22nd Annual Scientific Meeting of the Society of Thailand Main auditorium of Prasart Neurology Institute Bangkok, July 19th and 20th, 2018 Disclosure Form � Company Name Nature of Affiliation Epileptogenesis, and : � • Eisai, Nihon-Kohden, Otsuka, is it a paradigm shift? Endowed Department UBC Japan •

Akio IKEDA, MD, PhD • Eisai, UBC Japan, Otsuka, • Grants or Honorariums

Department of Epilepsy, Movement Disorders & Physiology Off-Label Product Usage

Kyoto University Graduate School of Medicine • None Kyoto, JAPAN

Schema of potentials Application of LFF and HFF � Time constant (TC) Slow potential shifts ictal DC shifts to filter out slow components BP(MRCP) DC EEG CNV: AC potentials Nerofeedback (NFB) � Slow EEG� Slow cortical potentials by VNS� �

EMG

LFF HFF� 0.3 Hz 70 Hz TC x LFF=0.16(1/2π) Steady potentials (DC potentials) (Sano, 1968) (TC=0.5 s) 0.1 sec 1.6 Hz � 1 sec 0.16 Hz 10 sec 0.016 Hz DC amplifier or long time constant(10sec) of AC amplifier needed�

Generalized spike and wave complex had also slow shifts�

ictal DC shifts:�

(Migraine) Nerofeedback (NFB) BP(MRCP) Slow cortical potentials CNV� Induced by VNS� ()�

(Kovac, Speckmann, Gorge et al., Prog. Neurobiol. 2017) “The lack of any detectable epileptiform pattern in the EEG of some patients in the course of focal activity is often adduced as an example of the limitations of EEG.” �

(Chatrian et al.,1964)�

The first report of PLEDs with epilepsia partialis continua

Periodic focal slow wave with described in 2 cases Time constants (TC) of 1 sec was employed, tin electrodes used

These slow transients may be the expression of local shifts TC=1 sec � of the steady potential (SP) of the .

Ictal DC shift preceded HFO in invasive recording. Wide-band EEG in clinical invasive recording (Imamura et al.,2011) (Kanazawa et al. 2015)(Daifu et al., 2016) 500Hz TC=10sec HFO� Neurons� 0Hz � DC shifts� Glia� 10 sec (Ikeda et al, Epilepsia,1996) HFO=high frequency oscillation� (Worrell et al., 2008)� (Worrell et al., 2008)�

(Ikeda et al., 1996) (Ikeda et al., 1996) 0.1Hz 1Hz 100Hz 1000Hz 0.1Hz 1Hz 100Hz 1000Hz

Hypothesis of active vs. passive DC shift in the ictal period , glia, and epilepsy: is it a paradigm shift?

onset 1) Glia vs. neurons, i.e., active- vs. passive DC shifts ・Conventional ictal pattern time ・HFO AMED study in Japan Surgical outcome

・DC shift 2) Interictal red slow, i.e., co-occurrence of slow and HFO ③ ② ① time 1960s� (drawn by the help of Murai) ③Active DC shift ② ① Passive DC shift 3) 2 types of ictal DC shifts endorsed by C10 sec vs. TC 2sec Clinical state Chronic epilepsy Acute symptomatic seizure T Clinical patients Habitual Non-habitual seizures 4）Is it recorded by scalp EEG ? Basic animals Chronic model Acute model Chronic pilocarpine model Acute pilocarpine model LGi1 rat model 5) Proposal of clinical practice parameter for recording and analysis of NER(=Noda epilepsy rat model) ictal DC shifts and HFO: invasive EEG (by Ikeda et al) (as the working hypothesis of AMED research team of Epilepsy and Glia, chaired by Maehara, Japan, 2015-2018) Expression of epileptogenicity by EEG Terminology: Ictal DC (direct current) shifts

1) By conventional EEG Also described as very slow, infra-slow, steady, spikes, sharp waves (pyramidal neurons)

Recorded by 2) By wide-band EEG (surrogate marker?) DC amplifier DC shifts DC shifts, slow shifts (pyramidal neurons, glia) AC (alternative current) amplifier Slow shits HFO or fast ripple activity (pyramidal neurons, long time constant, i.e. 10 sec ?) small low frequency filter (LFF) i.e., 0.016Hz

Hypothesis of active vs. passive DC shift in the ictal period onset Ictal activity by pentylentetrazole ・Conventional ictal pattern time ・HFO （acute animal model）� EEG�

・DC shift ③ ② ① time (drawn by the help of Murai) ③Active DC shift ② ① Passive DC shift Clinical state Chronic epilepsy Acute symptomatic seizure Clinical patients Habitual seizures Non-habitual seizures DC / EEG� Basic animals Chronic model Acute model Chronic pilocarpine model Acute pilocarpine model passive DC shifts?� (Speckmann & Elger, 1987) LGi1 rat model NER(=Noda epilepsy rat model)

(by Ikeda et al) (as the working hypothesis of AMED research team of Epilepsy and Glia, chaired by Maehara, Japan, 2015-2018)

EEG and intracellular recording in interictal and ictal periods Amplifying effects of glial cells on slow shifts (Ayala et al., 1973)

Glia Glia�

Intracellar potentials Neuron Neuron 2 Passive� 1 Passive� Glia Glia (Dysfunction of in CD, (Kuffler et al., 1966)� Ictal, sustained PDS Bordey et al., J Neurophysiol, 2001) Impaired K+ homeostasis is important both after seizure and � before seizure, that may triggger seizure?� Possible generator mechanism of Ictal DC shifts 3 Active (?)� Active vs. Passive mechanism

Passive generator mechanisms 1) Sustained paroxysmal depolarization shift (PDS) occurring in the epileptic neurons (1960s)

2) Augmented passive depolarization by increased [K+] of glia as the functional syncytium, associated with its dysfunction to salvage increased [K+] following seizure (1960s, 1990s)

Active generator mechanism 3) Active roles of glia: spontaneous slow oscillation of glia itself and decreased Kir4.1 channel in the astrocytes (after 2000) (Ohno et al, 2015)�

functional syncytium� （機能的合胞体） �

tripartite � Devinsky et al., 2013� (3者間シナプス）�

Subdurally recorded ictal EEG in Patient 1 TC=10sec, Focal, ictal slow shifts � �

(Ikeda et al, Brain,1999) Lt lateral epilepsy(oligodendroglioma) Subdural recording of ictal DC shifts (Lt. lateral TLE)� (T.C.= 0.1sec)� (T.C.=10 sec)� (T.C.=10sec)-35sec� -25sec -3sec (to clinical onset)�

A6� A6

A11 A11

B1,2

B4 B5� B5

B12,13 B12,13

(Ikeda et al.,Eplepsia,1996) (Ikeda et al, Epilepsia,1996)

Ictal DC shifts (invasive recording): summary

DC shifts at 1, 6� 1) Ictal DC shifts recorded by invasive electrodes, especially subdural ones, in humans were almost invariably recorded regardless of underlying etiology or epilepsy type.

2) 96 % of patients showed ictal DC shifts, incidence rate being 42~100% (87%) of seizures in each patient.

2) Its more restricted localization could aid in delineating ictal onset zone clinically before surgery presumably as a Clinical onset (aura) core epileptogenic zone, if present.

Conventional EEG� Sampling rate＝2kHz, TC=10s, time window=20,30,60sec, 2m, 3m, � HFF=300,1Hz �

TC 0.1 sec Conventional initial change

• (Imamura et al, 2011� A19 A20

B13

B18

D1 D2 D3

(Imamura et al, 2011）� Time frequency analysis on the EEG review station (+-60sec) Ictal DC shifts occurred earlier than or as early as ictal HFO. Collaboration with Nihon Kohden （HFO up to 600Hz and DC shifts shown together in the review station)� � � � DC onset plotted as compared with HFO onset at 0 sec.�

*=P<0.05�

Alligned iron sand � by the magnetic � field.� � Magnetic field=� modified extracellular� circumstances� (Kanazawa et al., Clin Neurophysiol, 2015)�

Occurrence rate of ictal DC shifts and HFO Study group of Epilepsy and Glia by AMED (Japan Agency for Medical Research and Development)

DC shifts HFO Clinical approach= multi-institutional study of DC shifts in epilepsy surgery Ikeda et al,1996,1999 96% of pts Maehara T,MD (Tokyo), Inoue Y, MD (Shizuoka), Shirozu H, MD 87% of sz (Niigata), Watanabe Y, MD (Tokyo), Ikeda A, MD (Kyoto)

Pathology Imamura et al, 2010 100% of sz 100% of sz Kakita Y, MD (Niigata) Wu et al, 2014 94% of sz 84% of sz Basic Kanazawa et al, 2015 75% of pts 50% of pts Ohno Y, PhD (Osaka), Okada M, MD (Mie): pharmacology 72% of sz 47% of sz Fukuda A, PhD (Hamamatsu), Koizumi S, PhD (Yamanashi): in Nakatani (AMED) 95% of sz 77% of sz vivo and in vitro analysis (2017)

Neuron, glia, and epilepsy: is it a paradigm shift? Irritative zone vs. epileptogenic zone� Red spike? Green spike?� 1) Glia vs. neurons, i.e., active- vs. passive DC shifts Interictal, � AMED study in Japan Surgical outcome

2) Interictal red slow, i.e., co-occurrence of slow and HFO Ictal, Ictogenesis� 3) 2 types of ictal DC shifts endorsed by TC10 sec vs. TC 2sec

4）Is it recorded by scalp EEG ?

5) Proposal of clinical practice parameter for recording and analysis of Red slow? (epileptic slow)� ictal DC shifts and HFO: invasive EEG (Lüders, 1992)� +5 Top：power spectrum +5 Co-occurrence of interictal slow &HFO B07 electrode Fig. 2. middle：ECoG(TC:0.3 sec, HFF:300 Hz) CJ3258YA.EEG −5 CJ3258YA.EEG −5 300 250300 Hz boom：ECoG(TC:10 sec, HFF:10 Hz) Spectrogram B07_REF01 200 B7: ictal focus 300 150 250 100 B07_REF01 200 50 150 100 TC: 0.3 sec 1 50 B07_REF01 +5+5 80secB07_REF01 HFF: 300 Hz

TC: 10 sec B07_REF01

B07_REF01

HFF: 10 Hz

CJ3258YA.EEG −-5 5

3570 3580 3590 3600 3610 3620 3630 3640 3570 3580 3590 3600 3610 3620 3630 3640 B07_REF01 B07_REF01 300 Hz B07_REF01 80secB07_REF01 +5

B07_REF01

B07_REF01

500

CJ3258YA.EEG −5 3650 3660 3670 3680 3690 3700 3710 3720 3650 3660 3670 3680 3690 3700 3710 3720 HFOs onset, convenonal inial change B07_REF01 400 B07_REF01 B07_REF01

B07_REF01 B07_REF01 300

80sec B07_REF01

B07_REF01

Seizure 200

3730 3740 3750 3760 3770 3780 3790 3800

500 3730 3740 3750 3760 3770 3780 3790 3800 DC shi onset Seizure B07_REF01 100 B07_REF01 400 B07_REF01 B07_REF01

B07_REF01

B07_REF01 B07_REF01 300

80sec

3810 3820 3830 3840 3850 3860 3870 3880 200 3810 3820 3830 3840 3850 3860 3870 3880 B07_REF01 B07_REF01 100 − B07_REF01 B07_REF01 500 μV B07_REF01 + B07_REF01 − 80secB07_REF01 500 μV 10 sec 500 μV + 3890 3900 3910 3920 3930 3940 3950 396010 sec 3890 3900 3910 3920 3930 3940 3950 3960 7200.1 (Inoue T et al., Annual congress of JCNS, 2014)� 7200.1 B07_REF01

B07_REF01

B07_REF01

Interictal slow shifts by macroinvasive electrodes Patient 1 (D1: the most core electrode within epileptic focus) Both interictal and ictal7200.1 500 Hz Interictal state may resemble HFO each other qualitavely Red slow?� 7200.1 0 D1 from the viewpoint of 500 µV slow/DC 330 μV It could be defined as shift DC shis and HFO. 1 sec > 200 microV 5 sec TC = 10 s, HFF = 6 Hz 1) 2) Ictal HFOs onset (Imamura et al., 2011 , Kanazawa et al., 2014 ) duration of 0.33-10sec 500 Hz

HFO� HFO always accompanied by HFO Neurons� 0 DC shifts onset 0.9 mV (>2SD of power at 80-200Hz from the baseline state) D1 slow/DC shift conventional onset spikes or sharp waves excluded Active DC shifts� 2 mV 10 sec Glia� � TC = 10 s, HFF = 1 Hz 22 sec 40

Neuron, glia, and epilepsy: is it a paradigm shift? Methods Subjects: 18 patients with intractable partial epilepsy ECoG recording: 1) Glia vs. neurons, i.e., active- vs. passive DC shifts #1. band-pass filter of 0.016-600 Hz sampling rate of 2,000 Hz AMED study in Japan #2. band-pass filter of 0.016-300 Hz sampling rate of 1,000 Hz Analysis software: EEG review program (Nihon Kohden, Tokyo, Japan) Surgical utcome o Selection of electrode: 1 electrode with the earliest ictal DC shifts with TC 10 sec 2) Interictal red slow, i.e., co-occurrence of slow and HFO Parameters: #1. the amplitude and duration from onset to peak of ictal DC shifts with TC 10 sec and TC 2 sec superimposed TC 10 sec. 3) 2 types of ictal DC shifts #2. the attenuation rate of the amplitudes endorsed by TC10 sec vs. TC 2sec = (TC10sec-TC2sec)/TC10sec×100 (%) TC: 10sec TC: 2sec Attenuation rate 4）Is it recorded by scalp EEG ? of amplitudes(vertical axis)

5) Proposal of clinical practice parameter for recording and analysis of ictal DC shifts and HFO: invasive EEG onset peak Duration with TC 10sec (horizontal axis) (Kajikawa�et al, JES annual congress,2017)� LVFA

Pre-spiking prior to LVFA

Burst of polyspikes prior to LVFA

Slow-wave /DC shifts prior to LVFA

Rhythmic spikes/ spike- wave

There/alpha sharp waves�

HPF=LFF=0.16Hz(TC of 1sec)�

Neuron, glia, and epilepsy: is it a paradigm shift?

250Hz� 125Hz� 250Hz� 1) Glia vs. neurons, i.e., active- vs. passive DC shifts AMED study in Japan Surgical outcome

2) Interictal red slow, i.e., co-occurrence of slow and HFO

3) 2 types of ictal DC shifts 125Hz� 250Hz� 125Hz� endorsed by TC10 sec vs. TC 2sec

4）Is it recorded by scalp EEG ?

5) Proposal of clinical practice parameter for recording and analysis of ictal DC shifts and HFO: invasive EEG

Scalp-recorded slow (DC) EEG with Lt focal motor seizures Hypothesis of active vs. passive DC shift in the ictal period Short duration Long duration onset ・Conventional ictal pattern time ・HFO

・DC shift ③ ② ① time 1960s� (drawn by the help of Murai) ③Active DC shift ② ① Passive DC shift (TC of 10sec) Clinical state Chronic epilepsy Acute symptomatic seizure Ictal DC shifts (scalp recording) Clinical patients Habitual seizures Non-habitual seizures 1) Incidence rate: 14~40% (22%) in 73 seizures. Basic animals Chronic model Acute model 2) Detected particularly when seizures were clinically Chronic pilocarpine model Acute pilocarpine model LGi1 rat model intense, but not in small seizures. (Ikeda et al., 1999)� NER(=Noda epilepsy rat model)

(by Ikeda et al) (as the working hypothesis of AMED research team of Epilepsy and Glia, -> Future advancement in recording condition is warranted. chaired by Maehara, Japan, 2015-2018)

Neuron, glia, and epilepsy: is it a paradigm shift? Standards as clinical practice parameters for recording and � analysis of ictal DC shifts and HFO� 1) Glia vs. neurons, i.e., active- vs. passive DC shifts � AMED study in Japan Initially made by AMED research group in Japan� Surgical outcome approved by JES, Japan Epilepsy Surgery Society� currently on the investigation by Japan Clinical Neurophysiology 2) Interictal red slow, i.e., co-occurrence of slow and HFO Society� � 3) 2 types of ictal DC shifts endorsed by TC10 sec vs. TC 2sec

4）Is it recorded by scalp EEG ?

5) Proposal of clinical practice parameter for recording and analysis of ictal DC shifts and HFO: invasive EEG

Standards as clinical practice parameters for recording and � Conclusion analysis of ictal DC shifts and HFO� � 1) Subdurally recorded ictal DC shifts could represent humoral Initially made by AMED research group in Japan� state of at least [K+] immediately before seizure onset. approved by JES, Japan Epilepsy Surgery Society, � currently on the investigation by Japan Clinical Neurophysiology 2) Active DC shifts were predominant in “chronic” epilepsy. Society� Decreased Kir 4.1 channels of the astrocytes is responsible for � DC shift generation.

3) Red slow, i.e., interictal slow with HFO may represent similar feature to ictal event.

4) Scalp EEG cold record both ictal DC shifts and HFO, but in a limited manner.

5) Proposal of clinical practice parameter for recoding and analysis of ictal DC shifts and HFO is warranted for further clinical application

Collaborators Bubble (=HFO)�Neuron� Kyoto University School of Medicine Department of Epilepsy, Movement Disorders and Physiology Shimotake A,MD, Hitomi T,MD, Inouchi M,MD Department of Neurology Neuron� Kajikawa S, MD Nakatani N, MD, Murai T, MD, Togawa J,MD, Inoue T,MD, Daifu M, MD, � Matsumoto R, MD Human Brain Research Center Matsuhashi M, MD, Mima T, MD spike� Department of Neurosurgery Kobayashi T, MD, Kikuchi T, MD, Yoshida K, MD, Miyamoto S,MD Ehime University School of Medicine Kunieda T,MD

Non-linear Neuro-oscillology (Grant‐in‐Aid for Scientific Research on Innovative Areas, MEXT) Red slow? (epileptic slow)� Tsuda I, PhD, Namiki T, PhD, Kitano K,PhD, Aoyagi T, PhD, Kitajo K, PhD

Epilepsy and Glia (AMED) Ohno Y, PhD, Sato K, MD (Osaka), Kakita A, MD, Kitaura H, PhD (Niigata), Maehara T, MD (Tokyo) Acute symptomatic sz: N� Big wave (=DC,infraslow)Glia� International Migraine with aura: G� Bernard C, PhD (Marseilla), La Van Quyen M, PhD (Paris), de Curtis M (Mirano) Epilepsy: N+G� Glia� Special thanks to Prof. Lüders H. (Case Medical School, OH, USA) Prof. Emeritus Shibasaki H. (Kyoto University, Japan)