Epilepsy Research (2014) 108, 327—330
journal homepage: www.elsevier.com/locate/epilepsyres
SHORT COMMUNICATION
Effects of amygdala—hippocampal stimulation on synchronization R. Tyrand a, M. Seeck a, C. Pollo b, C. Boëx a,∗ a b
Department of Neurology, University Hospital of Geneva, Switzerland Department of Neurosurgery, Inselspital, Bern, Switzerland
Received 4 April 2013; received in revised form 17 October 2013; accepted 12 November 2013 Available online 7 December 2013
KEYWORDS DBS; Epilepsy; Temporal lobe
Summary Changes in EEG synchronization, i.e., spatio-temporal correlation, with amygdala—hippocampal stimulation were studied in patients with temporal lobe epilepsy. Synchronization was evaluated for high frequency, 130 Hz, pseudo-monophasic or biphasic charge-balanced pulses. Desynchronization was most frequently induced by stimulation. There was no correlation between the changes in synchronization and the changes in interictal epileptiform discharge rates. Changes in synchronization do not appear yet to be a marker of stimulation efficiency in reducing seizures. © 2013 Elsevier B.V. All rights reserved.
Introduction Deep brain stimulation (DBS) of the amygdalo-hippocampal (AH) structure represents a potential therapeutic technique for patients with intractable epilepsy (Velasco et al., 2007; Boon et al., 2007; Boëx et al., 2011). Certain parameters of stimulation, such as stimulation frequency or waveform, can affect AH-DBS efficiency. These effects can be described by seizure frequency or interictal epileptiform discharge rates (IEDRs) (Boëx et al., 2007; Tyrand et al., 2012). In addition to these neurophysiological markers, changes in synchronization have also been used to analyze epileptic seizure dynamics (Schindler et al., 2007a,b; Jiruska et al., 2013 for a review). Regarding vagal nerve stimulation (VNS) or DBS in
∗
Corresponding author at: Department of Neurology, Hôpitaux Universitaires de Genève, CH-1211 Genève 14, Switzerland. Tel.: +41 (0)79 55 33 841; fax: +41 (0)22372 84 75. E-mail addresses: [email protected], [email protected] (C. Boëx).
epilepsy, different methods for quantifying synchronization have been used, including correlation matrix of electroencephalography (EEG, Schindler et al., 2007c), gamma-band desynchronization expressed with phase-locking (Sohal and Sun, 2011) or phase lag index (Fraschini et al., 2013). The objective of the present study was to evaluate whether changes in synchronization, as assessed via correlation matrix, could be used as a neurophysiological marker of AH-DBS efficiency. We analyzed the changes in synchronization with different waveforms, i.e., pseudomonophasic or biphasic, applied with high frequency (130 Hz) AH-DBS in patients with temporal lobe epilepsy (TLE), to determine if the changes in synchronization could correlate with the effectiveness of stimulation with each waveform.
Methods Patients Twelve patients were enrolled in the study at the time of invasive pre-surgical evaluation (Table 1). Eleven of the
0920-1211/$ — see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.eplepsyres.2013.11.024
328
R. Tyrand et al.
Table 1 Side of stimulation, seizure types and MRI of each patient: left (L); Right (R); complex partial seizure (CPS), secondary generalized seizure (SecGS), generalized tonic seizure (GTS); hippocampus sclerosis (HS), normal MRI (N) hippocampal sclerosis (HS), temporal lobe (TL). Changes in synchronization induced by the high frequency AH-DBS using biphasic or pseudomonophasic charge balanced pluses: desynchronization (D), synchronization (S), equivocal (E) and composite (C). Significant changes: ↑ increase, ↓ decrease, ↔ equivocal changes. Kolmogorov—Smirnov statistics in percent (i.e., maximum distance between empirical distribution functions, p = 0.001) indicated first for the average of the 5 smallest eigenvalues and second for the average of the 2 largest eigenvalues. NC: not conducted. M: channels of the intracranial EEG were not recorded. Patients
Stimulated side
Seizure types
MRI
Biphasic
Pseudomonophasic
S1 S2a S2b S3 S5 S6 S7 S9 S10 S12 S15a S15b S17a S17b S18a S18b
L R
CPS CPS; SecGS
N N
R L L R L L R R
GTS CPS; CPS CPS; CPS; CPS; CPS; CPS
No HS L FHS R HS R periventrical heterotopies L HS; R TL atrophy L HS; Cortical dayplasia R HS; Amygdala atrophy N
R
CPS; GTS
R HS
L
CPS
Neocortical TL atrophy
NC NC NC NC D ↔ ↓ 51 D ↑ 60 ↓ 02 D ↑ 45 ↓ 13 D ↑ 79 ↓ 80 C ↓ 88 ↓ 44 (sleep) S ↓ 40 ↔ S ↑ 81 ↔ D ↑ 100 ↔ D ↔ ↓ 46 D ↑ 11 ↔ E↔↔ D ↑ 48 ↓ 36
D ↑ 32 ↓ 31 D ↔ ↓ 27 D ↑ 52 ↓ 30 S ↓ 67 ↔ NC NC NC M D ↑ 12 ↓ 19 E↔↔ D ↑ 98 ↓ 24 D ↑ 97 ↔ D ↑ 25 ↓ 11 S ↓ 13 ↑ 04 E↔↔ D ↑ 08 ↓ 02
a b
SecGS SecGS SecGS GTS GTC
Phase width of 0.21 ms. Phase width of 0.45 ms.
patients also participated in the study reported by Tyrand et al. (2012). Patient Pt8 was left out of the present study because he was implanted with a single depth electrode which is insufficient to derive synchronization. In addition to the patients described in this previously study, patient S18 was also enrolled in our study (patient characteristics: 59-year-old female; right hand dominancy; drug resistant complex partial seizures; age of epileptic onset: 11 months; left hippocampal sclerosis and left corticosubcortical temporal atrophy; left and right temporal IEDs; intracranial iEEG: left amygdala—hippocampus > right; ictal onset: amygdala and left hippocampus; and stimulation: left hippocampus). The study was conducted according to the recommended ethical guidelines of the Declaration of Helsinki and was approved by the Ethical Committee of the University Hospital of Geneva. All subjects provided informed consent.
Stimulation parameters Synchronization was analyzed on the same group of patients and using the same stimulation parameters as those described by Tyrand et al. (2012). High frequency stimulation, i.e., 130 Hz, was applied for periods of two consecutive hours. At maximum two periods of two consecutive hours of stimulation was applied per day, and at max over two days. Charge-balanced, bipolar electrical stimulation was applied to the estimated epileptogenic zone (i.e., the first contact involved in seizures). Two different waveform stimuli were evaluated: the standard charge-balanced pseudomonophasic (Soletra
neurostimulators, M37021, Medtronic, Minneapolis, MN, USA) and the biphasic charge-balanced pulses (Grass S88X with the opto-isolator SIU-BI, Astro-Med, West Warwick, RI, USA). With the Soletra neurostimulators, the first phase is cathodic (negative), and its width is the one used as a stimulation parameter in the literature. The second phase, anodic (positive), is long width and low amplitude, depending on the cathodic phase width, in order to ensure charge balancing. The biphasic pulses have symmetrical amplitude and duration.
Changes in synchronization Synchronization was calculated from the correlation matrix of normalized iEEG (Müller et al., 2005). This method was chosen because it has been applied in the context of DBS stimulation, allowing comparison to previous results (Schindler et al., 2007c). Phase locking was also used in DBS (Sohal and Sun, 2011), but phase computation is subject to scepticism as EEG recordings are referenced to a common electrode (Guevara et al., 2005). The matrix is an instantaneous correlation of each pair of iEEG channels within a sliding time window. The temporal evolution of the average of the smallest and largest eigenvalues of the correlation matrix was computed (Schindler et al., 2007c). The smallest eigenvalues reflected an evolution that concerned only a subset of channels, whereas the largest eigenvalues reflected an evolution of the entire iEEG. To qualify the changes in synchronization induced by the electrical stimulation, the eigenvalues were compared between those computed from the iEEG of the 2-h
Effects of amygdala—hippocampal stimulation on synchronization stimulation period and those computed from the 1-h iEEG performed immediately prior to stimulation. The results were evaluated statistically using the Kolmogorov—Smirnov test with a significance threshold of p = 0.001. The changes in eigenvalues were categorized as four different states: synchronization (the largest eigenvalues increased, while the smallest eigenvalues either decreased or remained equivocal; or the largest eigenvalues remained equivocal while the smallest eigenvalues decreased), desynchronization (the smallest eigenvalues increased, while the largest eigenvalues decreased or remained equivocal; or the smallest eigenvalues remained equivocal while largest eigenvalues decreased), equivocal (both the smallest and largest eigenvalues remained equivocal), and composite (both the smallest and largest eigenvalues both increased or both decreased). In the present study, the changes in synchronization of the amygdala—hippocampal and fronto-orbital iEEG channels were calculated ipsilaterally to the stimulated zone. EEG, video recordings, and eigenvalue time series were visually inspected, and only periods without artefacts were selected. We assessed the correlation in changes in IEDRs with two underlying synchronization variables: the two Kolmogorov—Smirnov statistics representing the changes in the average of the five smallest eigenvalues and the changes in the average of the two largest eigenvalues.
Results The observed changes in synchronization are reported in Table 1. For the group of 12 patients, 24 periods of stimulation were analyzed (1—4 periods of stimulation per patient were available, Table 1). Desynchronization was the most frequent state induced by high frequency AH-DBS (66.5%; 10 patients with 16 out of 24 periods of stimulation), followed by synchronization (16.5%; 4 patients with 4 periods: S3, S17 with pseudomonophasic; S12, S15 with biphasic), equivocal (12.5%; 2 patients with 3 periods: S12 with pseudomonophasic; S18 with both waveforms) and composite (4.5%; S10 with one period with biphasic). These percentages remained constant when calculated separately for biphasic or pseudomonophasic pulses (desynchronization, 66.0%, in both cases; synchronization, 16.0%, in both cases; equivocal, 8.5% or 17.0%; and composite, 8.5 or 0.0%, respectively, for biphasic or pseudomonophasic pulses). Particular clinical or technical parameters could not be associated in particular to any of the changes in synchronization. There was no significant correlation between the changes in synchronization and the changes in IEDRs for biphasic or pseudomonophasic pulses (Table 2, IEDR changes were calculated for the same set of experiments; for a complete description of the results and discussion regarding IEDRs, refer to Tyrand et al., 2012).
Discussion In the present study, desynchronization was the predominant response to high frequency (130 Hz) AH-DBS regardless of the stimulation pulse waveform, bi-phasic or
329
Table 2 Correlation between changes in IEDR and changes in eigenvalues: Pearson correlation coefficients, r and pvalues.
Overall Biphasic Pseudomonophasic
5 smallest eigenvalues changes vs IEDRs changes r/p
2 largest eigenvalues changes vs IEDRs changes r/p
0.06/0.82 −0.22/0.63 0.44/0.15
0.37/0.11 0.39/0.39 0.07/0.82
pseudo-monophasic. Regarding the rate of stimulation, Schindler et al. (2007c), studying 10 patient using low frequency AH-DBS (1 Hz) has also demonstrated a predominantly desynchronization response. When combined with our results, this shows that changes in AH-DBS frequency do not significantly influence the EEG synchronization response. Regarding the site of stimulation, Sohal and Sun (2011) have also shown phase-locking desynchronization for gamma frequency with high frequency hippocampal stimulation as well as with cortical stimulation. Therefore, desynchronization has been observed with electrical stimulation regardless of the stimulation frequency, the site of stimulation or the pulse waveform. Only long phase duration (i.e., 0.5 ms and higher electrical charges than with shorter phases) was most often observed with synchronization (Schindler et al., 2007c). One recent study has reported gamma-band desynchronization expressed by a decrease in the mean phase lag index induced by vagal nerve stimulation (VNS) exclusively in patients with decreased seizure frequencies (Fraschini et al., 2013). Regarding thalamic stimulation, Zumsteg et al. (2006) did not identify rhythmic synchronization as a reliable neurophysiological marker of stimulation efficiency. Another neurophysiological marker of epileptogenicity, IEDR, has been shown to evolve similarly to seizure frequencies in association with VNS or DBS. For example, low frequency VNS can increase IEDRs, while high frequency reduces IEDRs (Olejniczak et al., 2001). Similarly, low frequency AH-DBS increases IEDRs and produce seizures (Boëx et al., 2007); high frequency AH-DBS can decrease IEDRs and reduce seizure frequency for some patients (Boëx et al., 2011, and for a review). While changes in synchronization were not different here between stimulation-pulse waveforms, IEDRs decreased more often with biphasic stimuli than with pseudomonophasic pulses (Tyrand et al., 2012). While desynchronization has been hypothesized to be associated with the suppression of IEDs (Jaseja, 2010), it has yet to be established as a marker of seizure reduction unlike IEDRs with VNS or DBS. Further assessment of the relationship between the changes in synchronization and the changes in seizure frequency will be necessary to establish synchronization as a clinically viable marker of stimulation effectiveness.
Conflict of interest statement The authors have no conflict of interest to disclose.
330
Acknowledgements This research was supported by the Swiss National Science Foundation (SNF grant no. 32-118385 and 320030-133080). The authors are especially grateful to the patients who agreed to participate in this study.
References Boëx, C., Seeck, M., Vulliémoz, S., Rossetti, A.O., Staedler, C., Spinelli, L., Pegna, A.J., Pralong, E., Villemure, J.G., Foletti, G., Pollo, C., 2011. Chronic deep brain stimulation in mesial temporal lobe epilepsy. Seizure 20 (6), 485—490. Boëx, C., Vulliemoz, S., Spinelli, L., Pollo, C., Seeck, M., 2007. High and low frequency electrical stimulation in non-lesional temporal lobe epilepsy. Seizure 16, 664—669. Boon, P., Vonck, K., De Herdt, V., Van Dycke, A., Goethals, M., Goossens, L., Van Zandijcke, M., De Smedt, T., Dewaele, I., Achten, R., Wadman, W., Dewaele, F., Caemaert, J., Van Roost, D., 2007. Deep brain stimulation in patients with refractory temporal lobe epilepsy. Epilepsia 48, 1551—1560. Fraschini, M., Puligheddu, M., Demuru, M., Polizzi, L., Maleci, A., Tamburini, G., Congia, S., Bortolato, M., Marrosu, F., 2013. VNS induced desynchronization in gamma bands correlates with positive clinical outcome un temporal lobe pharmacoresistant epilepsy. Neuroscience Letters 536, 14—18. Guevara, R., Velazquez, J.L., Nenadovic, V., Wennberg, R., Senjanovic, G., Dominguez, L.G., 2005. Phase synchronization measurements using electroencephalographic recordings: what can we really say about neuronal synchrony? Neuroinformatics 3 (4), 301—314. Jaseja, H., 2010. EEG-desynchronization as the major mechanism of anti-epileptic action of vagal nerve stimulation in patients with intractable seizures: clinical neurophysiological evidence. Medical Hypotheses 74, 855—856.
R. Tyrand et al. Jiruska, P., de Curtis, M., Jefferys, J.G., Schevon, C.A., Schiff, S.J., Schindler, K., 2013. Synchronization and desynchronization in epilepsy: controversies and hypotheses. Journal of Physiology 591 (Pt 4), 787—797. Müller, K., Neumann, J., Lohmann, G., Mildner, T., von Cramon, D.Y., 2005. Detection and characterization of changes of the correlation structure in multivariate time series. Physical Review E: Statistical, Nonlinear, and Soft Matter Physics 71 (4 Pt 2), 046116. Olejniczak, P.W., Fisch, B.J., Carey, M., Butterbaugh, G., Happel, L., Tardo, C., 2001. The effect of vagal nerve stimulation on epileptiform activity recorded from hippocampal depth electrodes. Epilepsia 42, 423—429. Schindler, K., Leung, H., Elger, C.E., Lehnertz, K., 2007a. Assessing seizure dynamics by analysing the correlation structure of multichannel intracranial EEG. Brain 130 (Pt 1), 65—77. Schindler, K., Elger, C., Lehnertz, K., 2007b. Increasing synchronization may promote seizure termination: evidence from status epilepticus. Clinical Neurophysiology 118, 1955—1968. Schindler, K., Elger, C.E., Lehnertz, K., 2007c. Changes of EEG synchronization during low-frequency electric stimulation of the seizure onset zone. Epilepsy Research 77 (2—3), 108—119. Sohal, V.S., Sun, F.T., 2011. Responsive neurostimulation suppresses synchronized cortical rhythms in patients with epilepsy. Neurosurgery Clinics of North America 22, 481—488. Tyrand, R., Seeck, M., Spinelli, L., Pralong, E., Vulliémoz, S., Foletti, G., Rossetti, A.O., Allali, G., Lantz, G., Pollo, C., Boëx, C., 2012. Effects of amygdalahippocampal stimulation on interictal epileptic discharges. Epilepsy Research 99 (1—2), 87—93. Velasco, A.L., Velasco, F., Velasco, M., Trejo, D., Castro, G., Carrillo-Ruiz, J.D., 2007. Electrical stimulation of the hippocampal epileptic foci for seizure control: a double-blind, long-term follow-up study. Epilepsia 48, 1895—1903. Zumsteg, D., Lozano, A.M., Wennberg, R.A., 2006. Rhythmic cortical EEG synchronization with low frequency stimulation of the anterior and medial thalamus for epilepsy. Clinical Neurophysiology 117, 2272—2278.
journal homepage: www.elsevier.com/locate/epilepsyres
SHORT COMMUNICATION
Effects of amygdala—hippocampal stimulation on synchronization R. Tyrand a, M. Seeck a, C. Pollo b, C. Boëx a,∗ a b
Department of Neurology, University Hospital of Geneva, Switzerland Department of Neurosurgery, Inselspital, Bern, Switzerland
Received 4 April 2013; received in revised form 17 October 2013; accepted 12 November 2013 Available online 7 December 2013
KEYWORDS DBS; Epilepsy; Temporal lobe
Summary Changes in EEG synchronization, i.e., spatio-temporal correlation, with amygdala—hippocampal stimulation were studied in patients with temporal lobe epilepsy. Synchronization was evaluated for high frequency, 130 Hz, pseudo-monophasic or biphasic charge-balanced pulses. Desynchronization was most frequently induced by stimulation. There was no correlation between the changes in synchronization and the changes in interictal epileptiform discharge rates. Changes in synchronization do not appear yet to be a marker of stimulation efficiency in reducing seizures. © 2013 Elsevier B.V. All rights reserved.
Introduction Deep brain stimulation (DBS) of the amygdalo-hippocampal (AH) structure represents a potential therapeutic technique for patients with intractable epilepsy (Velasco et al., 2007; Boon et al., 2007; Boëx et al., 2011). Certain parameters of stimulation, such as stimulation frequency or waveform, can affect AH-DBS efficiency. These effects can be described by seizure frequency or interictal epileptiform discharge rates (IEDRs) (Boëx et al., 2007; Tyrand et al., 2012). In addition to these neurophysiological markers, changes in synchronization have also been used to analyze epileptic seizure dynamics (Schindler et al., 2007a,b; Jiruska et al., 2013 for a review). Regarding vagal nerve stimulation (VNS) or DBS in
∗
Corresponding author at: Department of Neurology, Hôpitaux Universitaires de Genève, CH-1211 Genève 14, Switzerland. Tel.: +41 (0)79 55 33 841; fax: +41 (0)22372 84 75. E-mail addresses: [email protected], [email protected] (C. Boëx).
epilepsy, different methods for quantifying synchronization have been used, including correlation matrix of electroencephalography (EEG, Schindler et al., 2007c), gamma-band desynchronization expressed with phase-locking (Sohal and Sun, 2011) or phase lag index (Fraschini et al., 2013). The objective of the present study was to evaluate whether changes in synchronization, as assessed via correlation matrix, could be used as a neurophysiological marker of AH-DBS efficiency. We analyzed the changes in synchronization with different waveforms, i.e., pseudomonophasic or biphasic, applied with high frequency (130 Hz) AH-DBS in patients with temporal lobe epilepsy (TLE), to determine if the changes in synchronization could correlate with the effectiveness of stimulation with each waveform.
Methods Patients Twelve patients were enrolled in the study at the time of invasive pre-surgical evaluation (Table 1). Eleven of the
0920-1211/$ — see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.eplepsyres.2013.11.024
328
R. Tyrand et al.
Table 1 Side of stimulation, seizure types and MRI of each patient: left (L); Right (R); complex partial seizure (CPS), secondary generalized seizure (SecGS), generalized tonic seizure (GTS); hippocampus sclerosis (HS), normal MRI (N) hippocampal sclerosis (HS), temporal lobe (TL). Changes in synchronization induced by the high frequency AH-DBS using biphasic or pseudomonophasic charge balanced pluses: desynchronization (D), synchronization (S), equivocal (E) and composite (C). Significant changes: ↑ increase, ↓ decrease, ↔ equivocal changes. Kolmogorov—Smirnov statistics in percent (i.e., maximum distance between empirical distribution functions, p = 0.001) indicated first for the average of the 5 smallest eigenvalues and second for the average of the 2 largest eigenvalues. NC: not conducted. M: channels of the intracranial EEG were not recorded. Patients
Stimulated side
Seizure types
MRI
Biphasic
Pseudomonophasic
S1 S2a S2b S3 S5 S6 S7 S9 S10 S12 S15a S15b S17a S17b S18a S18b
L R
CPS CPS; SecGS
N N
R L L R L L R R
GTS CPS; CPS CPS; CPS; CPS; CPS; CPS
No HS L FHS R HS R periventrical heterotopies L HS; R TL atrophy L HS; Cortical dayplasia R HS; Amygdala atrophy N
R
CPS; GTS
R HS
L
CPS
Neocortical TL atrophy
NC NC NC NC D ↔ ↓ 51 D ↑ 60 ↓ 02 D ↑ 45 ↓ 13 D ↑ 79 ↓ 80 C ↓ 88 ↓ 44 (sleep) S ↓ 40 ↔ S ↑ 81 ↔ D ↑ 100 ↔ D ↔ ↓ 46 D ↑ 11 ↔ E↔↔ D ↑ 48 ↓ 36
D ↑ 32 ↓ 31 D ↔ ↓ 27 D ↑ 52 ↓ 30 S ↓ 67 ↔ NC NC NC M D ↑ 12 ↓ 19 E↔↔ D ↑ 98 ↓ 24 D ↑ 97 ↔ D ↑ 25 ↓ 11 S ↓ 13 ↑ 04 E↔↔ D ↑ 08 ↓ 02
a b
SecGS SecGS SecGS GTS GTC
Phase width of 0.21 ms. Phase width of 0.45 ms.
patients also participated in the study reported by Tyrand et al. (2012). Patient Pt8 was left out of the present study because he was implanted with a single depth electrode which is insufficient to derive synchronization. In addition to the patients described in this previously study, patient S18 was also enrolled in our study (patient characteristics: 59-year-old female; right hand dominancy; drug resistant complex partial seizures; age of epileptic onset: 11 months; left hippocampal sclerosis and left corticosubcortical temporal atrophy; left and right temporal IEDs; intracranial iEEG: left amygdala—hippocampus > right; ictal onset: amygdala and left hippocampus; and stimulation: left hippocampus). The study was conducted according to the recommended ethical guidelines of the Declaration of Helsinki and was approved by the Ethical Committee of the University Hospital of Geneva. All subjects provided informed consent.
Stimulation parameters Synchronization was analyzed on the same group of patients and using the same stimulation parameters as those described by Tyrand et al. (2012). High frequency stimulation, i.e., 130 Hz, was applied for periods of two consecutive hours. At maximum two periods of two consecutive hours of stimulation was applied per day, and at max over two days. Charge-balanced, bipolar electrical stimulation was applied to the estimated epileptogenic zone (i.e., the first contact involved in seizures). Two different waveform stimuli were evaluated: the standard charge-balanced pseudomonophasic (Soletra
neurostimulators, M37021, Medtronic, Minneapolis, MN, USA) and the biphasic charge-balanced pulses (Grass S88X with the opto-isolator SIU-BI, Astro-Med, West Warwick, RI, USA). With the Soletra neurostimulators, the first phase is cathodic (negative), and its width is the one used as a stimulation parameter in the literature. The second phase, anodic (positive), is long width and low amplitude, depending on the cathodic phase width, in order to ensure charge balancing. The biphasic pulses have symmetrical amplitude and duration.
Changes in synchronization Synchronization was calculated from the correlation matrix of normalized iEEG (Müller et al., 2005). This method was chosen because it has been applied in the context of DBS stimulation, allowing comparison to previous results (Schindler et al., 2007c). Phase locking was also used in DBS (Sohal and Sun, 2011), but phase computation is subject to scepticism as EEG recordings are referenced to a common electrode (Guevara et al., 2005). The matrix is an instantaneous correlation of each pair of iEEG channels within a sliding time window. The temporal evolution of the average of the smallest and largest eigenvalues of the correlation matrix was computed (Schindler et al., 2007c). The smallest eigenvalues reflected an evolution that concerned only a subset of channels, whereas the largest eigenvalues reflected an evolution of the entire iEEG. To qualify the changes in synchronization induced by the electrical stimulation, the eigenvalues were compared between those computed from the iEEG of the 2-h
Effects of amygdala—hippocampal stimulation on synchronization stimulation period and those computed from the 1-h iEEG performed immediately prior to stimulation. The results were evaluated statistically using the Kolmogorov—Smirnov test with a significance threshold of p = 0.001. The changes in eigenvalues were categorized as four different states: synchronization (the largest eigenvalues increased, while the smallest eigenvalues either decreased or remained equivocal; or the largest eigenvalues remained equivocal while the smallest eigenvalues decreased), desynchronization (the smallest eigenvalues increased, while the largest eigenvalues decreased or remained equivocal; or the smallest eigenvalues remained equivocal while largest eigenvalues decreased), equivocal (both the smallest and largest eigenvalues remained equivocal), and composite (both the smallest and largest eigenvalues both increased or both decreased). In the present study, the changes in synchronization of the amygdala—hippocampal and fronto-orbital iEEG channels were calculated ipsilaterally to the stimulated zone. EEG, video recordings, and eigenvalue time series were visually inspected, and only periods without artefacts were selected. We assessed the correlation in changes in IEDRs with two underlying synchronization variables: the two Kolmogorov—Smirnov statistics representing the changes in the average of the five smallest eigenvalues and the changes in the average of the two largest eigenvalues.
Results The observed changes in synchronization are reported in Table 1. For the group of 12 patients, 24 periods of stimulation were analyzed (1—4 periods of stimulation per patient were available, Table 1). Desynchronization was the most frequent state induced by high frequency AH-DBS (66.5%; 10 patients with 16 out of 24 periods of stimulation), followed by synchronization (16.5%; 4 patients with 4 periods: S3, S17 with pseudomonophasic; S12, S15 with biphasic), equivocal (12.5%; 2 patients with 3 periods: S12 with pseudomonophasic; S18 with both waveforms) and composite (4.5%; S10 with one period with biphasic). These percentages remained constant when calculated separately for biphasic or pseudomonophasic pulses (desynchronization, 66.0%, in both cases; synchronization, 16.0%, in both cases; equivocal, 8.5% or 17.0%; and composite, 8.5 or 0.0%, respectively, for biphasic or pseudomonophasic pulses). Particular clinical or technical parameters could not be associated in particular to any of the changes in synchronization. There was no significant correlation between the changes in synchronization and the changes in IEDRs for biphasic or pseudomonophasic pulses (Table 2, IEDR changes were calculated for the same set of experiments; for a complete description of the results and discussion regarding IEDRs, refer to Tyrand et al., 2012).
Discussion In the present study, desynchronization was the predominant response to high frequency (130 Hz) AH-DBS regardless of the stimulation pulse waveform, bi-phasic or
329
Table 2 Correlation between changes in IEDR and changes in eigenvalues: Pearson correlation coefficients, r and pvalues.
Overall Biphasic Pseudomonophasic
5 smallest eigenvalues changes vs IEDRs changes r/p
2 largest eigenvalues changes vs IEDRs changes r/p
0.06/0.82 −0.22/0.63 0.44/0.15
0.37/0.11 0.39/0.39 0.07/0.82
pseudo-monophasic. Regarding the rate of stimulation, Schindler et al. (2007c), studying 10 patient using low frequency AH-DBS (1 Hz) has also demonstrated a predominantly desynchronization response. When combined with our results, this shows that changes in AH-DBS frequency do not significantly influence the EEG synchronization response. Regarding the site of stimulation, Sohal and Sun (2011) have also shown phase-locking desynchronization for gamma frequency with high frequency hippocampal stimulation as well as with cortical stimulation. Therefore, desynchronization has been observed with electrical stimulation regardless of the stimulation frequency, the site of stimulation or the pulse waveform. Only long phase duration (i.e., 0.5 ms and higher electrical charges than with shorter phases) was most often observed with synchronization (Schindler et al., 2007c). One recent study has reported gamma-band desynchronization expressed by a decrease in the mean phase lag index induced by vagal nerve stimulation (VNS) exclusively in patients with decreased seizure frequencies (Fraschini et al., 2013). Regarding thalamic stimulation, Zumsteg et al. (2006) did not identify rhythmic synchronization as a reliable neurophysiological marker of stimulation efficiency. Another neurophysiological marker of epileptogenicity, IEDR, has been shown to evolve similarly to seizure frequencies in association with VNS or DBS. For example, low frequency VNS can increase IEDRs, while high frequency reduces IEDRs (Olejniczak et al., 2001). Similarly, low frequency AH-DBS increases IEDRs and produce seizures (Boëx et al., 2007); high frequency AH-DBS can decrease IEDRs and reduce seizure frequency for some patients (Boëx et al., 2011, and for a review). While changes in synchronization were not different here between stimulation-pulse waveforms, IEDRs decreased more often with biphasic stimuli than with pseudomonophasic pulses (Tyrand et al., 2012). While desynchronization has been hypothesized to be associated with the suppression of IEDs (Jaseja, 2010), it has yet to be established as a marker of seizure reduction unlike IEDRs with VNS or DBS. Further assessment of the relationship between the changes in synchronization and the changes in seizure frequency will be necessary to establish synchronization as a clinically viable marker of stimulation effectiveness.
Conflict of interest statement The authors have no conflict of interest to disclose.
330
Acknowledgements This research was supported by the Swiss National Science Foundation (SNF grant no. 32-118385 and 320030-133080). The authors are especially grateful to the patients who agreed to participate in this study.
References Boëx, C., Seeck, M., Vulliémoz, S., Rossetti, A.O., Staedler, C., Spinelli, L., Pegna, A.J., Pralong, E., Villemure, J.G., Foletti, G., Pollo, C., 2011. Chronic deep brain stimulation in mesial temporal lobe epilepsy. Seizure 20 (6), 485—490. Boëx, C., Vulliemoz, S., Spinelli, L., Pollo, C., Seeck, M., 2007. High and low frequency electrical stimulation in non-lesional temporal lobe epilepsy. Seizure 16, 664—669. Boon, P., Vonck, K., De Herdt, V., Van Dycke, A., Goethals, M., Goossens, L., Van Zandijcke, M., De Smedt, T., Dewaele, I., Achten, R., Wadman, W., Dewaele, F., Caemaert, J., Van Roost, D., 2007. Deep brain stimulation in patients with refractory temporal lobe epilepsy. Epilepsia 48, 1551—1560. Fraschini, M., Puligheddu, M., Demuru, M., Polizzi, L., Maleci, A., Tamburini, G., Congia, S., Bortolato, M., Marrosu, F., 2013. VNS induced desynchronization in gamma bands correlates with positive clinical outcome un temporal lobe pharmacoresistant epilepsy. Neuroscience Letters 536, 14—18. Guevara, R., Velazquez, J.L., Nenadovic, V., Wennberg, R., Senjanovic, G., Dominguez, L.G., 2005. Phase synchronization measurements using electroencephalographic recordings: what can we really say about neuronal synchrony? Neuroinformatics 3 (4), 301—314. Jaseja, H., 2010. EEG-desynchronization as the major mechanism of anti-epileptic action of vagal nerve stimulation in patients with intractable seizures: clinical neurophysiological evidence. Medical Hypotheses 74, 855—856.
R. Tyrand et al. Jiruska, P., de Curtis, M., Jefferys, J.G., Schevon, C.A., Schiff, S.J., Schindler, K., 2013. Synchronization and desynchronization in epilepsy: controversies and hypotheses. Journal of Physiology 591 (Pt 4), 787—797. Müller, K., Neumann, J., Lohmann, G., Mildner, T., von Cramon, D.Y., 2005. Detection and characterization of changes of the correlation structure in multivariate time series. Physical Review E: Statistical, Nonlinear, and Soft Matter Physics 71 (4 Pt 2), 046116. Olejniczak, P.W., Fisch, B.J., Carey, M., Butterbaugh, G., Happel, L., Tardo, C., 2001. The effect of vagal nerve stimulation on epileptiform activity recorded from hippocampal depth electrodes. Epilepsia 42, 423—429. Schindler, K., Leung, H., Elger, C.E., Lehnertz, K., 2007a. Assessing seizure dynamics by analysing the correlation structure of multichannel intracranial EEG. Brain 130 (Pt 1), 65—77. Schindler, K., Elger, C., Lehnertz, K., 2007b. Increasing synchronization may promote seizure termination: evidence from status epilepticus. Clinical Neurophysiology 118, 1955—1968. Schindler, K., Elger, C.E., Lehnertz, K., 2007c. Changes of EEG synchronization during low-frequency electric stimulation of the seizure onset zone. Epilepsy Research 77 (2—3), 108—119. Sohal, V.S., Sun, F.T., 2011. Responsive neurostimulation suppresses synchronized cortical rhythms in patients with epilepsy. Neurosurgery Clinics of North America 22, 481—488. Tyrand, R., Seeck, M., Spinelli, L., Pralong, E., Vulliémoz, S., Foletti, G., Rossetti, A.O., Allali, G., Lantz, G., Pollo, C., Boëx, C., 2012. Effects of amygdalahippocampal stimulation on interictal epileptic discharges. Epilepsy Research 99 (1—2), 87—93. Velasco, A.L., Velasco, F., Velasco, M., Trejo, D., Castro, G., Carrillo-Ruiz, J.D., 2007. Electrical stimulation of the hippocampal epileptic foci for seizure control: a double-blind, long-term follow-up study. Epilepsia 48, 1895—1903. Zumsteg, D., Lozano, A.M., Wennberg, R.A., 2006. Rhythmic cortical EEG synchronization with low frequency stimulation of the anterior and medial thalamus for epilepsy. Clinical Neurophysiology 117, 2272—2278.
