UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The design of a versatile magnetic stimulator Choi, Angela Sio-van 1990

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1991_A7 C56.pdf [ 5.41MB ]
Metadata
JSON: 831-1.0065454.json
JSON-LD: 831-1.0065454-ld.json
RDF/XML (Pretty): 831-1.0065454-rdf.xml
RDF/JSON: 831-1.0065454-rdf.json
Turtle: 831-1.0065454-turtle.txt
N-Triples: 831-1.0065454-rdf-ntriples.txt
Original Record: 831-1.0065454-source.json
Full Text
831-1.0065454-fulltext.txt
Citation
831-1.0065454.ris

Full Text

THE DESIGN OF A VERSATILE MAGNETIC STIMULATOR By ANGELA SIO-VAN CHOI B.A.Sc. The University of British Columbia, 1985 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Electrical Engineering We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA October 1990 ® Angela Sio-van Choi, 1990 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of E l e c t r i c a l Engineer ing The University of British Columbia Vancouver, Canada n , t p 5th October, 1990 DE-6 (2/88) ABSTRACT A magnetic stimulator consisting of a capacitor discharge and associated control circuits was designed and constructed for use in research and practice. This stimulator is capable of delivering an output voltage of 1000 V and output peak current of 8 kA to a coil. Depending on the coil used, the rise time of the current pulse can be varied in steps of 30 us or less. After each discharge, the stimulator measures the output current amplitude to within 0.2 kA and rise time to within 10 us, and displays their values on a liquid crystal panel. The stimulator can operate in pulsed or in repetitive discharge mode (1 pulse per 5 seconds). With minor modifications, the stimulator can operate up to 1600 V and provide up to 16 kA. The rise time can be changed in steps of 10 us or less. The capability to provide output current of variable rise time makes this stimulator is a versatile instrument for both research and clinical use of magnetic stimulation. This feature, which is found in no other stimulating instrument, is important for studying the effect of current rise time on magnetically induced stimulation. With this device we can determine the optimum current pulse shape for effective stimulation with the minimum energy. ii TABLE OF CONTENTS ABSTRACT ii LIST OF TABLES vi LIST OF FIGURES vii ACKNOWLEDGEMENT x INTRODUCTION 1 PHYSIOLOGICAL BACKGROUND AND ELECTRICAL STIMULATION 4 2.1 INTRODUCTION 4 2.2 PHYSIOLOGY OF PERIPHERAL NERVES 4 2.2.1 The Nerve Cell 5 2.2.2 Transmembrane Ionic Balance 6 2.2.3 The Equivalent Circuit Model of an Axon 8 2.2.4 Nerve Excitation 10 2.2.5 Propagation of the Action Potential 13 2.2.6 Neuromuscular Transmission 14 2.3 EXTRACELLULAR ELECTRICAL STIMULATION OF NERVES 16 2.3.1 Basic Principles of Extracellular Electrical Stimulation . .' 16 2.3.2 The Cable Theory 20 2.3.3 Current as a Stimulus 23 2.3.4 Extracellular Stimulation of Nerves in Practice 25 2.3.5 Stimulation of Muscles 26 MAGNETIC STIMULATION: THEORY AND PRACTICE 27 3.1 INTRODUCTION 27 3.2 THE THEORY OF MAGNETIC STIMULATION 28 3.3 THE PRACTICAL ASPECTS OF MAGNETIC STIMULATION OF PERIPHERAL NERVES 33 3.3.1 The Shape of the Coil Current pulse 33 3.3.2 Orientation and Location of the Coil 35 3.3.3 The Geometry and Construction of the Coil 38 3.4 MAGNETIC STIMULATION OF THE MOTOR CORTEX 40 3.5 MAGNETIC STIMULATOR TECHNOLOGY 40 3.6 SAFETY OF MAGNETIC STIMULATION : 41 3.7 OBJECTIVES 42 iii DESIGN OF THE MAGNETIC STIMULATOR 44 4.1 INTRODUCTION 44 4.2 THE BASIC STIMULATOR CIRCUIT 45 4.3 OBSERVATIONS AND PILOT STUDIES 49 4.4 PRELIMINARY DESIGN CONSIDERATIONS 51 4.4.1 Electrical Ratings of the Components 51 4.4.2 Protective Features 52 4.4.3 Operator Interface 53 4.5 TARGET SPECIFICATIONS OF THE MAGNETIC STIMULATOR 53 4.6 DETAILED DESIGN OF THE BASIC BLOCKS OF THE MAGNETIC STIMULATOR 54 4.6.1 The Power Supply Module 56 4.6.2 Capacitor Charge/discharge Module 59 4.6.3 The SCR Triggering Module 65 4.6.4 Operator Interface Module 66 4.6.5 Sensor Circuits . 67 4.6.6 The Microcontroller and Logic Module 71 4.6.7 SOFTWARE DESIGN 74 4.7 PRACTICAL DESIGN CONSIDERATIONS 81 4.7.1 Mechanical and Electrical Safety 82 4.7.2 Electromagnetic Interference 83 ENGINEERING TESTING AND PERFORMANCE EVALUATION 84 5.1 INTRODUCTION 84 5.2 LIMITATIONS OF THE PRESENT STIMULATOR 84 5.3 INSTRUMENTATION 85 5.4 THE 1600 V POWER SUPPLY 89 5.5 THE CAPACITOR CHARGE/DISCHARGE MODULE 93 5.5.1 Capacitor Discharge 93 5.5.2 The dl/dt and 1^ Measurements 95 5.5.3 The dVJdt and Measurements 96 5.6 THE SCR'S TRIGGERING CIRCUITS 101 5.7 THE MICROCONTROLLER AND THE LOGIC CIRCUITS 105 5.8 SENSORS 107 5.9 PERFORMANCE TEST I l l PROPOSAL FOR CLINICAL TESTING 115 6.1 INTRODUCTION 115 6.2 COIL SELECTION 116 6.3 Clinical Testing of Magnetic Stimulation 117 6.4 Techniques to Modify and Enhance Magnetic Stimulation 120 6.4.1 Magnetic Shield 120 6.4.2 Combination of Electrical and Magnetic Stimulation 122 iv CONCLUSION AND RECOMMENDATIONS 124 7.1 CONCLUSION 124 7.2 RECOMMENDATIONS FOR FURTHER WORK ON THE STIMULATOR 124 7.3 RECOMMENDATIONS FOR FURTHER WORK ON MAGNETIC STIMULATION 126 REFERENCES 128 v LIST OF TABLES Table 4.1 Coil current amplitude 1^ and rise time calculated with Vc = 1600 V, R = 50 mO and L = 5 /xH and 10 iiH 62 Table 4.2 The current ratings of the ten SCR's in the stimulator. See Figure 4.5 for the SCR designations 63 Table 4.3 The functions of the five ports of the 68HC11 microcontroller 73 Table 4.4 Function assignment to the different ports of the microcontroller 74 Table 5.1 The maximum calculated 1^ and dl/dt through the discharge SCR's and the rated values 96 Table 5.2 The relationship between and the other voltages in the circuit 99 Table 5.3 Theoretical coil current amplitude I9 and rise time assuming R = 40 mfi and L = 14.7 uH vs. the measured values 113 vi LIST O F F I G U R E S Figure 2.1 A typical nerve cell 5 Figure 2.2 A myelinated nerve 6 Figure 2.3 The sodium pump across a membrane 8 Figure 2.4 The equivalent transmission line network of the nerve and its membrane 9 Figure 2.5 The Na + conductance and K + conductance in relationship to the action potential 11 Figure 2.6 Absolute and relative refractory periods 12 Figure 2.7 The propagation of action potential by current flow 13 Figure 2.8 A typical neuromuscular junction 15 Figure 2.9 An unmyelinated nerve in the presence of a cathodal electrode. The applied electric field is transverse to the nerve 17 Figure 2.10 A unmyelinated nerve cell in the presence of an extracellular cathodal electrode. The applied field is longitudinal to the nerve 18 Figure 2.11 An externally applied current (ip) in the equivalent circuit of an axon. . . . 20 Figure 2.12 A strength-duration curve, with C set to 1 on the abscissa. Ir is the rheobase current and C is the chronaxie 24 Figure 2.13 Relationship between nerve stimulation threshold and the number of stimulating pulses 26 Figure 3.1 Magnetic stimulation of peripheral nerve 29 Figure 3.2 Cable model with an induced field ejx.t) 30 Figure 3.3 The equipotential lines of the axial derivative of the induced electric field. The points marked with + will be maximally hyperpolarized and the places marked with - will be maximally hypopolarized 32 Figure 3.4 The ideal (a) and the practically available (b) coil and induced currents. . . 35 Figure 3.5 The spatial and temporal effects of the coil current on magnetic stimulation of a nerve 37 Figure 3.6 The different coil shapes and directions of current flow 38 Figure 3.7 Safety limits for whole body exposure to time varying magnetic fields and the exposure the subject receives at ir/2 T within 100 s^ 42 Figure 4.1 Simplified circuit diagram of the charge/discharge circuit 46 Figure 4.2 Coil current and its derivative assuming that R = 50 mfi, L = 8 fxH and C = 1200 |*F 49 Figure 4.3 The six functional modules of the magnetic stimulator 55 Figure 4.4 The 1600 V DC power supply 58 Figure 4.5 The capacitor circuit 59 Figure 4.6 The SCR triggering circuit 66 Figure 4.7 The zero-crossing detector and the amplitude peak detector 69 Figure 4.8 The voltage sensor circuit 71 vii Figure 4.9 The microcontroller and the associated circuits 72 Figure 4.10 Flow Chart of the Software Mag_5.asm 77 Figure 4.11 Flowchart of the subroutine RUN 78 Figure 4.12 Flowchart of the subroutine MENU 79 Figure 4.13 The flowchart for the subroutine DATA 80 Figure 4.14 The flowchart for the subroutine CHARGE 81 Figure 5.1 The output of the current monitor with and without bias current 87 Figure 5.2 The voltage and current designations in the stimulator 88 Figure 5.3 The current provided by the filter and the AC line supply during charging 89 Figure 5.4 The fdter capacitor voltage, the discharge capacitor voltage and the charge current at the beginning of a charging cycle 91 Figure 5.5 Initial current delivered by the full-wave rectifier at the beginning of the charge cycle (V = 1000 V, C = 900 tiF) 92 Figure 5.6 The coil current, the diode current, the capacitor current and the capacitor voltage during discharge 94 Figure 5.7 Equivalent circuit of the capacitor charge/discharge module for the three operating modes: standby, charging, and discharging 98 Figure 5.8 The anode-cathode voltage of CSCR2 during charging 100 Figure 5.9 The anode-cathode voltage of CSCR2 during discharge 100 Figure 5.10 The anode-cathode voltage of DSCR2 during charging 100 Figure 5.11 The anode-cathode voltage of DSCR2 during discharge 100 Figure 5.12 The gate-cathode voltage Vgk of charge SCR's during charge 103 Figure 5.13 The gate-cathode voltage Vgk of charge SCR's during discharge 103 Figure 5.14 The gate-cathode voltage Vgk of discharge SCR's during charge 103 Figure 5.15 The gate-cathode voltage Vgk of discharge SCR's during discharge 103 Figure 5.16 The gate current Ig of a charge SCR during charging 104 Figure 5.17 The gate current Ig of a charge SCR during discharge 104 Figure 5.18 The gate current Ig of a discharge SCR during charging 104 Figure 5.19 The gate current Ig of a discharge SCR during discharge 104 Figure 5.20 The DC supply lines during discharge at Vc = 900 V, three banks of capacitors selected 106 Figure 5.21 The DC supply lines during discharge at Vc = 900 V, one bank of capacitors selected 106 Figure 5.22 The coil current amplitude 1^ measured by the stimulator vs. true 7^ . . . 109 Figure 5.23 Coil current rise time measured by the stimulator vs. true 110 Figure 5.24 Coil current at C = 900 uF at V = 250 V, 500 V, 750 V and 1000 V. . I l l Figure 5.25 Coil current at C = 100 pF, 300 uF, 500 uF and 900 pF at V = 1000 V 112 Figure 6.1 The optimum positioning of the coil over the median nerve. The site of depolarization is located at the point of maximum change of induced electric field along the nerve 117 viii Figure 6.2 The different positions of the tear-drop shaped coil over the median nerve in magnetic stimulation. Testing is performed at the angle 0 = 0° to 90° at 15° intervals 119 Figure 6.3 The different orientation of the coil in magnetic stimulation. Testing is performed at a and B = 0 ° to 90° at 15° intervals 119 Figure 6.4 A shield for use in magnetic stimulation of the arm. The magnetic field is reflected away from the diamagnetic shield. Only the part of the arm exposed has current induced in it 121 Figure 6.5 Magnetic stimulation combined with electrical stimulation 122 ix ACKNOWLEDGEMENT I wish to thank those who provided me with advice and supervision throughout this work. Dr. C. A. Laszlo provided me with the idea of the project and guided me throughout. Dr. W. D. Dunford has advised me on power electronics and measurement techniques. Dr. C. Hershler gave guidance on clinical applications, and Dr. L. M . Wedepohl taught me some special aspects of power electronics and power supply design. I would like to thank Mr. D. Fletcher for building the chassis of the stimulator, Mr. L. Welder for system integration and fabrication of the unit, and Mr. C. Sheffield for acquiring needed components for the project. I also would like to thank the British Columbia Science Council for funding the project through the Health Development Fund (HDF Grant 28-89) and the Natural Science and Engineering Research Council of Canada for providing me with the Postgraduate Scholarships for the first two years of my study. x C H A P T E R 1 I N T R O D U C T I O N Stimulation of peripheral nerves using electrical current is widely used in clinical practice to obtain diagnostic information about the neuromuscular system. In electrical stimulation electrodes are placed on the skin over the nerve, and current flowing through the electrodes passes into the muscular and nervous tissues. Current may also be induced in such neuromuscular tissues using rapidly changing magnetic fields. This is achieved by passing a current pulse through a coil placed outside the body in proximity of the nerve. This coil current generates an electromagnetic pulse that induces current in the tissues. Since this induced current is directly proportional to the time derivative of the coil current, the coil current amplitude and rise time determine the amplitude and duration of the induced current. Current injected into body tissues tends to disperse. Therefore, to stimulate deep nerves the current flowing through the skin must be high, and this can be very painful. In magnetic stimulation the electromagnetic pulse induces current directly around the nerves, and therefore deep nerves can be stimulated without causing much pain to the subject. This makes magnetic stimulation an attractive alternative to electrical stimulation in applications where the subject is 1 CHAPTER 1 sensitive to pain, where electrodes cannot be placed on the skin (burn cases), or where deep nerves are to be stimulated. Research on magnetic stimulation of nerves so far has focused on the various possible uses, the technique, and clinical observations. Most researchers use one of the few commercially available magnetic stimulator models in their work, and publications on the design of the stimulation coil and the generation of the current pulse have been few. The magnetic stimulator models available allow the user to change the coil current amplitude, but not the rise time. This limitation makes it impossible to see the effect of the rise time of the coil current on the stimulation. In electrical stimulators, the stimulating current is variable both in amplitude and in duration because stimulation of different nerves requires current pulses of different amplitude and duration. Similar capabilities should be available on magnetic stimulators for researchers and clinicians to achieve effective stimulating conditions with minimal electromagnetic exposure to the tissues. It is therefore the objective of this work to design and build a versatile magnetic stimulator which is capable of providing coil currents of variable amplitude and rise time. In this thesis we present the detailed design of such a stimulator. In Chapter 2, we discuss the neurophysiology of the nerve and its membrane, the process of nerve excitation, and action potential propagation. We also discuss the basic principles of electrical stimulation, and some practical aspects of this technique. 2 CHAPTER 1 Chapter 3 introduces a modified cable model for the nerve, and discusses the mechanism of magnetic stimulation. The practical aspects of the design of the stimulation coil, and the current pulse generator are also explored. In Chapter 4 we present the detailed design of a versatile magnetic stimulator. Based on published information and considering the practical limitations of available components, we establish the target specifications for the instrument. The actual design follows a modular approach. Special attention is paid to electrical safety, mechanical ruggedness, and operator convenience. In Chapter 5 the results of the testing of the performance of the magnetic stimulator are presented. Each module was tested both separately and in place to ensure that the components work properly when connected together and that they are operating within their rated values. The performance of the complete system was then compared with the target specifications. In Chapter 6 we present a plan for testing the performance of the stimulator in the clinical environment. This plan involves the stimulation of the median nerve both electrically and magnetically, allowing direct comparison of the performance of the two techniques for the first time. Finally, we discuss ideas that have the potential to make magnetic stimulation more focused and/or limit the exposure of the subject to the electromagnetic field. In Chapter 7 we present the conclusions and recommendations for further work. 3 CHAPTER 2 PHYSIOLOGICAL BACKGROUND AND ELECTRICAL STIMULATION 2.1 INTRODUCTION A general understanding of electrical stimulation is important in understanding magnetic stimulation. In this chapter, we will first discuss the physiology of a nerve and how it is excited. Then we will discuss extracellular electrical stimulation, including an equivalent circuit of a nerve (simplified cable theory). Finally, we will present the practical aspects of electrical stimulation, mostly on the location of the electrodes and the pulse shape of the stimulating current. 2.2 PHYSIOLOGY OF PERIPHERAL NERVES When a nerve is at its resting state, it maintains a voltage between the inside and outside of the nerve. This process is a dynamic one, involving both passive diffusion and active transport of ions vital to the nerve function. When a nerve is excited, this process is disrupted and a transient phenomenon is responsible for the ion flow. The excitation is then transmitted along the nerve as nerve impulses. 4 CHAPTER 2 2.2.1 The Nerve Cell The nervous system in mammals consists of a complex network of nerves. The nerves of the brain and the spinal cord form the central nervous system, and the nerves of the limbs etc. make up the peripheral nervous system [1]. A nerve cell is composed of the cell body (soma), the axon, and the dendrites. The soma contains the nucleus and performs many control and metabolic functions. The axon serves as the "actuator" that carries "messages" in the form of pulses to other nerves, muscles or organs. Dendrites connect to axons or soma of other nerves and function as "input terminals" (Figure 2.1). Since our primary interest is the stimulation of the peripheral nerves, we will discuss them in detail. Each peripheral nerve is a bundle of axons enclosed in connective tissues. In myelinated nerves, the axons are surrounded by discontinuous insulative tissues called myelin sheaths (Figure 2.2), which is absent in unmyelinated nerves. Figure 2.1 A typical nerve cell. 5 CHAPTER 2 Figure 2.2 A myelinated nerve. 2.2.2 Transmembrane Ionic Balance In each nerve cell there is a potential difference between the inside of the cell membrane and the outside, maintained by the higher internal concentration of anions. This potential difference is called the membrane potential (Vm). In the steady state, Vm is called the resting potential (Vm0), and it remains constant between -50 mV and -100 mV in nerves and skeletal muscle fibres. For the sake of simplicity, in the following discussions we will assume Vm0 is -90 mV. On either side of the cell membrane, there is a distribution of ions, mainly potassium (K+) cations, sodium (Na+) cations, chloride (CI) anions, and large protein ( A ) anions. Since there is a much higher concentration of K + ions inside the membrane than outside, these ions 6 CHAPTER 2 tend to move across the membrane under the influence of osmotic pressure. The membrane potential Vm, on the other hand, opposes the change in ionic distribution. In addition, when K + cations move across the membrane, the redistribution of ions sets up an electric potential which opposes the efflux of more cations. At equilibrium this electric potential, which is the K + equilibrium potential (F4), produces a force which is equal and opposite to that exerted by the osmotic pressure. To a lesser extent, the membrane is also permeable to other ions, especially to CI" and Na + ions. At equilibrium the CI" anion distribution depends on the distribution of the K + ions and E C 1 = Ej.. With the Na + cations a process other than passive diffusion is involved since they tend to flow into the cell due to both concentration difference and potential difference Vm. Were they allowed to stay in the cell, the cell would have to give up K + cations. Then Vm would start to rise, and gradually the difference in ion distribution would disappear. To keep the membrane at Vm0, the Na + ions which diffuse into the cell are actively transported out of the cell by a mechanism called the sodium pump. In addition, for each Na + ion pumped out of the cell, a K + ion is pumped back into the cell. This is called a coupled Na+ - K+ pump (Figure 2.3). The passive flow of K + ions and the active transport of cations by the Na + - K + pump enable the cell to maintain a dynamic balance in the intra- and extra-cellular ionic concentrations, and therefore Vm0. Since Vm0 is one of the driving forces for the Na + ions to flow into the cell, the larger Vm0 is, the more active is the sodium pump. 7 CHAPTER 2 2.2.3 The Equivalent Circuit Model of an Axon We can visualize an axon and its membrane as an electrical network [2] from their following properties: first, there is ionic movement (current) inside and outside of the nerve cell; second, there is current flow across the membrane and third, there is a potential difference across the membrane. We can then express these properties in electrical parameters (Figure 2.4). The circuit is similar to an electrical transmission line, and thus the name "cable theory". 8 CHAPTER 2 1 r rl dx r} dx Ie (EXTERNAL) o ' V X / V -V\ r2dx -Vm0 r2dx r x r2dx r 2 ^ /, (INTERNAL) Figure 2.4 The equivalent transmission line network of the nerve and its membrane. = cytoplasm resistance per unit length, = intracellular resistance per unit length, rm = membrane resistance per unit length, cm = membrane capacitance per unit length, = intracellular potential, v0 = extracellular potential, vm = membrane potential, Kin = resting potential, = intracellular current, h = extracellular current The intracellular resistance (r,) can be compared with the resistance of a wire: R - p - (2.1) A resistance of the wire in fi resistivity of the wire in fi per unit length length of the wire cross-section of the wire Since r{ is resistance per unit length, for a given cytoplasm resistivity it is only dependent on the cross-section of the axon. Therefore, the larger the axon size, the lower is rt. where R = P = I = A = 9 CHAPTER 2 Similarly, rm is inversely proportional to the surface area of the membrane and is therefore smaller in larger axons. 2.2.4 Nerve Excitation Nerve and muscle cells respond to various types of stimulus (electrical, mechanical, chemical etc.). In such cells Vm changes briefly to a positive value when the stimulus causes excitation. This transient membrane potential is called the action potential (AP). Cells in which an action potential can be triggered are called excitable cells. At its resting potential the sodium channels of the nerve cell membrane are closed. Moreover, when these channels are open they only allow the Na + ions to pass through the membrane but not the K + ions. At the beginning of the action potential, a stimulus causes Vm to rise to about -60 mV (called threshold), the sodium channels start to open and the Na+ ions rush across the membrane into the cell. This can be viewed as increased Na + conductance of the membrane and the consequent rapid rise of Vm because of influx of Na + ions. This increase in Na + conductance lasts less than 1 ms and then it starts to decrease. About 1 ms after the start of the action potential, the K + conductance begins to increase and there is an efflux of K + ions. Vm starts to decrease and may oscillate slightly before settling down to the resting potential. The rising phase of the action potential is called depolarization, and the falling phase is called repolarization. Figure 2.5 presents the time relationship among the Na + and K + conductances and the action potential. 10 CHAPTER 2 g (103Q"1cm2 ) 40 "30 "20 - 10 0 0 1 2 3 t(ms) Figure 2.5 The N a + conductance and K + conductance in relationship to the action potential. After depolarization the sodium channel system is inactive until Vm has dropped to -70 mV and stays below that voltage for several milliseconds. During the first millisecond or so the nerve is absolutely inexcitable and therefore this period is called absolute refractory period. The cell threshold decreases from a very high value to the normal value within the next few milliseconds (called the relative refractory period). It takes about 4 ms for most nerves to go from the start of the action potential to the excitable state again. V(mV) 20 - A ACTION POTENTIAL 0" \ -20- \ \ -40 - <W\ \ -60--80-V—'V l\ \x ' \ \ x / \ \ -100-11 CHAPTER 2 V (mV)" -100-0 1 2 3 t(ms) [ — ABSOLUTE *-\+ RELATIVE REFRACTORY PERIOD Figure 2.6 Absolute and relative refractory periods. 12 CHAPTER 2 2.2.5 Propagation of the Action Potential Once an action potential is triggered, the sodium pump sends anions across the membrane at the excited point (Figure 2.7, point A). These extra anions disturb the system, and to oppose the change cations flow in from the neighbouring and unexcited points (e.g. point B). Point B will end up having more cations inside the membrane and an electrotonic potential (potential caused by passive properties of the membrane when it is stimulated by currents that do not induce excitation) is set up. When the potential at B reaches the threshold value, this point also starts the process of depolarization and draws cations from the next point. The process is then repeated along the axon, to a new set of points and the action potential is propagated with no loss in amplitude. SITE OF POLARIZATION M E M B R A N E PROPAGATION OF ACTION POTENTIAL-Figure 2.7 The propagation of action potential by current flow. The distance between two points on the same nerve divided by the time taken for the action potential to travel from one of the points to another yields the propagation velocity of the 13 CHAPTER 2 nerve. The propagation velocity is influenced by several factors. Firstly, it depends on the magnitude of the Na + influx, because more current is available to excite the neighbouring points. Since Vm0 is one of the driving forces of the sodium pump, a higher Vm0 (e.g. -50 mV instead of -90 mV) reduces the current available and a lower conduction velocity results. Secondly, higher propagation velocity results if the membrane capacity is lower (electrotonic potential rises faster) and the membrane resistance is higher (electrotonic potential falls less with distance). Thirdly, the larger the nerve diameter, the larger is the membrane area for ions to flow through, and therefore more current is available and higher is the propagation velocity. In myelinated nerves, the propagation of excitation is slightly different. The myelin sheath gives the membrane low capacitance and high resistance. Between the Nodes of Ranvier the action potential propagates very fast and most of the delays occur at the uninsulated nodes (see Figure 2.2). Myelinated nerves therefore have much higher conduction velocities than unmyelinated nerves. 2.2.6 Neuromuscular Transmission Nerves terminate on muscles at neuromuscular junctions. When a neural action potential reaches a junction, the presynaptic terminal of the nerve releases a transmitter substance to the synapse. The transmitter substance diffuses into the subsynaptic membrane and causes chemical changes that trigger the action potential in the muscle, causing it to contract. 14 CHAPTER 2 AXON Figure 2.8 A typical neuromuscular junction. 15 CHAPTER 2 2.3 EXTRACELLULAR ELECTRICAL STIMULATION OF NERVES Electrical stimulation has been routinely used in the study of peripheral nerve functions [3] [4]. Either surface electrodes or needle electrodes for more localized extracellular stimulation can be employed, although the former are more commonly used. The surface electrode is usually round or square in shape, and its size is about 1 cm by 1 cm. The cathode is taped onto the skin directly above the nerve to be stimulated, and the anode is placed at a fixed distance away over the same nerve. Stimulation of the nerve is achieved by applying a voltage or current between the two electrodes. 2.3.1 Basic Principles of Extracellular Electrical Stimulation A nerve cell in the vicinity of a cathodal electrode will see a localized change in the extracellular potential. The cations inside and outside of the cell will move towards the electrode and the intracellular potential will decrease to try to maintain Vm at the resting potential. Since the membrane does not allow free efflux of ions, and the cytoplasm does not allow free migration of ions, the intracellular potential cannot keep up with the changing extracellular potential. If Vm reaches the threshold value, the nerve will be depolarized. If the applied electric field is transverse to the nerve, hyperpolarization and depolarization will occur next to each other (Figure 2.9), and no action potential will propagate. Figure 2.10 shows that nerve cell can only be stimulated if there is voltage gradient along it. For this reason, to stimulate a nerve with bipolar electrodes, the electrodes should be separated by some distance along the nerve. 16 CHAPTER 2 cathode ' distance Vf - V0 < Vmt : hyperpolarization " > VJHO : hypopolarization - V0> VT : depolarization Figure 2.9 An unmyelinated nerve in the presence of a cathodal electrode. The applied electric field is transverse to the nerve. V^ - = the intracellular potential V0 = the extracellular potential Vmg = the resting membrane potential VT = the threshold potential 17 CHAPTER 2 cathode Vt - V0< Vm0 : hyperpolarization ' Vo* v » o •' hypopolarization Vt - V0 > VT : depolarization Figure 2.10 A unmyelinated nerve cell in the presence of an extracellular cathodal electrode. The applied field is longitudinal to the nerve. V^- = the intracellular potential V0 = the extracellular potential Vm0 = the resting membrane potential VT = the threshold potential From Figure 2.10 it can be seen that the electrode current, which is also the current induced in the tissues for a given applied voltage, has to flow through the cytoplasm resistance and the membrane resistance. Large axons have low cytoplasm and membrane resistances, and are therefore more readily stimulated than smaller axons. The voltage applied is also another important factor that determines if the nerve will be excited. However, in experimental 18 CHAPTER 2 situations the electrode voltage does not give a good indication of what the voltage next to the nerve is, and its calculation from tissue resistivity can be difficult because most tissues are non-homogeneous and anisotropic. It is practically much easier to use constant current for stimulation since the amount of current involved is known and the voltages adjust themselves to the tissue and nerve "circuit" values. 19 CHAPTER 2 2.3.2 The Cable Theory . We can visualize electrical stimulation of a nerve as a current source applied between two points along the equivalent circuit of an axon (Figure 2.10). In the case of subthreshold conditions the electrical stimulus propagates along the axon passively. EXCITATION x=x p2 Ti dx Ie (EXTERNAL) T r dx\ z^z cmdx <^ I Vi r2dx r2 dx T X r2dx r2dx / , ONTORNAL) Figure 2.11 An externally applied current (i^ in the equivalent circuit of an axon. n = cytoplasm resistance per unit length r„ = intracellular resistance per unit length = membrane resistance per unit length cm = membrane capacitance per unit length V, = intracellular potential = extracellular potential vm = membrane potential Vmo = resting potential h = intracellular current h = extracellular current 20 CHAPTER 2 The extracellular membrane potential is dx dx dJi dx -Jr (2.2) e 1 -Itr2 (2-3) tn If / = Ie + I., in the region where there is applied current (ip) di . The change in membrane potential (Fm) can therefore be expressed as follows: (2.4) (2.5) dVm dVe dV{ n ^ —1 - — t 1 - (r.+r,)J, - r.J (2-6) Differentiating and substituting results in dx2 (2.7) The membrane current im can then be expressed as: 1 tn T\+T2 dx2 + r l l p (2.8) = lml +lm2 21 ~m r m dt tn Combining the above equations, we obtain the cable equation: where X = rm M (r1+r2) and m m CHAPTER 2 From Figure 2.10, the membrane current can also be described as: V dV i = -H + c — * (2.9) dx2 dt m 1 p (2.11) T = r C (2.12) By assuming a solution Vm - v(x,t)e~tlx (2.13) and let the initial condition be Vm(x,0) - 0 (Vm0 can be ignored since it is a DC quantity and we are looking for transient changes), equation (2.10) is reduced to ^ # 1 _ § 1 . _ / ( } ( 2 < 1 4 ) t dx2 dt where f(x,t) - —(r.X2^ (2.15) 22 CHAPTER 2 If ip is a rectangular pulse of amplitude I0 applied between the points xpl and xp2, then the solution is v(x,t) - rikI0[F(x-xp2,t)-F(x-xpl,t)] (2.16) where F(x,t) is an elaborate expression [5] involving error functions. From our point of view it is particularly relevant that I0 is directly involved in the solution. Once Vm reaches threshold, the analysis proceeds very differently [2]. We will not discuss it here since it is quite involved, and it is not essential for the understanding of the background of magnetic stimulation. The site of depolarization can be derived from the above analysis. Since there must be current flowing through the membrane to cause depolarization, the site of depolarization is located at where the membrane current (im) is at its maximum. Then, according to (2.4), without an externally applied current ip, im is at its maximum where the gradient of the axial intracellular current I{ is the largest. With an externally applied current, im is the maximum when the sum of the spatial gradient of the intracellular current and the applied current is the largest. 2.3.3 Current as a Stimulus A current pulse of finite duration can be used to stimulate a nerve. The same response may be obtained from current pulses of different amplitudes (strength) and durations whose 23 CHAPTER 2 relationship forms a strength-duration curve for the nerve [4]. Strength-duration curves follow the empirical equation 7 -7 , (1 + —) (2.17) tp current amplitude rheobase current (current amplitude at DC) chronaxie (duration at which I = 2 7r) current duration where 7 = = C = xp = 4 6 Duration 10 Figure 2.12 A strength-duration curve, with C set to 1 on the abscissa. Ir is the rheobase current and C is the chronaxie. If current is flowing out of a nerve cell at the site of depolarization, current must be flowing into the nerve cell at some other sites, hyperpolarizing the nerve. Very large stimulating current may cause the hyperpolarization to be large enough to block the propagation of action 24 CHAPTER 2 potential beyond it. It is called anodal blocking. A similar phenomenon occurs when the stimulating current is bipolar. The positive phase of the current tends to depolarize the nerve but the negative phase reverses the effect. If the amplitude of the negative phase is large enough, the nerve will not be polarized. Sometimes a stimulus is not enough to depolarize a nerve since Vm does not rise above the threshold. If a second stimulus is applied before Vm settles down to the resting potential again, the effect of the subthreshold depolarization sums up and may depolarize a nerve that would not be normally depolarized by one single stimulus. Figure 2.12 shows the relationship between threshold and the number of pulses [6]. 2.3.4 Extracellular Stimulation of Nerves in Practice In practice, constant current stimulation is generally used for electrical stimulation. The amplitude and duration of the current employed varies with the nerve to be stimulated. For example, if myelinated nerves are to be stimulated, current pulses of about 50 JUS should be used but if unmyelinated nerves are to be stimulated much longer pulses are needed. In most electrical stimulators, both the amplitude and duration can be varied so that the user can select the best current pulse for the application. 25 CHAPTER 2 0.4 J , 1 , , 1 r 0 2 4 6 Number of Pulses (N) Figure 2.13 Relationship between nerve stimulation threshold and the number of stimulating pulses. 2.3.5 Stimulation of Muscles Most of the above discussion on nerve cells can also be applied to stimulation of muscle cells. The muscle cell must also maintain a dynamic balance in intracellular and extracellular ionic concentrations for it to be excitable. However, there are a lot more C a + + ions involved in excitation. The excitation of the muscle cell follows the same pattern as that of a nerve cell except that it has a much higher threshold value. 26 CHAPTER 3 MAGNETIC STIMULATION: THEORY AND PRACTICE 3.1 INTRODUCTION Electrical and magnetic stimulation are similar theoretically in the sense that they both aim to create currents in excitable tissues. Therefore, in specifying the characteristics of the magnetic stimulator, we can make use of some of the experimental and theoretical principles established for electrical stimulation. While the literature on electrical stimulation is very large, there are few papers on the detailed design of magnetic stimulators. This is likely due to the great commercial potential that the magnetic stimulation technique possesses. 27 CHAPTER 3 3.2 THE THEORY OF MAGNETIC STIMULATION In magnetic stimulation, a stimulating coil is placed near the surface of the body over the nerve to be stimulated, and a current pulse passing through the coil is used to generate an electromagnetic pulse in the body tissues [7]. The induced electric field (e) can be expressed as [8]: e(r,r) (dlt(t))f VeNrjT J V-A) dt 4n (3.1) where e(r,t) = induced electric field It(t) = coil current i i 0 = permeability of non-magnetic material N = number of turns in the coil dV = differential element of the coil r = position where the electric field is calculated r' = position of dV 28 CHAPTER 3 We therefore modify the cable model described in Chapter 2 by introducing an additional term to the electric field internal to the neuron: E. - - _ i + cx(x,f) (3-2) where Et = the total electric field inside the axon Vt = the intracellular potential ex = the x-component (along the axon) of the induced electric field and -r^ - cx(x,0 - (r 1 + r 2 )7. - r , / (3.3) dx 29 CHAPTER 3 Ve rtdx r2dx r2 dx f l dx Ie (EXTERNAL) r_dx< T cmdx 'mO T T Vi r2dx T r2 dx T r2 dx \/\/S/—1 s /v^^y—'v/^ J e^ x, 0 dx | ex(x, 0 <£x h (INTERNAL) Figure 3.2 Cable model with an induced field ^ (x.t). The membrane current (im) can be expressed as: 1 m dex(x,t) . ^ 2 dx (3.4) lml + lm2 From Equation (3.1), we can reduce iml to di. (3.5) where k is a constant depending on the permittivity of the tissues and the geometry of the coil. 30 CHAPTER 3 The modified cable equation can then be expressed as: 6*K. m dV. m de x (3.6) - V. m dx2 dt dx where Y -de X (3.7) dx This result is similar in form to the cable equation we derived in Chapter 2, with the term ip replaced by y. As before, depolarization occurs at sites where im are at its maximum, and therefore where the spatial derivative of the induced electric field is maximum. Unlike in electrical stimulation where the site of depolarization is always under the cathode, in magnetic stimulation this site is not clearly defined. Using the analysis above one can determine a "volume of stimulation" in which the fibres will be stimulated. The boundary of this volume defines the "virtual cathode" in the tissues. The size and shape of this virtual cathode depends on the coil current and the coil shape. For a given coil configuration we can map y around the coil and find these points of interest. Figure 3.3 shows such a mapping for a round coil. 31 CHAPTER 3 Figure 3.3 The equipotential lines of the axial derivative of the induced electric field. The points marked with + will be maximally hyperpolarized and the places marked with - will be maximally hypopolarized. It can be seen from (3.5) that the induced current im2 is directly proportional to the first time derivative of the coil current (/,) and that it has an equivalent term in (2.11) in the analysis of electrical stimulation. Therefore, assuming that no other phenomenon is involved, one can theoretically apply most of the principles of specifying the current pulse in electrical stimulation to magnetic stimulation, except that this current pulse would be the electrode current in the former and the induced current in the latter. 32 CHAPTER 3 3.3 THE PRACTICAL ASPECTS OF MAGNETIC STIMULATION OF PERIPHERAL  NERVES The general description of the mechanism of magnetic stimulation as presented in the previous section is a simplified model of the events at the cellular level. In fact, many aspects of these events are incompletely understood. Nevertheless, much work has been done on the clinical use of magnetic stimulation of nerves, especially on the stimulation of the motor cortex. In some hospitals magnetic stimulation on the motor cortex are performed routinely. Magnetic stimulation of peripheral nerves is also of much clinical interest, and this is where we will focus our attention in this section. To stimulate peripheral nerves in practice, the coil is placed close to the limb over the selected nerve, and current pulses are passed through the coil. There are three factors that affect the stimulation: a. the shape of the coil current pulse; b. the orientation and location of the coil; and c. the geometry of the coil. 3.3.1 The Shape of the Coil Current pulse As mentioned in Section 3.2, the optimum induced current in magnetic stimulation should be the same as the optimum electrode current in electrical stimulation. Theoretically, the induced current should be monophasic for maximum efficiency [9]; repetitive magnetic 33 CHAPTER 3 stimulus should be available for nerves not excited by single stimuli. In addition, it should be possible to determine a strength-duration curve for every nerve using magnetic stimulation [10]. At present, not all of these can be realized because of the limitations of the available magnetic stimulators. In most stimulators, a bank of capacitors is charged to a certain voltage (usually 2 to 3 kV) and then discharged to transfer the energy to the coil. In most reports on magnetic stimulation, the coil current is a short pulse that rises to a peak in about 100 us and then decreases slowly to zero. Although some researchers would specify the coil current used in their research, most of them prefer to specify the maximum magnetic field measured. The magnetic field amplitude mostly quoted is about 2 T [11] [12] [13], which requires thousands of amperes to produce. Equation (3.1) shows that the induced current is proportional to the first derivative of the coil current. Therefore, to have a truly monophasic induced current, the coil current should rise to a peak and maintain the peak (Figure 3.4). But this is difficult to implement since the stimulator would have to continuously generate current in thousands of amperes and the coil must be able to carry such a current. For the same reason, it is difficult to deliver repeated stimuli to nerves not excited by a single stimulus. 34 CHAPTER 3 (a) (b) Figure 3.4 The ideal (a) and the practically available (b) coil and induced currents. /, = coil current Ij = induced current 3.3.2 Orientation and Location of the Coil It is desirable to place the coil over the nerve such that it operates with the highest efficiency, i.e., it stimulates and depolarizes with the least energy. According to the model presented in Section 3.2 [8], a round coil should be placed such that the nerve and the coil are on parallel planes, and not quite tangential to the nerve (Figure 3.3). Experiments have proved the above to be true by comparing the action potential generated by coil oriented in different ways [13] [14]. In the past, it was thought that stimulation occurs at the point of maximum magnetic field, e.g. at the centre of the coil, but this coil 35 CHAPTER 3 orientation induces current that flows transverse to the nerve and is therefore ineffective in exciting the nerve. As mentioned before, it is difficult to pinpoint the site of depolarization on the nerve in magnetic stimulation, and for the most part it is difficult to stimulate one nerve without affecting neighbouring nerves. It was reported that with careful placement of the coil, it is possible to selectively stimulate the median nerve in the arm without activating a neighbouring ulnar nerve [13], and with high current, the median nerve can be stimulated supramaximally. However, stimulation is very sensitive to the location of the coil [14]. In some experiments, it was even found that the action potential produced by magnetic stimulation had a higher amplitude than that produced by electrical stimulation [13] possibly because the stimulation excites motor axons in a temporal pattern which results in less phase cancellation between extracellular action potentials generated by different motor units. According to the analysis in Section 3.2, if a nerve is stimulated by a coil with the current flowing in one direction and then stimulated again with the current direction reversed, the latency periods from the two stimulations should differ by the distance between the sites of depolarization and hyperpolarization divided by the conduction velocity. By experiment it was found that the difference is really negligible [13] [15] and the conclusion was that the "virtual cathode" and "virtual anode" are located very close to each other. Because of this reason, the interaction of spatial and temporal effects of magnetic stimulation can cause error in propagation velocity measurements. Suppose a stimulation coil is placed over the 36 CHAPTER 3 nerve and the nerve is stimulated. At time t0 (Figure 3.5) the nerve is depolarized at location x0 and hyperpolarized at xv Two action potentials (API and AP2) propagate from x0 in opposite directions along the nerve with a finite speed. Since the coil current changes with time, the induced voltage along the nerve also changes. Depending on the temporal characteristics of the coil current /„ xx can be depolarized or hyperpolarized at tv Suppose Xj is depolarized at tlf then another two action potentials (AP3 and AP4) will propagate from Xj. At t2 AP4 will collide with API, which may propagate on or be cancelled. After t2 there may be two action potentials (AP2 and AP3) propagating in opposite directions with different apparent latency periods and origins (AP2 from x0 and AP3 from J C , ) . distance along axon A depolarized at t = tj depolarized p Figure 3.5 The spatial and temporal effects of the coil current on magnetic stimulation of a nerve. 37 CHAPTER 3 3.3.3 The Geometry and Construction of the Coil Although round coils are still widely used the 8-shaped coil is gaining acceptance rapidly. The 8-shaped coil concentrates the electric field at the intersection of the two wings of the coil, and theoretically produces a more focused stimulation [8] [16] [17]. This has been verified by experiment [8]. The tear-drop shaped coil gives a focused stimulation because the spatial derivative of the induced field is high at its vertex. 9 § ROUND 8-SHAPED OBLONG TEAR-DROP Figure 3.6 The different coil shapes and directions of current flow. In addition to shape, other issues must also be considered: a. the type of wire to be used; b. the type of winding: spiral or solenoidal; c. the number of turns and dimensions of the coil; and d. the type of wire insulation. Although stranded wires are lower in resistance, solid wire is preferred since no supporting structure is required to maintain the coil shape. If solid wire is used, there is no obvious 38 CHAPTER 3 advantage of selecting square wire over circular wire. The number of turns and dimensions of the coil directly affect the load presented to the stimulator. Because the windings should be very tight to produce a focused field, spiral winding is not recommended. Since the coil carries a large amount of current and usually a large voltage, it should be insulated very carefully. Because of the high current, the repulsion force among the windings is very high. At the same time, the coil heats up and expands since a lot of heat is generated. The insulation used should therefore be flexible and thermally conductive in addition to having good insulating properties. Unfortunately, insulation like this is not easy to find. In laboratory prototype, the electrician's tape has all the desirable properties mentioned [18], but for the coil to be used in a clinical environment, some sort of potting compound must be used. In designing or building the coil, the designer must compromise among several things: the efficiency of the coil, its focusing ability, the electrical limitation of the stimulator and the mechanical limitation of the insulation material. A large coil does not produce very focused field but it does not cause overload to the stimulator by drawing excessive current. A small coil may be focal but can overload the stimulator since the coil current is shorter in duration but higher in amplitude. Even if the stimulator can deliver the high current, the repulsion force among the windings may stress the potting compound so much that the it explodes [19]. 39 CHAPTER 3 3 A MAGNETIC STIMULATION OF THE MOTOR CORTEX Magnetic stimulation of the motor cortex is used largely in the diagnosis of motor neuron disorder [11] [12] [20] [21] [22] [23] although it has also been used to map the functions of the brain [24]. The stimulating coil is placed over the head close to the area to be stimulated. The stimulus is applied and the conduction time to the muscle innervated by the area is measured. Subjects with central demyelination diseases (e.g., multiple sclerosis) typically have a slower conduction time than normal subjects. Usually, electrical stimulation must be used first as a reference so that the exact location of stimulation can be determined. Once the reference is established, only magnetic stimulation is needed for further diagnosis on the same nerve. The main advantage of magnetic stimulation over electrical stimulation is that it is not painful. In electrical stimulation, current flows through the skin to the underlying structures and therefore can be very painful if the current is large as in the case of motor cortex stimulation. 3.5 MAGNETIC STIMULATOR TECHNOLOGY The current pulse used in most of the research is of the shape similar to that in Figure 3.4(b). The rise time of the current pulse is about 100 fj.s and the coil current is such that it produces a peak magnetic field of about 2 T at the centre of the coil. Stimulators are basically a bank of capacitors charged up to a high voltage, and then the energy is discharged to the coil. The 40 CHAPTER 3 electrical circuit is similar to that of a magnet charger or defibrillator. Two major models of magnetic stimulator models are used in research: Cadwell MES-10 and the Novametrix 200. In addition, there are also prototypes available from other manufacturers. All models make use of the principle of capacitor discharge using one bank of capacitors i.e., the rise time of the current pulse is not adjustable except by changing the coil. 3.6 SAFETY OF MAGNETIC STIMULATION There are now more and more studies on the short term and long term effects of magnetic stimulation [17] [25]. There has been no report of the long term effects of magnetic stimulation. In the short term, it was found that magnetic stimulation of the motor cortex may block the subject's ability to see and remember images (during the actual stimulation). Once the stimulus is taken away the subject returns to normal. The effects of long-time and short-time exposure to magnetic field is of concern to both the researchers and the subjects involved. The National Radiological Protection Board (NRPB) of Great Britain recommends that the maximum rate of change of magnetic field dB/dt to be less than 20 T/s for exposure durations (r p) exceeding 10 ms [6]. For a magnetic field of duration xp < 10ms, allowed exposure may be increased in the order of r~ 1 / 2 (Figure 3.7). The magnetic pulse produced in magnetic stimulation typically rises sinusoidally to a maximum magnetic field of 2 T within 100 ps. Adjusting for the fact that the pulse is not rectangular 41 CHAPTER 3 (equivalent to a rectangular pulse of duration 100 jus and amplitude T/2 T), the peak dB/dt is 15.7 kT/s which is over 10 times the safety limit (safety limit is 200 T/s). However, clinicians have tried stimulation on various parts of the motor cortex and around the neck but no obvious side effect was observed. ^ 10 • b 10 10 10 10° to1 magnetic stimulation safety limits by NRPB , . 3 , „ 4 . 5 10 10 10 10 Figure 3.7 Safety limits for whole body exposure to time varying magnetic fields and the exposure the subject receives at -JT/2 T within 100 lis. 3.7 OBJECTIVES Since all magnetic stimulator models used in research are limited to only providing current pulses of variable amplitude but not rise time, it is impossible to evaluate the effect of the current rise time on magnetic stimulation. It is the aim of this work to design and build a versatile magnetic stimulator that is capable of delivering coil current of variable rise time and 42 CHAPTER 3 amplitude, in order to provide researchers and clinicians with the capability to fully control the stimulus. From Figure 3.7, we can see that in current practice of magnetic stimulation, the subject is exposed to a very high magnetic field. It is therefore also our goal to minimize the exposure of the subject to magnetic stimulation by obtaining the optimized amplitude duration ratio in peripheral nerve stimulation and to use the results to study stimulation of motor cortex. 43 CHAPTER 4 DESIGN OF THE MAGNETIC STIMULATOR 4.1 INTRODUCTION In the design of our new magnetic stimulator, we must consider its use in both the research and the clinical environments. To be a research tool the instrument should be able to produce coil currents of different shapes, amplitudes and durations (or rise times) so that it can be used in the study of the various aspects of magnetic stimulation. For the instrument to be used in the clinical environment, reliability, safety, operator convenience, and simplicity of operation are important. Based on such considerations, the versatile magnetic stimulator we aim to create must have: a. a very high degree of reliability and safety; b. coil current rise time (t^) and amplitude (7^,) that are adjustable by the user over a wide range; and c. current pulses not affected by the frequency of operation. In this chapter, we will first describe the basic concepts of the magnetic stimulator, then the overall design and lastly, the detailed design of the electrical part of the stimulator. 44 CHAPTER 4 4.2 THE BASIC STIMULATOR CIRCUIT The function of a magnetic stimulator is to provide a short but intensive current pulse to the coil. There are many ways of generating such a pulse [26], including capacitor storage, inductor storage, battery storage and rotating machine. The most common way to use energy storage in a capacitor bank. This approach offers flexibility and high energy transfer-efficiency, although it also has the lowest energy to volume ratio. In addition, for pulses of duration in the order of milliseconds, the capacitor bank approach is the least expensive. Capacitors are also more reliable than inductors or batteries, and the ease with which they can be controlled makes them the choice in this application. The basic conceptual circuit of the stimulator is a capacitor charged by a source voltage when a switch (SWj) closes, and the capacitor discharges to an inductive coil when a second switch (SW2) closes (Figure 4.1). The resistor (R) represents the internal resistance of the capacitors, the coil, the connectors and the wires. It must be kept as low as possible to minimize energy loss and excessive heating in the discharge path. 45 CHAPTER 4 Figure 4.1 Simplified circuit diagram of the charge/discharge circuit. Extensive search of the literature indicated and was confirmed by our own experience that coil inductances in the range of 5-10 /iH, and capacitances in hundreds of jwF's, are required. With a low resistance, the circuit is underdamped and its output current (/,) during discharge is given by the standard equation: where V ( R \ — - exp 1 sinof oL { 2L ) (4.1) [LC 4L>) (4.2) 46 CHAPTER 4 with Y - -K 2 \ C L (4.3) and to0 -JLC (4.4) The first maximum of the current (1^) and the time taken to attain this maximum (t^) are: \ L exp -arcsin m.\J 1 -• (4.5) tt = \ILC—^—arcsiny/1 - y2 V7! - Y 2 (4.6) Since the circuit is underdamped, the current it will oscillate, and therefore induce an oscillating current during stimulation. As discussed in Chapters 2 and 3, monophasic current is more effective for stimulation. The induced current will be very close to monophasic if it decreases slowly after it reaches its peak. We can achieve this by placing a flywheel diode (d in Figure 4.1) directly across the coil. This diode will take the current when the capacitor voltage goes to zero (a short time after i, has reached its peak if r is small). The current /, then decays 47 CHAPTER 4 slowly in the diode-coil loop. Assuming that the diode is fully on instantaneously, the coil current will be: for t > ttl (4.7) where Ia and ta = the value of It when Vc = 0, = the time when V=0. The time derivative of the coil current is therefore: dt <x>L R ( R \ ( R \ 1 — exp 1 sinwf + w exp 1 coswf for t <; ttl (4.8) RT ( R . for t > ttl Figure 4.2 shows the output current I, and its derivative assuming R = 50 mfi, L = 8 ^H, and C = 1200 fiF. The peak of 7, was normalized to 1. 48 CHAPTER 4 0 100 200 300 400 500 600 t(us) Figure 4.2 Coil current and its derivative assuming that R = 50 mfi, L = 8 /tH and C = 1200 /tF. 4.3 OBSERVATIONS AND PILOT STUDIES In the published literature, descriptions of the design of magnetic stimulators or stimulating coils used are usually very brief, and detailed descriptions are non-existent. The general impression is that the researchers build and test a unit clinically without making a serious attempt to establish the optimum shape of the current. For this reason, we had to perform pilot studies to establish design parameters and also to prove design concepts. In particular, we have constructed a 450 V magnetic stimulator [27] using the design parameters published by Barker et al [28] as a guide. Our 450 V instrument has a total of 49 CHAPTER 4 ten 2200 uF capacitors, all charged and discharged as a unit. It has been tested with different stimulation coils, but we have not been able to get reliable nerve stimulation with this design. Nevertheless, this instrument allowed us to make different voltage and current measurements at the voltage of 450 V. This unit also gave us the opportunity to experiment with construction details to keep the inductance and resistance at a minimum. For example, R of 35 mfi was achieved. With this R, if we let L be 8 ttH, a maximum capacitance C of 1200 iiF is needed to produce a maximum rise time of 150 its. Our review of the literature yielded the following conclusions: a. in recently published papers, the magnetic stimulator voltage varies from 2 kV to 3 kV, and the very high voltage (20 kV) or lower voltage (450 V) approaches evident in earlier literature have been abandoned. 20 kV systems require more expensive components, and the insulation problems are more demanding. 450 V systems cannot deliver the current needed for reliable stimulation. b. a coil current rise time of 120 us is used very often in the magnetic stimulation of the motor cortex; and c. the reported peak magnetic field measured from coils that can successfully stimulate is about 2 T. We have decided that the coil would be have an inductance between 5-10 uH, and we would like to make 4, variable from 50 us to 150 us in steps of 10 us or less. The above information is 50 CHAPTER 4 adequate to formulate the specifications of our stimulator, including the maximum voltage, the maximum coil current amplitude 1^ and the range of the current rise time t^. A A PRELIMINARY DESIGN CONSIDERATIONS Before we set the final specifications of the stimulator, we must decide on its maximum operating voltage and current. We must also decide the features that it should have. 4.4.1 Electrical Ratings of the Components We control the amplitude and rise time of the output current pulse by varying the voltage supplied to the capacitors, and the amount of inductance and/or capacitance in the discharge circuit. The maximum design voltage is determined not only by technical consideration but also on the price and availability of the major components, namely, the capacitors, the switches and the wires. It was found that few components are available above the rated voltage of 2 kV, while many are available at 2 kV or below. The maximum voltage was therefore set to 1600 V, giving the components a safety margin of 400 V. To simplify calculations, we can approximate the coil current rise time by letting the circuit resistance R be negligible. By substituting R = 0 into Equation (4.6), we obtain LC - -t)p (4.9) 51 CHAPTER 4 This equation shows that in order to make variable, either L or C must be variable. Because it is easier to switch capacitor banks in the circuit then to change coils, we must determine the maximum capacitance needed and then divide this capacitance into several banks so they can be selected in different combinations. As we have shown earlier, to obtain a maximum t¥ of 120 /xs with L between 5-10 fxH, a total capacitance of about 1200 fxF is needed. It is desirable to be able to generate pulse trains so that researchers and clinicians can stimulate repeatedly with the pulse shape. This capability is also useful for discharging the capacitors before the system is switched off. 4.4.2 Protective Features Because of the high voltages and large currents involved in the stimulator, sensors that warn the operator of possible over-voltage can prevent major failures. If the supply voltage exceeds 1600 V, some of the components may be operating outside of their rated voltages and be damaged. A sensor sensing the supply voltage ensures that the user will be warned of such a problem. Another essential protective feature in the stimulator is a temperature sensor for the stimulating coil. During discharge a large amount of heat is dissipated in the coil which is tightly wound and is surrounded by insulation. If discharge is repeated rapidly the coil can 52 CHAPTER 4 become very hot and the insulation may break down. Therefore, a temperature sensor is needed to disallow discharge if the coil gets too hot. 4.4.3 Operator Interface The only electrical quantity we can measure directly is the voltage output of the power supply. However, this parameter does not tell us immediately what the output current looks like. In order to make this tool more useful for research purposes, the coil current rise time and amplitude 1^ should be measured and displayed to the user. 4.5 TARGET SPECIFICATIONS OF THE MAGNETIC STIMULATOR Based on our pilot studies and the above calculations we came up with the following specifications for our magnetic stimulator: a. maximum voltage = 1600 V b. maximum current = 16 kA c. coil current rise time is selectable in steps of 10 us or less d. two modes of operation: single shots and repetitive mode (at 1 or 2 Hz at reduced voltages) e. coil over-temperature protection f. over-voltage detection 53 CHAPTER 4 4.6 DETAILED DESIGN OF THE BASIC BLOCKS OF THE MAGNETIC STIMULATOR The magnetic stimulator is divided into six modules as shown in Figure 4.3: a. power supply; b. capacitor charge/discharge; c. SCR triggering; d. operator interface; e. sensors; and f. microcontroller and logic. 54 OPERATOR INTERFACE SCR TRIGGERING I MICRO-CONTROLLER AND LOGIC POWER SUPPLY I CAPACITOR CHARGE/ DISCHARGE I SENSOR Figure 4.3 The six functional modules of the magnetic stimulator. 55 CHAPTER 4 4.6.1 The Power Supply Module The power supply converts the line voltage of 115 V AC to a DC voltage variable from 0 to 1600 V. We have chosen the simple design of a full-wave rectifier. The line supply is connected to a variable transformer whose output is variable from 0 V to 115 V and is connected to the primary of a 120/1200 V transformer. The output of this second transformer is then full-wave rectified to give a 1600 V peak supply. Since a 120/1200 V transformer is not a standard item for most manufacturers and a custom one costs over $1200, we connect two 120/600 V transformers in parallel in the primary and in series in the secondary. Because one of the transformers is floating at 600 V above ground, it must be physically isolated from ground by insulators. The interwinding insulation of these transformers is rated up to 2000 V, and therefore it is judged to be adequate. At the output of the full-wave rectifier is a low-pass filter, followed by a power resistor which limits the charging current to the capacitors. The ratings of the transformer and the resistor were calculated such that the capacitors can be charged within 250 ms. The stimulator was originally designed to operate with a maximum frequency of 2 Hz at 1600 V. The rating of the variable transformer is calculated to be 3 kVA, and the transformers are also 3 kVA each. The capacitors in the fdter are electrolytic capacitors rated at 450 V and each has a voltage equalizing resistor placed across it. 56 CHAPTER 4 It should be pointed out that this power supply design is simple but the resulting unit is large and heavy since each 120/600 V transformer weighs 61 lb. A switching power supply is smaller, lighter and more efficient than a simple full-wave rectifier. It can also be electronically controlled, which makes it an even better choice in a microprocessor based system like this one. However, the design and building of such a power supply was outside the scope of this project. 57 120 V line input <r\j>-120 V 5 A Tl T2 120/600 T3 120/600 D. D„ D, R f i 1M2 > Rf2 1M2 ' R f 3 1M2 Rf4 1M2 "^OR" "-fl 2200 pF r _ f2 2^200 pF 2200 pF -^f4 "2200 pF 1600 V DC -O Figure 4.4 The 1600 V DC power supply. 58 CHAPTER 4 4.6.2 Capacitor Charge/discharge Module The capacitor charge/discharge module is composed of several banks of capacitors, each of which is each controlled by a set of switches for charge and discharge. To use the minimum number of switches possible, they are divided into five groups of unequal values of 50 uP, 100 uP, 200 up, 300 uP and 500 iiF respectively. The 50 uF is necessary to get a 10 us coil rise time resolution at low capacitance. The switches are silicon-controlled rectifiers (SCR's), and to prevent oscillations in the coil current, a flywheel diode is placed across the coil. o-1600 V max 50 uF °1 O-"100 uF '200 uF 400 uF 500 uF D l£o6 JL<?7 j£os A^Q9 A ^ Q 1 0 Figure 4.5 The capacitor circuit. 59 CHAPTER 4 4.6.2A The Selection of Capacitors Although electrolytic capacitors are used often for their compactness and low price, they are really unsafe for this application. Fast charging and discharging stresses them and may even cause internal shorts in which case the capacitors may explode. In this stimulator oil filled capacitors which are specially designed for pulse discharge purposes (e.g. laser power supply) are used. These capacitors have very low internal resistance and inductance. However, they have their own limitations in that they are bulky, expensive and if they are stored charged for extended periods of time their lifetime will be reduced. The capacitors used in this application are rated at 2 kV, and the standard stock capacitors come in 50 pF, 100 itF and 185 uF values. We have chosen to use the 100 u¥ and the 50 tiF units. Each capacitor is capable of providing discharge peak current of up to 25 kA, and the inductance of each capacitor has an internal inductance of about 0.1 uU as specified by the manufacturer. 4.6.2.2 The Selection of the SCR's The switches that control the charging and discharging of the capacitors are silicon-controlled rectifiers (SCR's). An SCR is a semi-conductor device that can be turned on with a pulse of current applied to its gate. The SCR will keep conducting current even if the gate current is removed as long as the current passing through it stays above the holding current value. There are three major types of SCR's: the phase control type which is designed to operate at 50 to 60 Hz, the inverter type which is 60 CHAPTER 4 designed to be used in high frequency switching circuits, and the gate turn-off (GTO) type which can be switched on and off by applying current pulses to its gate. Because each type of SCR is designed for a specific application, it has its advantages and disadvantages. The phase control SCR's have the highest current and voltage ratings among all three types of SCR's. But this type of SCR is optimized for 50 Hz to 60 Hz operation and therefore high frequency current can damage it by overheating one spot in the junction. The inverter type SCR is constructed with the gate connected to many spots on the junction and is therefore ideal for switching power supplies. The gate turn-off SCR's are the most difficult to construct. They have the lowest voltage and current ratings among all three types of SCR's. For sake of completeness it should be mentioned that we have thought about using vacuum relays as switches in this application but the high current involved, especially in the discharge circuit, would fuse the contact. 4.6.2.3 Calculation of the Ratings of the SCR's We can calculate the coil current amplitude 7^  and rise time from Equations (4.5) and (4.6). If any one of the five banks of capacitors is selected alone, with all other four inactive, it will carry some maximum current in the shortest rise time. We will therefore use these maximum currents to determine the current rated values of the SCR's (Table 4.1). 61 CHAPTER 4 Table 4.1 Coil current amplitude Ilp and rise time tlp calculated with V,. = 1600 V, R = 50 mfi and L = 5 /tH and 10 /iH. CQtF) L = 5 /iH L = 10 /iH K 0*s) M k A ) ttp (MS) /*(kA) 50 23.7 4.50 33.9 3.29 100 32.8 6.07 47.3 4.50 200 45.2 8.07 65.6 6.07 400 61.7 10.5 90.4 8.07 500 68.1 11.4 100.0 8.81 1250 100.0 15.3 150.0 12.3 The selection of each SCR is governed by several factors: the maximum voltage applied across it, the maximum current it conducts, and the maximum rate of change of current dl/dt (related to heat dissipation) that is impressed on it [29]. In our application the RMS current is very low but the transient current is very high. In fact, in the discharge circuit, all the currents are transient currents, and we have to operate the SCR's very close to the maximum transient current (ITS!A rating. If the device is not to fail prematurely the maximum dl/dt (governed by the hottest spot in the silicon) applied to the device must never exceed the rated value. Although switching type SCR's are probably more suitable for this application because of its fast turn-on time and high dl/dt capability, they are generally rated lower in voltage, more expensive and less available. Phase controlled SCR's that are rated very high in 7 r e M and dl/dt are used here. We selected the devices with lTSM about one and a half times to twice 62 CHAPTER 4 of the maximum peak current calculated above (L = 5 /iH). The maximum dl/dt impressed on each discharge SCR can be calculated by substituting t = 0 into Equation (4.8): dt (4.10) For Ve = 1600 V and L = 5 /iH, dl/dt is 320 A//ts. Based on the above calculations we selected ten 2000 V SCR's, five for charging and five for discharging. They are manufactured by the International Rectifier. The current rating for each of these SCR's are presented in Table 4.2. Table 4.2 The current ratings of the ten SCR's in the stimulator. See Figure 4.5 for the SCR designations. SCR Function Capacitor IRpart# dl/dt controlled (kA) (A//is) QrQs charging S23D20A0 5.7 800 HrQs discharge crc3 S30D20A0 11.3 800 Q9 discharge c4 S34D20A0 13.6 800 Qio discharge C5 S38B20A 15.5 800 63 CHAPTER 4 4.6.2.4 Calculation of the Rated Values of the Flywheel Diodes During discharge, the capacitor voltage drops from the full voltage to zero, and will start to reverse if there is no flywheel diode placed across the coil. The flywheel diode is forward biased when the capacitor voltage starts to reverse, and the coil current will start flowing through the diode instead of the capacitors. The current then decays exponentially to zero. The coil current is almost at its peak when the diode takes over, and therefore the diode must be rated at 2 kV and the full current rating of the stimulator (20 kA). We used three diodes, each rated at 2 kV and surge current of 12.8 kA and connected them in parallel. 4.6.2.5 Limitation on the Charging/Discharging Frequency Because the charging circuit is an RC circuit, the charging current surges to its maximum when the SCR's first switch on and then the current decays exponentially. During charging, each selected charge SCR's conducts as long as the SCR is forward biased and the current flowing through it is higher than its holding current (about 250 mA). When the charging current drops below this level, the capacitors are charged slowly by current pulses of duration of 50 /*s every 500 fis (whenever the SCR triggering circuits are on). To ensure that the capacitors would be fully charged before discharge, and to prevent the event in which the charge SCR's are still on when the discharge SCR's are gated (the voltage source is short-circuited to the coil), no discharge is allowed until five seconds after the beginning of the charge cycle. This 64 CHAPTER 4 limits the maximum repetitive rate of the charge/discharge cycles to 1 pulse per 5 seconds. 4.6.3 The SCR Triggering Module To turn on the SCR's, a pulse of current must be provided to the gate of the device and the SCR must be forward biased. In this application, the SCR's are controlling a circuit at 1600 V maximum whereas the triggering currents to their gates are provided by a 12 V circuit [30] [31]. Therefore, the high voltage circuit must be isolated from the low voltage circuit by some means. Although opto-isolators are generally insulating up to 3000 V, but they require a separate power supply on the high voltage side to switch them on. A pulse transformer directly couples the excitation current from the primary side to the secondary side, eliminating the need for an extra power supply, and is therefore ideal for this application. To turn on the SCR's, a gate current of 250 mA of duration at least 10 us is required. To reduce the SCR turn on time delay, the gate drive current is a rectangular pulse of 1 A peak. The circuit consists of a driver integrated-circuit DS0026 controlling a MOSFET (Figure 4.6). The current which flows through the primary side of the pulse transformer is coupled to its secondary side, driving the gate of the SCR. A forward diode is used in series with the SCR gate to prevent reverse current from damaging the SCR. Whereas the discharge SCR's are only triggered once for every discharge, the charge SCR's are triggered by current pulses. This is done to limit the power delivered to the SCR gate by the driver 65 CHAPTER 4 circuit. If the gate current is on continuously the rated gate power of 2 W will be exceeded. Moreover, it will also overload the driver circuit. The triggering frequency is arbitrary but it should be much higher than 120 Hz. The higher the frequency is, the faster the charging is completed. Triggering the charge SCR's at 2 kHz with a duty cycle of 20% was found by experiment to be a good compromise between the charging time and the amount of load imposed on the driver circuit. INPUT AT ABOUT 2 - 3 V DS0O26 10 R - ^ / w — 12 V IE IRF 630 5R SCR GATE > > SCR CATHODE Figure 4.6 The SCR triggering circuit. 4.6.4 Operator Interface Module This peripheral circuit consists of a six-key keypad and a small liquid crystal display (LCD) panel. The user can view and change the status of the stimulator by entering commands from the keypad. The six keys and their functions are: RUN request for discharge INC/CHARGE charge or increase quantity specified in MENU 66 CHAPTER 4 DEC decrease quantity specified in MENU MENU brings up menu to change various quantities RESET reset quantities to default values in MENU DATA shows the measured and 1^ in the previous discharge A small dot-matrix LCD display panel capable of displaying a maximum of 16 characters shows the status of the stimulator and other messages. The display has an internal ROM for the standard fonts and its own driver circuit. All the user is required to provide are the supply voltages and the data representing the characters to be displayed. The characters are sent from the microcontroller as serial data before being converted to parallel data for display. 4.6.5 Sensor Circuits There are three sensor circuits in the stimulator: a. the current sensor that measures amplitude and rise time (7^ , and t^) of the coil current (7,). b. the temperature sensor that senses the temperature of the discharge coil; and c. the voltage sensor that measures the voltage across each group of capacitor. All sensors provide input to the microcontroller and the logic module. The first two sensors are only monitored before and during discharge, but the voltage sensor is being monitored constantly. In fact, in the case of over-voltage the microcontroller will be immediately 67 CHAPTER 4 interrupted. All SCR gating circuits will be disabled and a warning message will be displayed until the voltage has dropped below 1700 V. 4.6.5.1 The Current Sensor The current monitor is a toroid type current transformer with a distributed internal resistance (Pearson 101). To measure a current waveform the current carrying conductor is passed through the centre of the toroid. The input current is reproduced at the output with a 0.01 V/A ratio. This current monitor can measure up to 50 kA (with a bias current of 1.8 mA into the monitor) with an accuracy of 1% and a droop rate of 0.0001 %//xs, and it saturates at a maximum current-time (IT) product of 2.5 Amp-sec. Without the bias current it will operate up to 1/3 of the maximum IT product. In the discharge circuit, the wire carrying the coil current J, is made to pass through the centre of the toroid and a small bias current of about 1.8 mA is fed to the current monitor. The output of the current monitor is scaled down by a 1:16 divider and low-pass filtered before it serves as the input to a peak amplitude detector (or simply peak detector) and a peak time detector (Figure 4.7). 68 CHAPTER 4 COIL PEAK TIME DETECTOR CURRENT MONITOR I VOLTAGE DIVIDER ZERO-CROSSINQ DETECTOR ' TIMER MICRO LOW-PASS PEAK FILTER DETECTOR ADC Figure 4.7 The zero-crossing detector and the amplitude peak detector. The peak time detector consists of a differentiator, and a zero-crossing detector [32]. When the coil current I, peaks, its differentiated output is zero and the zero-crossing detector (ZCD) changes state. The period of time from the beginning of the discharge to the time the ZCD changes state is the current rise time t^. When the ZCD changes state, the A/D converter in the micro is activated to convert the output of the peak amplitude detector to a digital representation of its value (1^). Although the current sensor works well, there is a small timing error that comes with the design. If an ideal differentiator were used, t¥ will be exact within the tolerance of the measurement. However, we have chosen to use an operational amplifier 69 CHAPTER 4 differentiator, and to ensure that this circuit is stable a small capacitor must be placed at its input [33]. This introduces a few degrees of phase lag in the circuit output, ie, the output zero-crossing lags the input peak by a few /AS. Overall, the error in measured is < 5% of the true value. Fortunately, the amplitude detection is not affected, since the peak detector, which has very low leakage, is still holding the peak amplitude value when the zero-crossing is detected. No noticeable error in the measurement of 1^ was observed. 4.6.5.2 The Temperature Sensor and the Voltage Sensor The temperature sensor is a precision diode (LM335) attached to the coil to measure Tc. The output of the diode is filtered and then fed to a comparator whose output is normally low. The comparator switches to a high output if Tc > 45 °C. The sensor has a hysteresis of 10°. The voltage sensor is connected across the filter capacitors located at the output of the full-wave rectifier. It consists of a 200k resistor connected in series with the input light-emitting diode of an opto-isolator. The current flowing through the diode is equal to the voltage to be measured divided by 200kQ. This current controls the diode output light intensity. The photo-transistor in the opto-isolator senses the light and produces an output voltage proportional to the light intensity. If the voltage across the power supply is larger than 1800 V, the sensor generates a hardware interrupt to the micro 70 CHAPTER 4 which stops triggering currents to all SCR's. Operation will resume if voltage drops to below 1700 V. Figure 4.8 The voltage sensor circuit. 4.6.6 The Microcontroller and Logic Module This module is responsible for controlling and monitoring the following circuits: a. the SCR triggering module; b. the monitoring of variables that must not exceed design limits such as coil temperature (Tc) and capacitor voltage (Vc); c. the measurement of the output current rise time (t^) and amplitude (1^); d. the monitoring the keypad for any operator input; and e. the display of messages. 71 KEYPAD DISPLAY MICRO-CONTROLLER AND LOGIC CIRCUITS 1600 V PEAK POWER SUPPLY CHARGE SCR'S TRIGGERING TEMPERATURE SENSOR DISCHARGE SCR'S TRIGGERING CAPACITOR CHARGE/ DISCHARGE MODULE COIL VOLTAGE SENSOR Figure 4.9 The microcontroller and the associated circuits. 72 CHAPTER 4 The versatile Motorola 68HC11 microcontroller was chosen for this application. It has five ports, including input and output timers, A/D converters and serial I/O lines. The following table summarizes the ports available and any special functions that they may have. Table 4.3 The functions of the five ports of the 68HC11 microcontroller. Port # of input only lines # of output only lines # of I/O lines Special functions A 3 4 1 timers B 0 8 0 none C 0 0 8 none D 0 0 6 serial data I/O E 8 0 0 A/D converter 73 CHAPTER 4 Table 4.4 summarizes the functions implemented, and the port assignment: Table 4.4 Function assignment to the different ports of the microcontroller. Function Port Serial Parallel Interrupt SCR selection D X SCR triggering B X Temperature sensing E X Voltage sensing E X X measurement A X X Itp measurement E X Keypad E X Message for display D X Display enable B X This leaves most of port A and port C free for any future expansion. 4.6.7 SOFTWARE DESIGN The software is designed to integrate the operation of all the modules through the I/O ports of the microcontroller (micro). It controls the selection of capacitors and the timing for charging and discharging. The sensor outputs are monitored in case the operating parameters may be exceeded. It also monitors the keypad for any command entered by the operator, and appropriate messages are displayed after each action. The output current rise time and amplitude are measured after each discharge and are displayed on the LCD panel. 74 CHAPTER 4 After the micro comes out of reset, the software initializes the registers that control the operation of the micro. This include defining the location of the RAM's and registers, enabling the A/D converter, and setting up port D for serial data transfer. The micro then sets the SCR selection to the default value (850 u¥) and sets the mode of operation to single pulse charge/discharge. All charge and discharge SCR's are disabled. The message "READY" is displayed and the micro waits for keypad entries. This is called the standby mode. The function of each key is summarized in the section "Operator Interface Module". If the key RUN is pressed, the stimulator enters the discharge mode. The micro first checks the temperature of the coil. If Tc > 45 °C, it will stop the triggering currents to all charge SCR's and wait for 5 ms to ensure that the SCR's are truly off before triggering the discharge SCR's. Right after the discharge has been enabled, the micro checks the mode of operation. If it is in single pulsed mode, the micro stores the value of the internal timer and enables the interrupt controlled by input timer 2 (TIC2) which is connected to the zero-crossing detector. If it is in the multiple discharge mode, the micro waits for 2 ms and then immediately jumps to the charging routine, and the interrupt controlled by TIC2 is enabled only after the last discharge. When the zero-crossing detector goes high, the micro is interrupted and the timer value is automatically stored in a register (TIN2). The time difference between the beginning of discharge and when interrupt occurs is the rise time of the current. In the interrupt service routine, the A/D converter is enabled, and the peak amplitude of the current is obtained from the peak detector. The rise time and amplitude are 75 CHAPTER 4 then displayed on the LCD panel, each staying on the display for 2 seconds before the message "READY" is displayed and the micro is again ready to accept new keypad entries. The key "MENU" sets up the MENU mode, which shows the operator the parameters that can be controlled and modified. The first parameter displayed is the mode of operation. The operator can select single pulse or multiple pulse (10 pulses within 50 seconds) mode by the key "INC" or "DEC". If the key "MENU" is pressed again, the total capacitance in the circuit is shown. The operator can modify it by the keys "INC" and "DEC". If the key "MENU" is pressed again the MENU mode will be ended and the message "READY" displayed again. Within the MENU mode, the key "RESET" resets the stimulator to the default mode, i.e., single pulse operation at a total capacitance of 850 fiF. The keys "DEC" and "RESET" will only be recognized inside the MENU mode. If the micro is not in the MENU mode when the key "INC" is pressed, the micro will enable the triggering circuits of the selected charge SCR's. The stimulator is operating in the charging mode. The message "CHARGING" will be displayed for 5 seconds before "READY" appears again. Within these five seconds, the micro will not accept any key entries. In the MENU mode, the key "INC" is used to modify the parameter displayed. The key "DATA" is used to retrieve the rise time and amplitude of the current resulting from the last discharge. Each parameter will stay on the LCD panel for 2 seconds. 76 CHAPTER 4 Figure 4.10 shows the flowchart of the main program of the software, and Figure 4.11 to Figure 4.14 show the flowcharts of the four major subroutines. INITIALIZE CPU REGISTERS 1 INITIALIZE OPERATING PARA. GOTO SUBROUTINE CHARGE Figure 4.10 F low Chart of the Software Mag_5 .asm. 77 CHAPTER 4 RUN S T A R T COIL T E M P . O K A Y ? D I S A B L E A L L C H A R G E S C R ' s E N A B L E D I S C H A R G E E N A B L E T IMER INPUT E N A B L E T IMER I N T E R R U P T M E A S U R E f„ I M E A S U R E /„ S T O R E f ^ A N D /„ IN R A M D I S P L A Y M E S S A G E "HOT COIL" G O T O S U B R O U T I N E SCAN U P D A T E P U L S E C O U N T E R P U L S N G O T O S U B R O U T I N E CHARGE G O T O S U B R O U T I N E DATA Figure 4.11 Flowchart of the subroutine RUN. 78 CHAPTER 4 START DISPLAY MODE PARAMETERS I DISPLAY CAP PARAMETER GOTO SUBROUTINE SCAN Figure 4.12 Flowchart of the subroutine M E N U . 79 CHAPTER 4 fOATA~~~\ n r OBTAIN f„ FROM RAM DISPLAY tp OBTAIN /„ FROM RAM DISPLAY /„ GOTO SUBROUTINE SCAN Figure 4.13 The flowchart for the subroutine DATA. 80 CHAPTER 4 fCHARGE\ \ S T A R T / OBTAIN C A P . P A R A M E T E R I A C T I V A T E C H A R G E S C R ' s DISPLAY M E S S A G E "CHARGING" DISPLAY M E S S A G E " P U L # " P U L S N Figure 4.14 The flowchart for the subroutine CHARGE. 4.7 PRACTICAL DESIGN CONSIDERATIONS While we wish to build a stimulator that satisfies all stated performance criteria we must accomplish this within time, funding and other constraints. Specifically, for the electrical design of the stimulator, we must consider the electrical ratings of the components available, and the 81 CHAPTER 4 overall electrical safety of the stimulator. For the mechanical design, we must consider the ruggedness of the chassis, the mechanical safety, and the operator interface. The electrical and mechanical designs of the stimulator are closely inter-related. An additional constraint is the long lead time in the delivery of high voltage SCR's and other components. 4.7.1 Mechanical and Electrical Safety There are mechanical as well as electrical safety factors that need to be considered during design. Although every effort has been made to ensure that the stimulator is designed to the highest standard, there is always the danger of component failure. Therefore, the stimulator must be enclosed properly to protect the user. The mechanical safety features are built into the design of the chassis. Since the stimulator chassis must contain over 100 lb of components (each 100 fiF capacitor weighs 7 lbs and each 120/600 V transformer weighs 61 lbs), the frame is made of 1" stainless steel angles welded together. Although we never had experience with high voltage capacitors exploding, we were warned to take precautions [34]. The four side panels and the top panel are 1/4" plexiglass, while the bottom panel of the chassis is stainless steel. For electrical safety, all components that must "float" electrically are isolated from ground by nylon standoffs and attached to the chassis so that they cannot inadvertently come in contact with the operator of the stimulator. For the same reason, the wire carrying the discharge current is enclosed in a heat shrinkable tubing and the stimulating coil is potted 82 CHAPTER 4 in epoxy 21AC-7C manufactured by Industrial Formulators of Canada. The frame and the bottom panel are all connected to the safety ground to protect the user from electrical shock. 4.7.2 Electromagnetic Interference Interference generated within the chassis is either from (1) noise generated during switching of the SCR's or (2) electromagnetic field generated during charge/discharge. When the SCR's switch on and off, the noise generated can be enough to cause misfiring of the other SCR's. Therefore, the wires that connect the triggering circuits to the SCR's must be short and twisted together to minimize coupling. In cases where short wires are not really possible, and interference is especially serious, shielded cable must be used. 83 C H A P T E R 5 E N G I N E E R I N G T E S T I N G A N D P E R F O R M A N C E E V A L U A T I O N 5.1 INTRODUCTION Before we test the performance of the stimulator in the clinical environment, we must perform engineering testing to ensure that all components work properly. Wherever it was possible, testing was performed on each module as a stand-alone unit. However, because most modules require input signals and/or loads to perform their actual functions, all modules were tested "in place" as integral parts of the complete stimulator. The quantities we are interested in are mostly currents, voltages, and time duration of events. In this chapter, we present the results of the measurements made in the various modules in the stimulator. Finally, we also compare the system performance with the target performance. 5.2 LIMITATIONS OF THE PRESENT STIMULATOR The stimulator design allows for five banks of capacitors in total. These five banks of capacitors can be selected in any combination to give coil currents of different rise time (t^) and amplitude (Iip). However, we only have three 2000 V SCR's at the time of prototying and testing because 84 CHAPTER 5 the manufacturer encountered production problems. We assembled the system using three available 2000 V SCR's as discharge SCR's, and use readily available 1200 V SCR's as charge SCR's. The system we tested therefore has three banks of capacitors, with values 96.8tiF (C2), 295.7 iiF ( Q , and 489.9 tiF (C3) (for simplicity we would take them as 100 tiF, 300 /xF, and 500 iiF). These capacitance values were chosen so that the total capacitance is large but the resolution can be as low as 100 ttF, and so we can control the duration of the coil current rise time to within a few microseconds. Because the charge SCR's are rated to only 1200 V, most of the testing has been done at a maximum of 1000 V. The results at 1600 V can be extrapolated from these results. 5.3 INSTRUMENTATION For voltage measurements we used a Tektronix 2232 storage scope, Beckman 310B digital multimeters, and voltage isolators made in the department of Electrical Engineering, UBC. The UBC voltage isolators have very good noise properties, but the slew rate of the output OP-AMP is limited to about 150 mV/fis. During discharge, we are looking at signals of maximum rate of change of 250 mV/iis, and therefore, the UBC isolators can only be used to measure slowly changing voltages, such as the charging voltage. To measure fast changing but low magnitude voltages (<400 V), the storage scope can be directly connected to some parts of the stimulator. At higher voltages there is no satisfactory way of performing this measurement. 85 CHAPTER 5 Because there is a lot of electromagnetic "interference" generated in the stimulator during discharge, the measurements must be carefully conducted to ensure that the interference coupled into the measuring circuit is minimal. For example, in measuring the gate current to the charge SCR's, the scope must be operated in differential mode, and the leads must be tightly twisted together to minimize pickup. We found that if this was not done carefully, the scope displayed large output spikes which could be reduced by rearranging the leads. Moreover, mistriggering of SCR's were also observed when the scope leads were haphazardly connected. The current was measured by a current monitor (Pearson model 101) which has an output of 0.01 WA. The conductor carrying the current to be measured is passed through the centre of the current monitor. Following the manufacturer's specifications, a small bias current of about 1.8 mA must be supplied to the current monitor to enable it to operate properly up to an input current of amplitude-duration (IT) product of 2.5 Amp-sec. If this is not done, the current monitor output will saturate at an input current of amplitude-duration product higher than 0.8 Amp-sec. Therefore, a -12 V supply is connected to the output of the current monitor through a 6k8 resistor. This gives a bias current of 1.76 Ma to the current monitor, keeping it from saturating. Figure 5.1 shows the output of the current monitor with and without the bias current. Measurements were performed on the system loaded by a test coil consisting of 8 turns of tightly wound #16 gauge solid wire. The outer diameter of the coil was 3.5 in. The inductance of this coil was calculated to be about 10 uH. 86 CHAPTER 5 All the results shown in the figures of this chapter are output plots from a Tektronix 2232 storage scope. We used the KERMIT communication software to transfer the plots from the scope into a 386 PC, and then used DrawPerfect to add comments and/or change the linewidth in the plots to improve readability. The symbols used in the results Figure 5.4 to Figure 5.21 are defined in Figure 5.2. J,(kA) 0 0.4 0.8 1.2 1.6 2.0 time (ms) Figure 5.1 The output of the current monitor with and without bias current. 87 CHAPTER 5 Figure 5.2 The voltage and current designations in the stimulator. he = the current supplied by the 115 V AC line supply. = the current flowing through the high voltage capacitors h the current flowing through the flywheel diodes h - the gate current of the SCR I, = the current supplied to the high voltage capacitors = the charging current I, = the coil current = output current vAC = the 115 V AC line supply voltage vak = the anode-cathode voltage of an SCR vc = the voltage across the high voltage capacitors = the coil voltage = output voltage v, the voltage at the output of the power supply the voltage across the filter capacitors 88 CHAPTER 5 5 A THE 1600 V POWER SUPPLY During charging, the capacitor charge/discharge module and the power supply cannot be tested separately. Figure 5.3 shows the sequence of the events. 120 V AC he-™1 AC 120 V AC POWER SUPPLY MODULE 120/1200 V TRANS-FORMER AND FULL-WAVE RECTIFIER V A _ L / =7 CAP MODULE (a) POWER SUPPLY MODULE 120/1200 V TRANS-FORMER AND FULL-WAVE RECTIFIER 'AC / = J + /' * / AC Rs CAP MODULE (b) Figure 5.3 The current provided by the filter and the AC line supply during charging. (a) When the charge SCR's are just beginning to conduct, the filter capacitors supply all the charging current. (b) After the initial current supply to the high voltage capacitors, the voltage of the filter capacitors drops and the AC line supply supplies current to both the filter capacitors and the high voltage capacitors. IAC = the current supplied by the 115 V AC line supply 7A C' = the current supplied by the full-wave rectifier Ij = the current supplied by the filter capacitors Is = the current supplied to the high voltage capacitors 89 CHAPTER 5 Assuming that the charging circuit is lossless, at the beginning of the charging cycle, the current drawn by all capacitors is Therefore, the charging current peaks at t = 0. For Rs = 50 R, Vs - 1000 V and Vc = -100 V, Is should peak at 22 A, but the measured value was 14 A (Figure 5.4). This indicates that in addition to the charging resistance of 50 fi, there is a distributed circuit resistance of about 28.6 fi, which includes the resistances of the wires, the filter capacitors, and the high voltage capacitors. As expected, the charging current decays exponentially from t = 0. Within the first 50 ms, the high voltage capacitors are charged up to 70% of the full charge. Full charge is complete within the next 5 s. (5.1) where and V. y, c the voltage across the filter capacitors, the voltage across the high voltage capacitors, the current-limiting resistance, the total capacitance selected in the capacitor charging/discharging module. 90 CHAPTER 5 Figure 5.4 The filter capacitor voltage, the discharge capacitor voltage and the charge current at the beginning of a charging cycle. I, = the current supplied to the high voltage .capacitors = the charging current V = the peak output voltage of the full-wave rectifier = the supply voltage Vt = the voltage across the high voltage capacitors V, = the voltage at the output of the power supply = the voltage across the filter capacitors Because the initial charging current Is drawn is so high, we were concerned that the AC line voltage (VAC) may drop. For this reason, we also performed measurements on VAC and the current drawn from the line supply (IAC). Figure 5.5 shows the VAC and IAC measured at the beginning of a charging cycle. When charging is started, the filter capacitors provides the large initial current surge and therefore its voltage (Vs) drops rapidly. The line supply starts delivering 91 CHAPTER 5 current to both the filter capacitors and the high voltage capacitors. The magnitude of IAC depends on the instantaneous line voltage, the filter capacitor voltage and the voltage across the high voltage capacitors. Because the line voltage is stepped up by a ratio of 10 in the stimulator, the current drawn from the supply (IAC) is 10 times the combined charging current to both Cs and C. Figure 5.5 shows that IAC reaches a maximum of 30 A for a supply voltage V and a capacitance C of 900 iiF. We found that the line voltage was not affected in any way. BEGINNING OF CHARGING V = 1000 V, C = 900 u.F time (ms) Figure 5.5 Initial current delivered by the full-wave rectifier at the beginning of the charge cycle (V = 1000 V, C = 900 /xF). C = the total capacitance in the charge/discharge module I, = the charging current V = the peak output voltage of the full-wave rectifier = the supply voltage Ve = the high-voltage capacitor voltage Vs — the filter capacitor voltage 92 CHAPTER 5 5.5 THE CAPACITOR CHARGE/DISCHARGE MODULE This module can be in one of three states: charging, discharging, and standby. The testing was performed on the module when it was in either charging or discharging modes. Our testing aimed to confirm that the capacitors and the SCR's are not overloaded in any way and that the performance criteria are met. There are several parameters that we need to test to ensure that the SCR's are operating within their rated values. These parameters are a. the maximum current each SCR conducts, b. the maximum rate of increase of the current flowing through the SCR (maximum dl/dt) c. the maximum voltage across the SCR (maximum V^, and c the maximum rate of voltage change across an SCR ( maximum dVJdt). 5.5.1 Capacitor Discharge Since capacitor charging has already been discussed in Section 5.4, we will only discuss capacitor discharge in this section. At the beginning of the discharge (t0 < t < tt in Figure 5.6) the coil current follows Equation (4.1) in Chapter 4 until the capacitor voltage reaches zero (t = *,). When the capacitor voltage starts reversing, the flywheel diodes become forward biased and start conducting. Since the phase control SCR's used in this application have a maximum turn-off time of 150 us, the commutation (switching the current from one device to another) does not occur instantaneously. During commutation (t1> t > 93 CHAPTER 5 t2), current is shared among the flywheel diodes and the discharge SCR's. Meanwhile, Vc keeps decreasing until the SCR's are completely turned off (t = t2). Then the current keeps flowing in the diode-coil loop until it decays to zero (t > t2). time (LIS) Figure 5.6 The coil current, the diode current, the capacitor current and the capacitor voltage during discharge. /„ = the current flowing through the high voltage capacitors = the current flowing through the flywheel diodes 7( = the coil current = output current V = the peak output voltage of the full-wave rectifier = the supply voltage Vc = the voltage across the high voltage capacitors If this were a totally lossless circuit, all the energy stored in the capacitors would be transferred to the coil (an inductor) when the coil current peaks, ie, the capacitor voltage 94 CHAPTER 5 would be zero when the coil current is at its maximum. In the measurements the maximum of coil current I, does not coincide with the zero of capacitor voltage Ve, indicating that the resistance in the discharge path is not negligible. We can calculate the resistance with the simultaneous equations set up in Chapter 4. Equations (4.4) and (4.5) are simultaneous equations that can be solved for the circuit resistance R and inductance L since we know what the supply voltage and capacitance are. We will discuss this in detail in Section 5.9. 5.5.2 The dl/dt and 7^  Measurements As we have discussed in Chapter 4, the maximum heating that an SCR junction can withstand is represented by the maximum rate of change of current (dl/dt) through the SCR. We can calculate the maximum value of dl/dt for the coil current It from Equation (4.7): dt V = l l (5.2) Table 5.1 shows the measured Itp and at V = 1000 V and the calculated 1^ and dl/dt at V = 1600 V. 95 CHAPTER 5 Table 5.1 The maximum calculated I, and dl/dt through the discharge SCR's and the rated values. C(uP) (us) V = 1000 V Cal. I¥ at V = 1600V (kA) rated ITSM (kA) Cal. dl/dt at V = 1600 V (A/us) rated dl/dt of the SCR (Alus) 100 64 2.4 3.8 9.5 123.1 800 300 92 3.9 6.2 9.5 123.1 800 500 118 5.0 8.0 15.5 123.1 800 We can see from Table 1.1 that according to the calculations, both 1^ and dl/dt are within the rated values. 5.5.3 The dVnk/dt and Vah Measurements The two most important voltage related ratings of an SCR are the maximum repetitive off-state voltage (absolute maximum anode-cathode voltage for each SCR = 2000 V) and the critical rate-of-rise of voltage (critical dVJdt for each SCR = 700 V//cs). An SCR will be damaged if exceeds the rated maximum. If dVJdt across an SCR exceeds the rated value, the SCR may switch on without gate triggering. This will also damage the SCR. Thus we must verify that the SCR's are operating within the voltage ratings. We have selected to make all the measurements on the charge and discharge SCR's connected to C2. Both Cj and C} were selected throughout the measurements, and C2 was the only bank 96 CHAPTER 5 selected and unselected. The measurements then allow us to calculate the dVJdt parameter. Figure 5.7 shows the equivalent circuits of the SCR's during charging and discharging. Table 5.2 shows the relationship between and the other voltages. 97 CHAPTER 5 Ci \ Vg T» T-c 2 \ \ \ \ \ (a) ••s - v w + \ \ \ \ Ic-Id=It-0 (b) (V) \ \ \ \ (c) \ \ (d) Figure 5.7 Equivalent circuit of the capacitor charge/discharge module for the three operating modes: standby, charging, and discharging. (a) standby mode (b) charging mode (c) discharge mode (Vc > 0) (d) discharge mode (Vc < 0) /„ = the current flowing through the high voltage capacitors Id = the current flowing through the flywheel diodes Ia = the charging current /, = the coil current = output current = the anode-cathode voltage of an SCR Vc = the voltage across the high voltage capacitors Vs = the voltage at the output of the power supply = the voltage across the filter capacitors 98 CHAPTER 5 Table 5.2 The relationship between and the other voltages in the circuit. status of circuit of non-selected charge SCR of selected charge SCR of non-selected discharge SCR K*of selected discharge SCR standby v. -v. v.-v. vc vc charging vs-vc Vf vc Vc discharging vs-vc vs-vc vs-v0 Vf where VaJ t= the anode-cathode voltage of an SCR, V,. = voltage across the high voltage capacitors, Vf = forward voltage drop across an SCR when it is on (about 1 V), Vg = voltage across the stimulating coil, and Vs = voltage across the filter capacitors. A conducting SCR is equivalent to a short circuit and a non-conducting SCR is equivalent to an open circuit. Figure 5.8 to Figure 5.11 verify that the V^ of SCR's as described in Table 5.2. They show that the most rapid rise of voltage occurs at the beginning of each charging cycle, during which the circuit is a DC source charging a capacitor through a resistor. v . - y . h - " " * ) ( 5 - 3 ) where Rs = current limiting resistor at the output of the filter. Therefore, dt V (5.4) 99 CHAPTER 5 which is the largest when Vs is at its maximum (1600 V). For this circuit, dV^dt is never larger than lV/iis. V - 120 V OF A NON-SELECTED CHARGE SCR " DURING CHARGING ' Vat o p A SELECTED CHARGE SCR - DURING CHARGING V- I0OV V I I I I I I .v OF A NON-SELECTED CHARGE SCR DI <* JRTNG DISCfl [ARGE ^ 'a* OF >URIN< x — A SEL 3 DISC JBCTBI HARD ) CHA] i tOE S< z VERT. SCALE: 50 V/div HORZ. SCALE: 50 ms/div VERT. SCALE: 50 V/div HORZ. SCALE: 50 y«Miv Figure 5.8 The anode-cathode voltage C S C R 2 during charging. of Figure 5.9 The anode-cathode voltage C S C R 2 during discharge. V..1, of v- 100 v I I I I I I I OF A NON-SELECTED DISCHARGE SCR _ DURING CHARGING OF A SELECTED DISCHARGE SCR . DURING CHARGING V BEGINNING OF DISCHARGE VERT. SCALE: 50 V/div HORZ. SCALE: 50 ms/div OF A SELECTED DISCHARGE SCR DURING DISCHARGE -I 1 1 V ^ OF A NON-SELECTED DISCHARGE SCB DURING DISCHARGE VERT. SCALE: 50 V/div HORZ. SCALE: 50 iWdiv Figure 5.10 The anode-cathode voltage D S C R 2 during charging. of Figure 5.11 The anode-cathode voltage of D S C R 2 during discharge. 100 CHAPTER 5 5.6 THE SCR'S TRIGGERING CIRCUITS There are two major factors that can cause mistriggering in SCR's. Firstly, The noise generated when SCR's switch can couple into the triggering circuits of the other SCR's. Secondly, noise at the anode of an SCR may be coupled by the capacitance in the SCR to the gate. The physical layout of the circuits is such that the charge SCR's are located quite far from the triggering circuits and therefore long wires must be used in the connection. This increases the potential for electromagnetic interference "pick-up". To prevent the SCR's from mistriggering, shielded wires are used in the gating circuit to the SCR connections. In addition, a small capacitor is placed across the output of each gating circuit. To prevent the interference from directly affecting the circuits driving the triggering circuits, filter capacitors are distributed on the transformer board. The measured gate-cathode voltages Vgk for an SCR during charge and discharge are presented in Figure 5.12 to Figure 5.13. Since SCR's are current-controlled devices, it is the induced gate current Ig that would cause mis-triggering. However, mis-triggering only occurs when: (1) the SCR is forward biased, (2) the combination of induced gate-cathode voltage and gate current provide enough gate bias and energy. During charging, the induced gate-cathode voltage Vgk is less than 0.5 V (Figure 5.12 and Figure 5.14), although the induced gate current Ig can be as high as 400 mA (Figure 5.16 and Figure 5.17). According to the manufacturer's specifications, Vgk should be between 1.2 to 4 V for triggering when Ig is 800 mA, and therefore it is unlikely that mistriggering will occur to the SCR's. During discharge, the induced gate-cathode voltage can be as high as 10 V and 101 CHAPTER 5 induced gate current 700 mA for the unselected discharge SCR's, but they are slightly reverse biased. For the charge SCR's, the induced gate-cathode voltage and the gate current are not enough to cause triggering. 102 CHAPTER 5 V - 1000 V t OP A U N G ( SHLEt 2HARC 7TED < ZHARC ESCR D U ING \ ^ V ft O F JRING A N O N CHAR -SELEI GING ?THD ( ;HARC B SCR DI —t— VERT. SCALE: 1 V/div HORZ. SCALE: 10 (is/div Figure 5.12 The gate-cathode voltage of charge SCR's during charge. V- 1000 V - V - T O F A SELECTED C H A R G E SCR DURING DISCHARGE O F A NON-SELECTED C H A R G E SCR ' DURING DISCHARGE. VERT. SCALE: 2 V/div HORZ. SCALE: 02 ms/div Figure 5.13 The gate-cathode voltage Vgk of charge SCR's during discharge. V - 1000 V 1 1 I 1 I I V R T O F A SELECTED DISCHARGE SCR f DDI 0NGCHARG 1 ING •1— 1 1 1 O F A NON-S ELEC1 ING "ED DI SCRAP G E S C R DUI UNO C HARG ,1 — L -1 VERT. SCALE: 0.5 V/div HORZ SCALE: 02 ms/div Figure 5.14 The gate-cathode voltage Vgk of discharge SCR's during charge. SCR SWITCHES ON SCR SWITCHES OFF V - 1000V - Vgk O F A SELECTED DISCHARGE SCR" DDRINO DISCHARGE VgIC O F A NON-SELECTED DISCHARGB SCR DURING DISCHARGE. VERT. SCALE: 10 V/div HORZ. SCALE: 0.2 ms/div Figure 5.15 The gate-cathode voltage Vgk of discharge SCR's during discharge. 103 CHAPTER 5 V- 1000 V 1 1 1 1 1 / . OF A SELECTED CHARGE S CR DL [RING CHAR< 3ING 1 y r~ i / . O F A N O l 4-SELE CHAD 1CTED .GING CHAR r,F, K SCRD URTNG -4. y ~r— V - 1000 V I 1 1 1 1— It OF A SELECTED CHARGE SCR DURING DISCHARGE . SCR DURING DISCHARGE VERT. SCALE: 1 A/div HORZ. SCALE: 10 |«/div VERT. SCALE: 0.2 A/div HORZ. SCALE: 0.1 ms/div Figure 5.16 The gate current Ig of a charge SCR during charging. Figure 5.17 The gate current Ig of a charge SCR during discharge. SCR SWITCHES ON SCR SWITCHES OFF V - 1000 V 1, OF A SELECTED DISCHA RUE y ISCHA RGB -lt Oh A IMON-SbLBCTKD D ' SCR DURING CHARGING J .1 y t 1 1 V - 1000 V OF A SELECTED DISCHARGE -SCR DURING DISCHARGE / SCR DURING DISCHARGE VERT. SCALE: 40 mA/div HORZ, SCALE: 0.1 msj&v VERT. SCALE: 1 A/div HORZ. SCALE: 0.1 ms/div Figure 5.18 The gate current Ig of a discharge Figure 5.19 SCR during charging. The gate current Ig of a discharge SCR during discharge. 104 CHAPTER 5 5.7 THE MICROCONTROLLER AND THE LOGIC CIRCUITS Besides the switching noise generated by the SCR's, noise generated at junctions of connectors, and radiation from the wires in the discharge path, can all affect the microcontroller and the logic circuits. In both 5 V and 12 V lines, large spikes occur whenever the discharge SCR's switch on. Although the spikes may be as large as 0.5 V, the logic circuits performed their functions with no errors. The -12 V line was not affected much because it is not supplying the SCR gating circuits, and is therefore less spread out then other power supply wires. Every time an SCR is triggered, its gating circuit draws about 1 A from the 12 V line, and if all capacitor banks are selected, a total of 5 A must be supplied by the 12 V line whenever the gating circuits are active. This forces the DC power supply into current limiting mode which lasts for about 500 /us during discharge. Although 100 ttF filter capacitors are placed all over the transformer board, they do not eliminate this problem. If three banks of capacitors are selected (Figure 5.20), the 12 V line will be decreased by 0.4 V at the end of the discharge enable pulse before it starts to rise to the nominal value. However, since the 12 V line supplies current to only the SCR triggering circuits and the sensors, this voltage drop will not affect the overall operation of the stimulator. Figure 5.21 shows that with only one bank selected, current limiting is insignificant. 105 CHAPTER 5 BEGINNING OF DISCHARGE ENABLE PULSE END OF DISCHARGE ENABLE PULSE 1 ^ 12 V / 5 V ^ -1 2 V V - 1000 V, C - 900 uF VERT. SCALE: 1 V/div (diff. mode, AC coupling) HORZ. SCALE: 1 ms/div Figure 5.20 The DC supply lines during discharge at Vc = 900 V, three banks of capacitors selected. BEGINNING OF DISCHARGE ENABLE PULSE 1 , 12 V 5 V •~1— I V - ' ' V - 1000 V, C - 100 (iF VERT. SCALE: 1 V/div (diff. mode, AC coupling) HORZ. SCALE: 1 ms/div Figure 5.21 The DC supply lines during discharge at Vt = 900 V, one bank of capacitors selected. 106 CHAPTER 5 5.8 SENSORS There are three sensors in the stimulator that need to be calibrated and tested for performance: a. the voltage sensor which senses the filter capacitor voltage (Vc) and changes state when this voltage is larger than 1800 V; b. the temperature sensor which senses the coil temperature (Tc) and changes state if the temperature is higher than 45 °C; and c. the current sensor which senses the coil current rise time (t^) and amplitude (7^ ,). Because some of the sensors have been designed to sense conditions that are outside of the normal operating conditions of the stimulator, such conditions cannot always be provided for the calibration of the sensors. In particular, to calibrate the temperature and voltage sensors we had to use simulated conditions. The voltage sensor used in this design incorporates an opto-isolator (Figure 4.7). Its function is to set an interrupt signal to the microcontroller if the input voltage is larger than 1600 V. Since we could not provide a supply that is larger than 1600 V, we could not calibrate the voltage sensor directly. Therefore, we reduced the input current control resistor from 200 k to 2 k and calibrated the sensor with a variable 25 V power supply. The sensor output changes to low at an input voltage of 17.59 V. The voltage sensor switches back to high when the input voltage drops to below 16 V. We can conclude that in its final form, the voltage sensor switches to low at about 1760 V and back to high at about 1600 V. 107 CHAPTER 5 The temperature sensor is a temperature sensitive diode connected to a variable resistor. The flat surface of the diode was placed in contact with the coil, and a small amount of heat conductive compound (Wakefield 120-2) was placed on the contact surface for better heat conduction. The tip of a temperature probe (Fluke 80T-150U), also in contact with the coil, was placed next to the diode. The coil was heated up with the aid of a heat gun, but care was taken not to blow air directly on the diode or temperature probe. The heat was also delivered in bursts every five seconds, simulating the repetitive discharge operating mode. The temperature sensor was calibrated to switch to high at 46°C and to switch back to low at 43.6°C. During calibration, it takes about 30 seconds for an uninsulated coil to cool from 46°C to 43.6°C. When we calibrated the current sensor, we assumed that the current monitor (Pearson 101) was accurate to within 1 % of the true value. The output of the current monitor was displayed on the scope and the coil current amplitude 7^  and rise time read off manually. These values were then compared to those values displayed on the LCD panel. In the stimulator, 7^  is measured by first scaling down the current monitor output, and then feeding the reduced signal to the A/D converter in the microcontroller. The measured coil current amplitude 7^  can be easily calibrated to an accuracy of 0.2 kA for any amplitude (Figure 5.22). 108 CHAPTER 5 As discussed in Chapter 2, the differentiator introduces a small phase error in the output, and therefore the measured coil current rise time is usually longer than the true value. The values of displayed were found to be within 10 tis of those measured on the scope (Figure 5.23). 109 CHAPTER 5 60 80 100 120 140 160 180 true ttp (us) Figure 5.23 Coil current rise time tlp measured by the stimulator vs. true tt 110 CHAPTER 5 5.9 PERFORMANCE TEST The major feature of this stimulator that distinguishes it from the other stimulators is that: it can produce coil current pulse of varying amplitude (1^) and rise time (t^) by varying the supply voltage and the capacitance in the circuit. Figure 5.24 and Figure 5.25 show the coil current shapes when one of the parameters (V or C, respectively) is kept constant while the other is varied. Figure 5.24 shows that for different Vs, 1^ increases accordingly but is not affected in any way. Therefore, as expected, stays constant regardless of the value of V if C is held constant. Figure 5.25 shows that as C becomes larger, 1^ and tv also becomes larger. Thus, we have confirmed experimentally that we can vary both the coil current amplitude 1^ and rise time just by changing the circuit capacitance C and/or the supply voltage V. 7,(kA) 0 0.2 0.4 0.6 0.8 1.0 time (ms) Figure 5.24 Coil current at C = 900 jiF at V = 250 V, 500 V, 750 V and 1000 V. I l l CHAPTER 5 /,(kA) time (ms) Figure 5.25 Coil current at C = 100 /iF, 300 /xF, 500 jtF and 900 fiF at V = 1000 V. If we measured the coil current amplitude 1^ and rise time t^, we can calculate the equivalent inductance and resistance in the circuit since we know the values of C and V since \ L exp :arcsin\/1 - y 2 (5.5) t.„ = J LC arcsini/1 - y 2 (5.6) 112 CHAPTER 5 From these equations we calculated the average resistance (JR) to be 40 mO and inductance (L) to be 14.7 juH in the circuit. Table 5.3 shows the 1^ and calculated by substituting these values into Equations (5.5) and (5.6) as well as the values measured off the displays on the scope. Table 5.3 Theoretical coil current amplitude Ilp and rise time t,p assuming R = 40 mf l and L = 14.7 /tH vs. the measured values. C(uF) V(V) cal. 7„ (kA) meas. Itp (kA) cal. ^ (/is) meas. (us) 96.8 1000 2.4 2.3 57 55 295.7 1000 3.9 3.8 98 100 489.9 1000 4.9 4.8 124 124 882.4 1000 6.2 6.2 163 162 882.4 250 1.6 1.6 163 162 882.4 500 3.1 3.1 163 162 882.4 750 4.7 4.7 163 162 882.4 1000 6.2 6.2 163 162 We can see that the theory can predict the coil current amplitude to within 0.1 kA and rise time to within 2 /ts. We therefore conclude that there is no change in the physical properties of the components in the presence of large currents, ie, no large signal phenomena. 113 CHAPTER 5 The performance of this prototype is therefore: a. maximum voltage = 1000 V b. maximum current = 6 kA (with the test coil) c. tfp is selected in a maximum resolution of 30 us d. two modes of operation: single shots and repetitive mode (1 pulse per 5 seconds) e. coil over-temperature protection f. over-voltage protection g. coil current amplitude and rise time measurements to 0.2 kA and 10 us accuracy. Comparing the performance of this unit to the target specifications in Section 4.5, some differences are noted. In particular, it is impossible to cycle charge/discharge at 1 Hz as the target specification calls for because of the limitation of the charging process. When we specified the rated values of the SCR's in the circuit, it was assumed that the circuit resistance R is 50 mfi, and the minimum load inductance (which results in the highest 7^ ) was assumed to be 5 uH. Since R was found to be only 40 ml), the minimum load inductance for the circuit must be larger than 5 uH if we want to ensure that the SCR's are operating within their rated values. This minimum load inductance was calculated to be 8 uH. The amplitude of the current delivered at this inductance with 900 uF and 1000 V is 8 kA. 114 CHAPTER 6 PROPOSAL FOR CLINICAL TESTING 6.1 INTRODUCTION In clinical testing of the magnetic stimulator, our goal is to establish a strength-duration curve for the median nerve, and to compare the results with that established by the standard electrical stimulation technique. We must be able to magnetically stimulate the median nerve without affecting a neighbouring nerve. The design of the magnetic coil is vital in achieving this. We propose to experiment with two coil shapes: the tear-drop shaped coil and the 8-shaped coil because they have been proven to produce more localized stimulation. In addition to the shape, the outer diameter of the coil and the number of turns will also determine whether the coil produces focused stimulation. However, we must also keep in mind that the maximum current rating of the stimulator is only 8 kA for this prototype. If the coil inductance is too low, the peak output current may exceed this value and the stimulator may be damaged. 115 CHAPTER 6 6.2 COIL SELECTION For the coil shapes we have selected the tear-drop shaped coil and the 8-shaped coil because they have been proven to perform better than plain circular coil. The calculation of the inductance of a tear-drop shaped coil would be quite involved. Since its shape is not very different from that of a circular coil, we will estimate its inductance by calculating the inductance of a circular coil of the same coil area. The inductance of a circular coil can be estimated by the following: In c - 1.75 r \ w ) where rc = radius of coil rw - - radius of wire N = number of turns in the coil fi0 = permeability of non-magnetic material An 8-shaped coil is two round coils connected in series. We will ignore the mutual inductance between the two wings and assume that the total inductance is twice of that of each wing. In published literature, a coil inductance of about 8 /iH is widely used and we will use this inductance as a starting value. The wire used to form the coil is a magnet wire of radius 1/16". Knowing the inductance (L) and the radius of the wire (rw), the only two parameters that need to be determined are the radius of the coil and the number of turns in the coil. 116 CHAPTER 6 6.3 Clinical Testing of Magnetic Stimulation For the first part of the testing, we will compare magnetic stimulation with electrical stimulation in the stimulation of peripheral nerves. The peripheral nerve selected is the median nerve (in the arm) which controls the Thenar Eminence muscle (muscle at the base of the thumb). The coil will be placed over the nerve in the position that is theoretically optimum (Figure 6.1) and the median nerve will be stimulated and the action potential of the Thenar Eminence will be recorded. The muscle innervated by the ulnar nerve will also be monitored to ensure that the ulnar nerve is not stimulated as well. The stimulator will be set to a supply voltage of 400 V, a capacitance of 800 uF, and single pulse operation as its default operating conditions. THEORETICAL OPTIMUM POSITION OF THE NERVE FOR STIMULATION Figure 6.1 The optimum positioning of the coil over the median nerve. The site of depolarization is located at the point of maximum change of induced electric field along the nerve. 117 CHAPTER 6 The median nerve will be stimulated and the action potentials of the two muscles will be observed. If the ulnar nerve is also stimulated, it can be that the stimulus is too large or the coil shape and/or position need improving. We will investigate then by decreasing the stimulus or moving the coil. If neither of these helps to improve the focality of stimulation then the coil must be redesigned. If the median nerve can be stimulated without affecting the ulnar nerve, we will perform measurements to experimentally determine the optimum coil position (reliable stimulation with the minimum stimulus energy). The stimulator will be set to its default operating conditions. The median nerve will be repeatedly stimulated with increasing stimulus until the amplitude of the action potential recorded does not increase with increased stimulus. The median nerve is then maximally stimulated and the action potential recorded from the Thenar Eminence is called an M-wave. We will measure the amplitude, the latency (time difference between stimulus and the onset of M-wave) and the propagation velocity of the action potential. After this reference coil current pulse has been established, the coil will then be moved around to different positions over the nerve (Figure 6.2) and then to different orientations with respect to the nerve (Figure 6.3) and the response recorded. For each position and orientation, we will adjust the stimulator output until an M-wave is recorded or until the stimulator is operating to its limits. From these measurements we will find the optimum position for stimulation without affecting a neighbouring ulnar nerve. After we have located the optimum location and 118 CHAPTER 6 orientation, we will vary the output of the stimulator to obtain the S-D curve of the median nerve. coo. POSITION OF NERVE Figure 6.2 The different positions of the tear-drop shaped coil over the median nerve in magnetic stimulation. Testing is performed at the angle 9 = 0° to 90° at 15° intervals. Figure 6.3 The different orientation of the coil in magnetic stimulation. Testing is performed at a and B = 0° to 90° at 15° intervals. 119 CHAPTER 6 After we have collected the results from magnetic stimulation, we will use electrical stimulation on the same nerve. For both types of stimulation, we will compare latency, the amplitude of the M-wave, and the propagation velocity, and therefore compare their performance. 6.4 Techniques to Modify and Enhance Magnetic Stimulation In the research of magnetic stimulation of peripheral nerves and the motor cortex, most researchers simply rotate or move the coil to get the best response possible, and there has been very little effort in developing "accessories" for the technique. Like conducting gel for electrodes in electrical stimulation, there must be devices of similar functions in magnetic stimulation. These devices can help in focusing the electric field or shielding it from certain parts of the body. 6.4.1 Magnetic Shield One device that could be tested is a shield that would shield the electric field from stimulating parts adjacent to the stimulated area. Although this device may not be very useful in stimulation of the peripheral nerves, it can be quite useful for stimulation of the motor cortex. Suppose it is desired to stimulate a deep structure and the coil used is very small. The field induced by the other parts of the coil (not directly over the structure) may stimulate a more superficial structure. If this is not tolerable, one can consider using a diamagnetic shield covering all the area of neighbouring parts and only exposing the part to be stimulated 120 CHAPTER 6 (Figure 6.4). Since the magnetic field will be reflected off this shield, the subject's exposure to the electromagnetic field will be minimized. A tank of 1 % saline solution can be used to simulate the biological tissues for preliminary studies. An electric field sensor placed inside the tank can be used to map the field induced by the coil current. The measurements will be performed with and without a diamagnetic shield around the tank. The two field mappings will give an indication whether the diamagnetic shield minimizes field exposure or focus the field. coil Figure 6.4 A shield for use in magnetic stimulation of the arm. The magnetic field is reflected away from the diamagnetic shield. Only the part of the arm exposed has current induced in it. It = the coil current It = the current induced in the tissues. 121 CHAPTER 6 6.4.2 Combination of Electrical and Magnetic Stimulation According to Bickford et al [35], when they combined electrical and magnetic stimulation on an ulnar nerve (Figure 6.5), they obtained a response about 2 to 4 times the amplitude of either stimulation alone. To verify if this is due to simple summation, they stimulated the same nerve simultaneously with two electrical stimuli but the response enhancement obtained was less than when the two stimulation techniques were combined. Bickford et al suspected that it was due to the Hall effect, but since Hall effect is a semi-conductor phenomenon, we suspect that it is more likely an electrochemical reaction. Figure 6.5 Magnetic stimulation combined with electrical stimulation. ItUc = the current provided by the electric stimulator, I; = the current induced by the magnetic stimulator, /, = the coil current. 122 CHAPTER 6 We propose to repeat the experiment on the median nerve. The output of the electrical stimulator and magnetic stimulator will each be adjusted to half of the output with which the median nerve is stimulated supramaximally. The outputs of the two stimulators will be synchronized and the median nerve will be stimulated. The stimulators will be adjusted until the median nerve is supramaximally stimulated. We will then measure the amplitude, latency and propagation velocity of the M-wave as before. This experiment will show whether combined electrical and magnetic stimulation really gives more focused stimulation and whether a larger action potential is obtained. 123 C H A P T E R 7 C O N C L U S I O N A N D R E C O M M E N D A T I O N S 7.1 CONCLUSION We have built a versatile magnetic stimulator that is capable of delivering coil current pulses of variable rise time and amplitude. The operating voltage of this prototype is 0 to 1000 V. The range of the rise time and amplitude depends on the coil used. With a test coil of inductance 14 jitH the coil current rise time ranges from 50 us to 160 ps and the current amplitude ranges from 0 to 6 kA. The stimulator can discharge in single pulses or 1 pulse per 5 seconds for 10 times. There is over-voltage protection and coil over-temperature protection. The user interface allows the user to change the capacitance and the mode of operation via a keypad. 7.2 RECOMMENDATIONS FOR FURTHER WORK ON THE STIMULATOR We have achieved our goal of building a stimulator capable of delivering current pulses of variable rise time and amplitude. However, because of the time and funding constraint, we have 124 CHAPTER 7 used the simplest design that works. To improve on the performance and size of the stimulator, the existing design can be modified in several areas. These include a new power supply, and a low cost temperature sensor. There are also additional features that can be added to improve the performance of the stimulator. The power supply in the existing design is a variable transformer whose output is stepped up and then rectified. This can be replaced by a AC to DC converter with a constant current output, whose amplitude is controlled by the output of a timer in the microcontroller. The high-voltage capacitors can then be charged with constant current, and the final capacitor voltage will depend on the amplitude and duration of this current. Therefore, all the operating parameters of the stimulator, including the operating voltage (capacitor voltage), the mode of operation and circuit capacitance, are controlled by the microcontroller. The user can have the convenience of specifying all these parameters via the keypad. If large scale testing is to be done, a PC can be used to control the microcontroller as well as perform functions such as data acquisition. Other advantages of this power supply include: less stress on the transformers and the charge SCR's because there is no initial current surge during charging; and a smaller, lighter, and less expensive unit. With this new power supply, the voltage across each bank of capacitor will have to be monitored to ensure it never exceeds 1600 V. The current monitor used in the stimulator has a dynamic range of 20,000 A with an accuracy of ± 1 %. This is a high quality current monitor originally intended for instrumentation use. It can be replaced by a low cost Hall-effect sensor. However, it is probably difficult to build a 125 CHAPTER 7 single Hall-effect sensor with a dynamic range of 20,000 A. There can be two sensors, however: one sensing from 0 to 2000 A and one from 2000 A to 20,000 A. Because of time and cost limitations, many features that would enhance the performance of the stimulator were not implemented. These features include: a. short circuit detection within the stimulator, especially with the SCR's; b. automatic capacitor discharge when the stimulator is switched off; c. voltage sensing across each capacitor bank for display on the LCD panel; and d. open circuit detection to stop all operation if the coil is disconnected; e. load sensing to sense the inductance value of the coil; f. user specifying the coil rise time and current and the software automatically calculates the capacitance and the supply voltage needed; g. doubly insulating the chassis for added mechanical and electrical safety. 7.3 RECOMMENDATIONS FOR FURTHER WORK ON MAGNETIC STIMULATION Magnetic stimulation is a new technique largely unexplored. Our stimulator provides a versatile tool for research of the technique. The neurophysiology of magnetic stimulation and its biological effect are still largely unknown. It has been assumed so far that stimulation is caused by the induced current in the nerve tissues, 126 CHAPTER 7 but the actual mechanism has not been proven experimentally. Detailed understanding of the mechanism will provide insight into the biological effect of magnetic stimulation. Although much research has been done on the effect of low frequency and microwave frequency magnetic field on animals and models of humans, there has been little work on the effect of pulsed magnetic fields. Systematic study of such pulsed field is important in establishing standards for the safe use of magnetic stimulation. The overall technique of magnetic stimulation needs refinement before its potential in general clinical use can be realized. In electrical stimulation of nerves clinicians have good control over the technique: they know how and where to stimulate a nerve and can accurate predict the site of depolarization. Magnetic stimulation should be just as easy to use. The stimulation coil shape must be designed to focus the stimulation such that the site of depolarization can be accurately predicted. The optimum current amplitude and rise time needed to stimulate different nerves must also be known. As we have discussed in Chapter 6, there may be devices and/or processes that can enhance the effectiveness of magnetic stimulation. They can serve to focus the field onto the area to be stimulated or shield the rest of the body from the magnetic field. 127 REFERENCES [I] R.F. Schmidt, Fundamentals of Neurophysiology, Springer-Verlag, 1978. [2] V. A. Marsocci, "Electrical network modelling of active membranes of nerves," CRC Critical Rev. in Biomed. Eng., vol. 8, no. 2, pp. 135-194, 1982. [3] J. Kimura, Electrodiagnosis in Disease of Nerve and Muscle: F.A. Davis Co., 1983. [4] M.M. Patterson, R.P.Kesner, Electrical Stimulation Research Techniques, Academic Press, pp. 2-36, 1981. [5] D. Noble, "Conductance mechanisms in excitable cells," Biomembranes, vol. 3, F. Kreuzer and J. F. G. Siegers, Eds., Plenum Press, New York, 1972. [6] J.P. Reilly, "Peripheral nerve stimulation by induced electric currents: exposure to time-varying magnetic fields," Med. & Biol. Eng. & Comp., pp. 101-110, March 1989 [7] R.F. Harrington, Time Harmonic Electromagnetic Fields, McGraw-Hill Book Company, pp. 93, 1961. [8] L .G. Cohen, B J . Roth, J. Nilsson, N. Dang, M. Panizza, S. Bandinelli, W. Friauf, and M . Hallett, "Effect on coil design on delivery of focal magnetic simulation. I. Technical Considerations," Electroencephalogr. clin. Neurophysiol., vol. 75, pp. 350-357, 1990. [9] L.A.Geddes, "Stimulation of excitable tissue with time varying magnetic fields," presented at the IEEE Eng. in Med. & Biol. Soc. 10th Ann. Int. Conf., pp. 918-921, 1988. [10] C. Reuter, J.H.Battocletti, J.Myklebust, and D. Maiman, "Magnetic stimulation of peripheral nerves," presented at the IEEE Eng. in Med. & Biol Soc. 10th Ann. Int. Conf., pp. 928-929, 1988. [II] A.T. Barker, I.L.Freeston, RJalinous, and J.AJarratt, "Magnetic stimulation of the human brain and peripheral nervous system: An introduction and the results of an initial clinical evaluation," Neurosurgery, vol. 20, pp. 100-109, 1987. [12] J.AJarratt, "Magnetic stimulation for motor conduction," AAEE Symposium pp. 27-33, 1986. 128 [13] P.J.Maccabee, V.E.Amassian, R.Q.Cracco and J.A.Cadwell, "An analysis of peripheral motor nerve stimulation in humans using the magnetic coil,", Electroencephalogr. & Clin. Neurophysiol., vol. 70, pp. 524-533, 1988. [14] B.A. Evans, W.J. Litchy, and J.R. Daube, "The utility of magnetic stimulation for routine peripheral nerve conduction studies," Muscle & Nerve, vol. 11, pp. 1074-1078, 1988. [15] C. Reuter, J.H. Battocletti, J. Myklebust, and D. Maiman, "Magnetic stimulation of peripheral nerves," presented at IEEE Eng. Med. Biol. Soc. 10th Ann. Int. Conf., pp. 928-929, 1988. [16] D.Cohen and B.N.Cuffin, "Developing a more focal magnetic stimulator. Part I: Some basic principles," to be published. [17] S.Chokroverty (Ed), Magnetic Stimulation in Clinical Neurophysiology, Butterworths, Stoneham, 1990. [18] J.A.Cadwell, private communication. [19] D.Cohen, private communication. [20] P.J.Maccabee, V.E.Amassian, R.Q.Cracco, J.B.Cracco, andB.J.Anziska, "Intracranial stimulation of facial nerve in humans with the magnetic coil," Electroencephalogr. & clin. Neurophysiol., vol. 70, pp. 350-354, 1988. [21] V.E.Amassian, R.Q.Cracco and P.J.Maccabee, "Focal stimulation of human cerebral cortex with the magnetic coil: a comparison with electrical stimulation," Electroencephalogr. & clin. Neurophysiol., vol. 74, pp. 401-416, 1989. [22] R.Q.Cracco, V.E.Amassian, P.J.Maccabee and J.B.Cracco, "Comparison of human transcallosal responses evoked by magnetic coil and electrical stimulation," Electroencephalogr. & clin. Neurophysiol., vol. 74, pp. 417-424, 1989. [23] T.N.Schriefer, C. W.Hess, K.R.Mills, and N.M.F.Murray, "Central motor conduction studies in motor neuron disease using magnetic brain stimulation," Electroencephalogr. & clin. Neurophysiol., vol. 74, pp. 431-437, 1989. [24] L.G.Cohen, B.Bandinelli, S.Lelli, and M.Hallett, "Non-invasive mapping of hand motor somatotopic area using magnetic simulation," J. Clin. Neurophysiol., vol. 5, pp. 371-372, 1988. [25] V . E . Amassian, R.Q. Cracco, P.J. Maccabee, J.B. Cracco, A. Rudell and L. Eberle, "Suppression of visual perception by magnetic coil stimulation of human occipital cortex," Electroencephalogr. & clin. Neurophysiol., vol. 74, pp. 458-462, 1989. 129 [26] H. Knoepfel, Pulsed High Magnetic Fields, North Holland Publishing Company, Amsterdam, pp. 130-155, 1970. [27] P. Geyer, Systems Lab Report, University of British Columbia, Department of Electrical Engineering, 1986. [28] M . J. R. Poison, A. T. Barker and I. L. Freeston, "Stimulation of nerve trunks with time-varying magnetic fields," Med. & Biol. Eng. & Comput., vol. 20, pp. 243-244, 1982. [29] D. R. Grafham and F. B. Golden, SCR Manual, 6th Edition, Prentice-Hall, USA, pp. 35-70, 1979. [30] Motorola TMOS Power MOSFET Transistor Data, Rev. 2, 1988, Chapter 6: Gate Drive Requirements. [31] D. R. Grafham and F. B. Golden, SCR Manual, 6th Edition, Prentic-Hall, USA, pp. 71-122, 1979. [32] R. F. Graf, The Encyclopedia of Electronic Circuits, 1st Edition, Tabs Books, USA, pp. 729, 1985. [33] W. G. Jung, IC Op-Amp Cookbook, 3rd Edition, Howard W. Sams & Company, USA, pp. 440-445, 1988. [34] J. Cadwell, private communication. [35] R.G. Bickford, M . Guidi, P. Fortesque, and M . Swenson, "Magnetic stimulation of human peripheral nerve and brain: response enhancement by combined magnetoelectrical technique," Neurosurgery, Vol. 20, No. 1, pp. 110 - 116, 1987. 130 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0065454/manifest

Comment

Related Items