EM, Boyer K and Campbell FA (2008). Pain in hospitalized children: a prospective cross-sectional survey of pain prevalence, intensity, assessment and management in a Canadian pediatric teaching hospital. Pain Res Manag 13, 25–32. Toth V, Hejjel L, Fogarasi A, Gyimesi C, Orsi G, Szucs A, Kovacs N, Komoly S, Ebner A and Janszky J (2010). Periictal heart rate variability analysis suggests long-term postictal autonomic disturbance in epilepsy. Eur J Neurol 17, 780–787. Van de Vel A, Cuppens K, Bonroy B, Milosevic M, Jansen K, Van Huffel S, Vanrumste B, Cras P, Lagae L and Ceulemans B (2016). Non-EEG seizure detection systems and potential SUDEP prevention: State of the art: Review and update. Seizure 41, 141–153. Weil S, Arnold S, Eisensehr I and Noachtar S (2005). Heart rate increase in otherwise subclinical seizures is different in temporal versus extratemporal seizure onset: support for temporal lobe autonomic influence. Epileptic Disord Int Epilepsy J Videotape 7, 199–204. Weiss SA, Alvarado-Rojas C, Bragin A, Behnke E, Fields T, Fried I, Engel J and Staba R (2016). Ictal onset patterns of local field potentials, high frequency oscillations, and unit activity in human mesial temporal lobe epilepsy. Epilepsia 57, 111–121. Wittling W, Block A, Schweiger E and Genzel S (1998). Hemisphere asymmetry in sympathetic control of the human myocardium. Brain Cogn 38, 17–35. Zhang ZH, Rashba S and Oppenheimer SM (1998). Insular cortex lesions alter baroreceptor sensitivity in the urethane-anesthetized rat. Brain Res 813, 73–81.

29

Legends

Table 1. Demographic and clinical features of the 9 patients suffering from TLE.

Table 2. Summary of implanted brain structures for all patients. R = Right; L = Left. For each structure, X indicates that the brain structure has been explored by depth electrodes.

Table 3. Statistical results of repeated-measures analysis of variance for RR intervals, low frequency (LF), high frequency (HF) power and LF/HF ratio of RR intervals according to time, seizure lateralization and interactions between those effects.

Table 4. Statistical results of linear mixed-effects models for high frequency SEEG activity (HFA) according to time, anatomical structures, seizure focus lateralization and structure lateralization, and interactions between these effects.

Figure 1: Example of SEEG recordings of intracerabral and cardiac activity during a right TL seizure. Up: Localization of depth electrodes on T1 MRI slices. Black dots correspond to the artifact created by the electrode contacts. All depth electrodes implanted in the right hemisphere are visible on the sagittal slice (1). Coronal views of electrodes implanted in: right amygdala (2), bilateral hippocampus (3), right insula and left hippocampus (4). Down: Raw EEG activity recorded by contacts located in hemisphere ipsilateral to the seizure focus (right frontal lobe, insula, amygdala, hippocampus and lateral temporal neocortex) and 30

in the contralateral hemisphere (Contralateral). The bottom trace is the electrocardiogram. Part I shows the onset of the ictal EEG discharge. The low-voltage fast discharge at seizure onset is restricted to the amygdala (arrow). Part II shows the involvement of the hippocampus with a low-voltage fast activity (arrow). Part III shows the propagation of the ictal activity to the frontal lobe, the temporal neocortex and the insula. The contralateral propagation is moderate and late. Note that the tachycardia (arrow) begins just after the involvement of the hippocampus and before propagation to insula. Initial automatized extraction of ECG data was subsequently checked by visual inspection, so as undetected QRS, ectopic beats or artifacts were manually corrected.

Figure 2: Example of ECG signal analysis for one seizure. Representative examples of A) R-R intervals in ms (RR), B) Low frequency (LF, red) and high (HF, green) frequency variations of RR intervals in ms 2, C) LF/HF ratio of RR intervals before, during and after ictal tachycardia; D) Slope of RR variations used to determine the start and end of the plateau. To identify each step of the ictal tachycardia based on RR intervals, cardiac reactivity was divided in main 3 periods: the attack, marked by the continuous decline in RR (cardiac acceleration), the plateau, marked by the RR stabilization, and the return, marked by a gradual return to the baseline RR. The time boundaries between the different periods of tachycardia are marked by vertical blue lines in the figure: line 1 indicates the ictal tachycardia onset; line 2 indicates the start of the plateau; line 3 indicates the end of the plateau; line 4 indicates the end of tachycardia. Please see the methods for details on time boundaries determination.

31

A basal period (at least 120 s) considered as stable within the 180 s preceding the seizure and a 30 s recovery period were also analyzed, which are illustrated by the grey shaded areas before and after the tachycardia (for details, see methods). In A, B and C the 6 analyzed periods are marked by grey shaded areas (a- baseline, b- pretachycardia, c- attack, d- plateau, e- return, f- recovery). This representative example illustrates the cardiac reflex activation induces by temporal lobe seizures, implying a time-ordered parasympathetic withdrawal followed by a sympathetic drive.

Figure 3. Summary of RR intervals ad RR variability during the course of ictal tachycardia across the 9 patients. A) RR intervals, B) low (LF) and (C) high frequency (HF) powers, and (D) LF/HF ratio according to six defined basal, pre-tachycardia, attack, plateau, return and recovery periods (mean ± SD) and according to the lateralization of seizures (green : right hemisphere seizures, red : right hemisphere seizure). Data for RR, LF, HF and LF/HF are normalized in percentage relatively to the basal period values. For each period of tachycardia ▪ symbols refer to significant (p<0.05) differences with baseline period (for details, see results). Temporal lobe seizures induce a cardiac reflex activation, implying a time-ordered parasympathetic withdrawal followed by a sympathetic drive.

Figure 4. Variations of SEEG High Frequency Activity (HFA) during the course of ictal tachycardia across the 9 patients. 32

In A) HFA are presented for all analyzed structures and in B) HFA in anterior hippocampus are presented according to seizures and hippocampus lateralization, during basal, pretachycardia, attack, plateau, return and recovery periods (HFA value, High Frequency activity value, mean ± SD). For each period, HFA values are positive if high frequency activity is above the mean basal values and negative if high frequency activity is below the mean basal values. For each period of tachycardia ▪ symbols refer to significant (p<0.05) differences with baseline period (for details, see results). Graphs in A) show that HFA are significantly higher than baseline in amygdala and anterior hippocampus during the pre-tachycardia, attack and plateau periods ▪: p<0.05 for anterior hippocampus; ▪: p<0.05 for amygdala. Graphs in B) show that HFA varies significantly with time in hippocampus (p < 0.001) with structures x seizures lateralization (p < 0.001) and time x structures x seizures lateralization (p < 0.001). See results for details.

33

Table 1

PATIENTS

GENDER, AGE

EPILEPSY ONSET (YEARS) 26

PATIENT 1

M, 32

PATIENT 2

F, 37

PATIENT 3

F, 22

PATIENT 4

M, 26

PATIENT 5

F, 27

PATIENT 6

M, 21

PATIENT 7

F, 50

PATIENT 8 PATIENT 9

F, 23

4

F,36

23

8

8

11

16

8

17

ICTAL SEMIOLOGY

MRI

Pallor, stop activity, loss of consciousness

Right hippocampal sclerosis and right anterior frontal flair hypersignal Right ventricular dilatation and posterior cortical flair hypersignal Right hemispheric atrophy and hippocampal sclerosis Normal

Elementary visual hallucination, oroalimentary automatisms, left arm dystonic Tachycardia, pallor, cold sensation Cephalic aura, abdominal constriction, hypersalivation, nausea, laryngeal striction, vomiting. Postictal aphasia Loss of consciousness, amnesia, postictal confusion Left oculo-cephalic version, left facial clonus, secondary generalization Epigastric aura, oroalimentary automatisms, left arm dystonic, postictal aphasia Epigastric aura, oroalimentary automatisms, postictal aphasia Mutism, stop activity

N°AND SIDE OF SEEG ELECTRODES

EPILEPTOGENIC ZONE

SURGERY

ENGEL CLASS

FOLLOW-UP (MONTHS)

12R, 2L

Right mesiotemporal structures

Anterior temporal lobectomy

1a

26

11 R, 2L

Right temporooccipital region

No

13R, 1L

Right mesiotemporal structures

Anterior temporal lobectomy

1a

48

10L, 2R

Left amygdala and temporal pole

Amygdalectomy and polectomy

1a

69

Anterior temporal lobectomy

2b

52

Anterior temporal lobectomy

2b

45

Temporal polectomy

2a

21

Normal 11R, 3L Right temporal cortical malformation Left parietal cavernoma, left hippocampal sclerosis Bilateral hippocampal sclerosis Right periventricular heterotopy

11R, 3L

8L, 5R 9L, 1R 12L, 3R

Right mesiotemporal structures Right posterior temporal neocortex Left temporomesial structures Left temporomesial structures Left anterior temporal neocortex

No

No

Table 2 Patients

Seizure’s side

S. 1

R

S. 2 S. 3 S. 4 S. 5 S. 6

R R L R R

S. 7

L

S. 8

L

S. 9

L

Total

Struture’ side

Ant. Hippocampus

L

X

R

X

L

X

R

X

L

X

R

X

L

X

R

X

L

X

R

X

X

L

X

X

R

X

X

L

X

X

R

X

X

L

X

X

X

X

X

X

X

X

4 3

4 3

Post. Hippocampus

Ant. Insula

Ant-Inf Insula

Post. Insula

Amygdala

X

X

Lat. temporal X

X

X X

X

X

X X

X X

X

X

X

X

X

X

X

X

X

X X X

X X

X

X X

X X

X

X X

X

X

X

R L R L R

X

X 9 7

3 4

X 3

3

9

4

3

8

Table 3 Effect

t

p

104.5

< 0.001

Seizure lateralization

0.1

0.714

Interaction

2.0

0.084

Time

2.0

< 0.001

Seizure lateralization

0.1

0.745

Interaction

0.2

0.944

Time

3.1

0.010

Seizure lateralization

1.0

0.314

Interaction

0.2

0.954

Time

9.0

< 0.001

Seizure lateralization

0.1

0.801

Interaction

0.2

0.954

RR intervals Time

HF power

LF power

LF/HF ratio

Table 4 Effect

t

p

Time

59.9

< 0.001

Structure

23.3

< 0.001

Structure*Time interaction

2.7

< 0.001

Time

59.9

< 0.001

Seizure focus lateralization

0.7

0.742

Structure lateralization

0.2

0.690

Time*Seizure focus lateralization interaction

1.3

0.417

Time* Structure lateralization interaction

1.5

0.092

Seizure focus lateralization*Structure lateralization interaction

31.4

Time*Seizure focus lateralization*Structure lateralization interaction

4.4

< 0.001 < 0.001

37

Abstract

OBJECTIVE: Due to limited information from scalp electroencephalographic (EEG) recordings, brain areas driving changes in cardiac rhythm during Temporal lobe (TL) seizures are not clearly identified. Using stereotactic EEG (SEEG) recordings, we aimed at identifying which of the brain regions involved in autonomic control trigger ictal tachycardia. METHODS: The neural activity of several mesial temporal lobe structures including amygdala, hippocampus, insula, and lateral temporal lobe recorded with SEEG were collected during 37 TL seizures in 9 patients, using indices based on High Frequency Activity (HFA). R-R intervals (RR) monitoring and time-frequency spectral analysis were performed to assess parasympathetic (High frequency power (HF)) and sympathetic (Low frequency/High frequency (LF/HF) ratio) reactivities. RESULTS: Tachycardia was associated with a significant increase in LF/HF ratio and decrease in HF. Autonomic cardiac changes were accompanied by simultaneous SEEG signal changes with an increase in seizure-related HFA in anterior hippocampal formation and amygdala, but not in insula. CONCLUSION: In our sample, TL seizures are thus accompanied by an early decrease in parasympathetic control of cardiac rhythm and by an increase of sympathetic tone, concomitant to seizure activity in anterior hippocampus and amygdala. SIGNIFICANCE: These results support a pivotal role of hippocampus and amygdala in tachycardia occurring during TL seizures.

38