A low-power implantable device for epileptic seizure detection and neurostimulation

A low-power implantable device for epileptic seizure detection and neurostimulation
of 4
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  Table I. Biphasic stimulation parameters Stimulation Location Stimulation parameters Frequency On Off OLNS [2] ATN 145 Hz 1 min 5 min CLNS [3] EZ 100-500 Hz 1 sec 1 min ATN 100-200 Hz 2.5 sec 1 min OLS: Open loop neurostimulation, CLNS: Closed-loop neurostimulation, ATN: Anterior thalamic nucleus, and EZ: Epileptogenic zones   A Low-power Implantable Device for Epileptic Seizure Detection and  Neurostimulation Muhammad Tariqus Salam 1 , Dang Khoa Nguyen 2 , and Mohamad Sawan 1   1 Polystim Neurotechnologies Laboratory, École Polytechnique de Montréal, Québec 2  Neurology Service, Department of Medicine, Notre-Dame Hospital (Centre Hospitalier de l'Université de Montréal), Québec  Abstract –  In this paper, we present the design of a low-power closed-loop neurostimulator (CLNS) as an adjunctive treatment for patients with refractory partial epilepsy. The CLNS combines epileptic seizure detection with simultaneous electrical stimulation feedback. The system amplifies the neural signal with adjustable gain, detects epileptic low-voltage fast-activity, and generates programmable stimulation currents. The implemented seizure detector is based on a detection algorithm validated in Matlab tools and was tested using intracerebral electroencephalographic (iEEG) recordings from a patient with drug-resistant epilepsy. The amplifier, epileptic-seizure detector and electric stimulator were implemented using CMOS 0.18-µm process. The iEEG were assessed by the proposed CMOS building blocks and the predefined seizure suppression biphasic electrical stimulations were administrated at 2 to 3 sec after electrographical seizure onsets. The simulated power consumption of the CLNS has showed that the system could run on a button cell battery for more than 8 years. I.   I  NTRODUCTION Epilepsy is the second most common neurological disorder and approximately 50 million people worldwide have epilepsy. Despite currently available treatments (antiepileptic drugs, resective epilepsy surgery and vagus nerve stimulation), ~30% patients continue to have disabling seizures. There is a growing interest in device-based therapy for epilepsy as recent studies using implantable devices (RNS system, Neuropace Inc. and Sante trial, Medtronic)  for the treatment of epilepsy have shown promising results [1].  Proof-of-concept experiments conducted in animals and humans with epilepsy have demonstrated that focal electrical, thermal, or pharmacological manipulations of the epileptogenic zone can suppress seizure activity, paving the way for a novel approach to the treatment of epilepsy [1].  Given the experiences to date,  focal brain stimulation therapy has proven to be an effective therapy and has been receiving attention as an alternative therapy for refractory patients. There are two main approaches to brain stimulation: open loop neurostimulation (OLNS) and closed-loop neurostimulation (CLNS). The OLS consists of scheduled delivery of stimulation [2] while in CLNS, stimulation is triggered upon detection of a seizure [1]. The advantages of the closed-loop stimulation over the open-loop method are (i) lower number of treatments, (ii) less adverse effects, (iii) high efficiency because of rapid access, and (iv) ability to review recordings to monitor seizure frequency [3]. Some of these experiments have demonstrated the feasibility of responsive treatment [3]. The responsive device analyzes the intracerebral EEG (iEEG) , identifies seizures at their onset and triggers focal treatment to the epileptogenic zone to abort the seizure, whether by electrical stimulation, cooling, or drug release. To do so, an efficient seizure detection algorithm is required for accurate seizure onset detection without false alarms.  Several mathematical models have been developed to detect seizures . These models were developed using desktop computers for off-time data processing and cannot be employed in a low-power implantable microchip due to the heavy computation. Some seizure detection algorithms have  been proposed to be implemented in custom integrated circuits [4-6]. These are reliable seizure detection systems sensitive enough to detect seizures early on but also specific enough to prevent unwarrantedly triggered focal intervention [4-6]. Apart from the seizure detection performance, the other concerns are the safety of chronic electrical stimulation and the determination of optimal stimulation parameters. Table I shows parameters used in pilot trials. Moreover, Osorio et al. demonstrated the feasibility and short-term safety of automated high-frequency electrical stimulation in blocking seizures using an external prototype [3]. However, several issues remain to be addressed, such as maximum charge density, stimulation mode, current distribution, shunting of injected current, impedance of cerebral tissue, and the region affected by stimulation. In this paper, we present a low-power CLNS for the treatment of epilepsy. The CLNS includes intracranial electrodes, a data acquisition system, a digital signal processor in order to process real-time iEEG, and a programmable electrical stimulator. The proposed seizure detection algorithm is designed for patients with partial seizures characterized by  progressive increase of low-voltage fast-activity on iEEG recordings. The electrical stimulator has tuneability to the  biphasic stimulation parameters. The performance of the  proposed CLNS is verified using iEEG recordings from a  patient with refractory partial epilepsy. The system is expected to be reliable in an implantable device without risking false detections. II.   S EIZURE ONSET DETECTION ALGORITHM  Early investigations of seizure detection research were aimed at finding consistent changes and seizure onset patterns in iEEG prior to clinical manifestation. Several patterns in 978-1-4244-7270-3/10/$26.00 ©2010 IEEE 154    Fig. 1. Invasive EEG recordings of a patient with refractory focalepilepsy and signals analysis: (a) start of seizure activity ischaracterized by low-amplitude fast activity, (b) frequency analysisof the iEEG, (c) mean absolute amplitude analysis of the iEEG, and(d) seizure onset detection. Fig. 2. The biphasic stimulation parameters: (a) two subsequent  burst of stimulations, (b) the biphasic stimulation, (c) the stimulation waveform profile, and (d) ) limitation of maximum  pulse width with injected current of biphasic stimulation.   Table II Stimulation parameters Parameter Value Parameter Value T  1  1 sec – 1 min T  2  1 min – 5 min T  3  2 – 10 msec T  4   100 µ - 11msec  I  STIM    0 – 58 mA  I  GM   0 – 12 mA iEEG were reported, such as low-voltage fast-activity, high-voltage fast-activity or rhythmic spiking [7]. The seizure onset  pattern usually lead to a variety of behavioral manifestations; however, it may sometimes be clinically silent (i.e. electrical seizure) especially if the discharge remains very focal (without spread), is brief (few seconds) and occurs in non-eloquent cortex. Time-frequency analysis of iEEG can provide precise seizure onset information. Fig. 1 shows time-frequency analysis of epileptic iEEG (Fig. 1(a)) recorded from the right medial temporal lobe (hippocampus) using depth electrodes. This analysis at seizure onset (at the 160 sec timepoint) demonstrates a sudden increase in frequency (Fig. 1(b)) and  progressive amplitude increase (Fig. 1(c)) of the iEEG signal leading up to clinical manifestations. Because the increase in frequency and amplitude may vary from patient to patient according to the underlying substrate, the type of intracerebral electrodes used, and their locations with respect to the epileptogenic zone, the seizure detection algorithm requires several adjustable parameters for optimal sensitivity and specificity. Signal analysis of the seizure onset detection demonstrates that the early modulation and proper rectification of iEEG can identify low-voltage fast-activity of the signal efficiently. The iEEG is analyzed over a certain time frame and a higher number of the detections indicates an upcoming seizure event (Fig. 1(d)). III.   S TIMULATION PARAMETERS  Two modes of cortical electrical stimulation have been shown feasible, safe, tolerable, and efficacious in humans:  bipolar or unipolar stimulation. The unipolar mode tends to stimulate a wider region while the bipolar mode is effective in  producing localized current flows [8]. The charge-balanced asymmetric biphasic stimulation (Fig. 2 (a) – (c)) can avoid damaging electrochemical processes of the brain because the cathodic pulse (  I  STIM  1 ) of the biphasic stimulation (Fig. 2(c)) activates the neurons and the anodic pulse (  I  STIM  2 =  I  STIM  1 /4) removes the delivered charges of  I  STIM  1 . However, two subsequent opposite pulses could prevent the generation of an action potential. Thus, a short time delay ( T  4 ) in between the  pulses is required to propagate the action potential away from the stimulation site before removing the injected charge. It is generally agreed that limiting the maximum charge density ( σ MAX ) to 60 µC/cm 2  per phase can avoid tissue damage [9]. Fig. 3(d) illustrates the electrical brain stimulation safe region using standard subdural electrodes (diameter: 5mm and interelectrode spacing: 10 mm). The safety region is defined  by Eq. (1). GM  MAX   I  AT   σ   = 4  (1) where, T  4  is pulse width, A is area of the electrode and  I  GM   is injected current through the gray.   During the biphasic stimulation using bipolar subdural electrodes, there is significant shunting of current through the cerebrospinal fluid (CSF) because of higher (approximately four times) current density in CSF than the gray matter [8]. Therefore, the stimulated current  I  STIM   across the electrodes can be estimated (Eq. (2)) from the resistive model of brain [7]. CSF CSF GM GM STIM   R R R I  I  )(  +=  (2) where,  R GM   and  R CSF   are resistance of gray mater and CSF, respectively. Fig. 2(d) shows linear relationship between  I  STIM   and  I  GM  . The proper stimulation parameters need to be individualized during the presurgical iEEG study for maximum efficacy prior to the implantation of a stimulation device. The range of promising biphasic stimulation  parameters are shown in Table II.   The stimulation is delivered at close proximity to the seizure onset, directly to (i) anterior thalamic nucleus (ATN) by depth electrodes (Fig. 3) suitable for the patients with two or more epileptogenic zones (EZ) and (ii) epileptogenic tissues by subdural electrodes (inset of Fig.3) suitable for patient with one or more discrete 155  Neural signalSeizure detector DetectionTimer Stimuli generator Output stage Di     s a b l     e enable YesNoFrequency Generator  I  STIM  [ T  1- 4 ] F  SZ V  DA V  ST  V  REF I  STIM     D  e  v   i  c  e   S  u   b   d  u  r  a   l  e   l  c  e   t  r  o   d  e  s (a)(b) (c) Amplifier  Seizure detector    Stimuli generator Output stage V  ST  V  SO      D   E   M   U   X Fig. 4.   The proposed integrated CLNS: (a) the implant configuration, (b) block diagram, and (c) flowchart of the system.   Fig. 5. Schematic diagrams of the electrical stimulator: stimuli generator, output stage and frequency generator.   T  1  T  2 Fig. 6. Timing process of two biphasic stimulation burstsgeneration.   epileptogenic zone [3]. IV.   S YSTEM IMPLEMENTATION  The proposed implantable CLNS provides continuous long-term monitoring of iEEG on multiple simultaneous channels from epileptogenic zones. Fig. 4(a) illustrates the implant configuration of CLNS and the functional block diagram (Fig. 4(b)) presents the architecture of the CLNS. The device will  be implanted within the skull and interface directly with the recording/stimulation sites. The flowchart of the CLNS (Fig. 4(c)) shows that bipolar biphasic electrical stimulation therapy is activated in response to seizure detection. Details of the CLNS are described below.  A.    Amplifier A new chopper stabilized preamplification method was introduced in our previous work [6], where the demodulator and low-pass filter of a classical chopper stabilized  preamplifier [9] was replaced by a high-pass filter. The amplifier has 80 dB midband gain and 17 kHz (2 kHz to 19 kHz) bandwidth with 6µVrms input-referred noise. Moreover, the variable gain of the amplifier can emphasize certain amplitude range of the neural signal.  B.   Seizure detector The detector [6] consists of voltage level detectors, digital demodulators and a high-frequency detector. The adjustable threshold voltage identification and tunable   recognition of high-frequency activities provide unique detection criteria for a specific patient. Moreover, digitally-controlled low-power CMOS circuits perform accurate seizure detection. C.   Stimuli generator The stimuli generator (Fig. 5) is a current mode DAC that comprises 60 identical current sources activated by a binary-to- inverse thermometer code generator (B2T). Each of the current sources (I L ) is designed to supply 16 µA and maximum output current (  I  SG ) is 1 mA.  D.    Frequency generator The frequency generator contains 4 counters and a 3-to-8 decoder (Fig 5). Fig. 6 demonstrates the timing process of frequency generator in the response of seizure detections V  SO . The stimulation (  I  STIM ) duration is T  1 , while another seizure detection begins after T  2 , if the seizure does not terminate. The restimulation is subject to a safety constraint of no more than five stimulations per detection [3]. The counter 1 and 2 generate V   DA and   V  ST   in Fig. 6, respectively. On the other hand, the counter 3 defines the frequency (1/  T  3 ) of the stimulation and a state machine (counter 4 and decoder) generates the stimulation waveform. Following sections discuss in further detail.   Fig. 3. The electrical stimulations are delivered in close proximity to the seizure onset, directly to anterior thalamic nucleus by depth electrodes and the inset shows the stimulations are delivered to epileptogenic tissue by subdural electrodes.   156        V       i     n       (    µ      V      )      V       S      0       (      V      ) Fig. 7.   The iEEG recordings of a patient with refractory focalepilepsy and zoom inset output of the seizure detector.      I    G   M      (    m     A     )       E       1       E       2       E       3       E       4       E       5       E       6       E       7      E       i       E       0       0      0      1      S       1      S       2      S       3       1      0      0      0      0      1      0      1      0      0      1      0      1      0      0      1      0      0      0      1      0   Fig. 8.   Output current of the stimulator for binary encoded amplitude.  E.   Output stage The final stage (Fig. 5) of the stimulator activates the  bipolar electrodes using switches S 1 , S 2 , and S 3 . The state machine controls the switches and generates the biphasic waveform. The current mirror circuits are designed to supply  I  STIM1  = 4  I  STIM2  and maximum  I  STIM1  is 58 mA. V.   R  ESULTS  A 24 year-old male with drug-resistant partial epilepsy candidate for epilepsy surgery underwent an intracranial study to better delineate the epileptogenic zone. Several seizures were recorded, all srcinating from the right hippocampus and spreading to the lateral temporal neocortex and the insula (Fig. 7). The seizure onsets were marked by an epileptologist (DKN). The seizure onset was characterized by an initial low-voltage tonic alpha activity (at the 25 sec mark) evolving into rhythmic spiking. The signal recorded during a seizure (inset of the Fig. 7) was fed into the CLNS to test the performance. The input signal V  in  was modulated, amplified, and analyzed in frequency and time domain in order to detect abnormalities in the signal. For this particular seizure, the onset V SO was detected at 28 sec. The effective individualized biphasic stimulation was administrated to a dummy tissue at regular interval ( T  3 ) in response to the early detection of seizure development. The stimulation parameters (  I  STIM , T  1 , T  2 , T  3 , and T  4 ) are widely variable. Fig. 8 shows that the range of injected current into the gray matter is 0 to 12 mA and resolution is 200 µA. The state machine (E 0 , E 1 , E 2 , …, E 7 ) of frequency generator (Fig. 5) controls the switches (S 1 , S 2 , and S 3 ) to generate the biphasic stimulation waveform (Fig. 8). The  power consumption is dominated by the stimulation  parameters and it would vary patient-to-patient. However, the whole device with the highest stimulation setting (  I  GM   = 12 mA, T  4  = 1 sec and 3 stimulations per week) could theoretically run on a rechargeable lithium-ion button cell  battery (LIR2025: dimension 20 x 2.8 mm) for 8 years. VI.   C ONCLUSION In this paper, we have described the design of a low-power closed-loop stimulator that is hypothesized to improve the electrical brain stimulation technique for the treatment of epilepsy. The CLNS uses the properties of the measured iEEG to trigger a predefined seizure suppression biphasic electrical stimulation upon seizure onset. The performance of the CLNS tested in a patient with refractory epilepsy has shown early detection of seizure with subsequent responsive stimulation. Such stimulation of the epileptogenic focus may hopefully disrupt the seizure progression and propagation to the adjacent regions. Further validation of this low-power implantable system responsive to ictal low-voltage fast activity patterns is underway. A CKNOWLEDGMENTS The authors are grateful for support from the NSERC, the Canada Research Chair in Smart Medical Devices, and the EEG technicians at Notre-Dame Hospital, Montréal. R  EFERENCES   [1] S. C. Schachter, J. Guttag, S. J. Schiff, D. L. Schomer, and Summit Contributors, Advances in the Application of Technology to Epilepsy: The CIMIT/NIO Epilepsy Innovation Summit,  Epilepsy & Behavior  , vol. 16, 3–46, 2009. [2] R. Fisher, V. Salanova, T. Witt, R. Worth, T. Henry, R. Gross, K. Oommen, I. Osorio, et al. , Electrical stimulation of the anterior nucleus of thalamus for treatment of refractory epilepsy, Epilepsia, vol. 51(5):899–908, 2010. [ 3] I. Osorio, M. G. Frei, S. Sunderam, J. Giftakis, N.C. Bhavaraju, S. F. Schaffner, and S. B. Wilkinson, Automated seizure abatement in humans using electrical stimulation,  Ann of Neurology , vol. 57(2), 258-68, 2005. [4] M. T. Salam, M. Sawan, and D. K. Nguyen, Low-Power Implantable Device for Onset Detection and Subsequent Treatment of Epileptic Seizures: A Review,  Journal of Healthcare Engineering  , vol. 1(2), 2010. [5] M. T. Salam, M. Sawan, D. K. Nguyen, and A. A. Hamoui, “Epileptic Low-Voltage Fast-Activity Seizure-Onset Detector,”  IEEE-BIOCAS  , 2009. [6] M. T. Salam, M. Sawan, D. K. Nguyen, and A. Hamoui, “Low-power CMOS-based epileptic seizure onset detector,”  IEEE-NEWCAS  , 2009. [7] D. K. Nguyen, S. S. Spencer, and B. D. Robert, Invasive EEG in  presurgical evaluation of epilepsy, Chapter 62 of The Treatment of Epilepsy, pp. 633-658, 2009. [8] S. S. Nathan, S. R. Sinha, B. Gordon, R. P. Lesser, and N. V. Thakor, Determination of current density distributions generated by electrical stimulation of the human cerebral cortex, Electroencephalography and Clinical Neurophysiology, vol 86(3), pp. 183-92, 1993. [9] T. L. Skarpaas and M. J. Morrell, Intracranial stimulation therapy for epilepsy,  Neurotherapeutics , 6(2), 238-43, 2009. [10] B. Gosselin, M. Sawan, and C. A. Chapman, "A low-power integrated  bioamplifier with active low-frequency suppression," IEEE Transactions on Biomedical Circuits and Systems, v 1, n 3, pp. 184-192, 2007. 157
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks