UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The use of impedance plethysmography to predict the onset of blood flow beneath a tourniquet cuff McConnell, Gordon 1988

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

Item Metadata

Download

Media
831-UBC_1988_A7 M32.pdf [ 10.04MB ]
Metadata
JSON: 831-1.0097900.json
JSON-LD: 831-1.0097900-ld.json
RDF/XML (Pretty): 831-1.0097900-rdf.xml
RDF/JSON: 831-1.0097900-rdf.json
Turtle: 831-1.0097900-turtle.txt
N-Triples: 831-1.0097900-rdf-ntriples.txt
Original Record: 831-1.0097900-source.json
Full Text
831-1.0097900-fulltext.txt
Citation
831-1.0097900.ris

Full Text

THE USE OF IMPEDANCE PLETHYSMOGRAPHY TO PREDICT THE ONSET OF BLOOD FLOW BENEATH A TOURNIQUET CUFF By Gordon McConnell B.A.Sc, Queen's University, Kinston, Ontario, 1985 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE i n THE FACULTY OF GRADUATE STUDIES (Department of E l e c t r i c a l Engineering) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA A p r i l 1988 (c) Gordon McConnell, 1988 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 U B C T £ \ d ( \ L - ^ N g r l M ^ g C The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date AP&iL. 2 . ^ \qee ABSTRACT Pneumatic tourniquets are i n common use i n hospitals to stop the flow of blood to a limb or d i g i t and provide a bloodless s u r g i c a l s i t e . The r i s k of injury to the tissue beneath the tourniquet cuff from the high pressure i n the cuff could be minimized with a tourniquet system capable of maintaining the cuff pressure at the occlusion pressure, which i s defined to be the minimum pressure that w i l l stop the flow of blood past the cuff . In the work described i n t h i s thesis, a novel impedance-based method was developed to estimate the occlusion pressure while blood flow i s arrested by a tourniquet cuff. This should f a c i l i t a t e the development of actual p r a c t i c a l tourniquet systems capable of safely and r e l i a b l y maintaining the tourniquet cuff pressure near the occlusion pressure for the duration of a surgical procedure. The impedance of the tissue beneath the cuff was measured with a s p e c i a l l y designed impedance plethysmograph. The electrode configurations used with the plethysmograph were evaluated with a computer model of the limb. A rel a t i o n s h i p was found between the impedance of the underlying tissue and the cuff pressure r e l a t i v e to the occlusion pressure that could be used to predict the nearness of the onset of blood flow. Eighteen subjects were used i n the t r i a l s to estab l i s h the r e l a t i o n s h i p . The occlusion pressure was t y p i c a l l y 150 mmHg and the standard error of estimation was 16 mmHg. To establ i s h the relationship, the tissue pressure p r o f i l e along the arm beneath the cuff was measured with a special transducer and controlled by varying the pressures i n the bladders of a dual-bladder tourniquet cuff used to arrest blood flow. An algorithm based on the rel a t i o n s h i p was developed to control the bladder pressures according to the impedance of the underlying tissue. The f e a s i b i l i t y of a tourniquet system that uses impedance plethysmography to keep the cuff pressure as close as possible to the occlusion pressure while s t i l l preventing blood flow was demonstrated with a single subject by c o n t r o l l i n g the pressures i n a dual-bladder tourniquet cuff placed about his arm using the algorithm. Sources of interference and a r t i f a c t encountered i n the study are discussed, and techniques for removing the a r t i f a c t presented. The major contributions of the research i n t h i s thesis were as follows: 1) the development of a computer model for predicting the performance of electrode configurations; 2) the development, design, implementation, and evaluation of a novel impedance plethysmograph; 3) the measurement of the actual tissue pressure p r o f i l e along the surface of the arm beneath various occlusive c u f f s ; 4) the discovery of a quantitative rel a t i o n s h i p between the pressure i n a tourniquet cuff and the impedance of the underlying tissue which for the f i r s t time enables the prediction of the occlusion pressure while the tourniquet cuff pressure i s above the occlusion pressure; i v 5) the i d e n t i f i c a t i o n and p r i o r i t i z a t i o n of the sources of a r t i f a c t and interference together with the development of approaches for detecting and handling them; and 6) the establishment of the f e a s i b i l i t y of c o n t r o l l i n g the pressure i n a tourniquet cuff by incorporating the above into a p r a c t i c a l adaptive tourniquet system. V TABLE OF CONTENTS ABSTRACT i i LIST OF TABLES x LIST OF FIGURES x i ACKNOWLEDGMENT x i i i 1. INTRODUCTION 1 1.1 TOURNIQUET RELATED HAZARDS 1 1.2 THESIS OBJECTIVES 1 1.3 CONTRIBUTIONS OF THE RESEARCH 3 1.4 THESIS OUTLINE 4 2. BACKGROUND 2 2.1 DESCRIPTION OF TOURNIQUET SYSTEMS AND USAGE 6 2.2 OUTLINE OF TYPES OF TOURNIQUET INJURIES 8 2.3 DETECTION OF BLOOD FLOW BENEATH TOURNIQUET CUFFS 10 2.4 PROPERTIES OF IMPEDANCE PLETHYSMOGRAPHY RELEVANT TO BLOOD FLOW MEASUREMENT 11 3. PROBLEM DEFINITION 16 3.1 THE SIGNIFICANCE OF ADAPTIVE CONTROL OF THE CUFF PRESSURE IN TOURNIQUET CUFFS 16 3.2 DEVELOPMENT OF A MODEL TO EXPLAIN IMPEDANCE CHANGES ASSOCIATED WITH BLOOD FLOW UNDER TOURNIQUET CUFFS .18 3.3 IMPEDANCE SIGNALS ASSOCIATED WITH BLOOD FLOW UNDER TOURNIQUET CUFFS 22 3.3.1 P u l s a t i l e impedance 22 3.3.2 Reference to normalize the impedance data 24 3.3.3 Sources impedance related noise and a r t i f a c t 26 3.3.4 Impedance related a r t i f a c t from motion 28 3.4 CLINICAL CONSIDERATIONS FOR EQUIPMENT USED IN IN AN ADAPTIVE TOURNIQUET SYSTEM 31 3.5 OVERVIEW OF APPROACH TO PROBLEM 31 4. COMPUTER MODEL OF THE IMPEDANCE CHANGE IN THE LIMB 33 4.1 THEORY 33 4.2 AN EXAMPLE OF MODEL UTILIZATION 36 4.3 DESCRIPTION OF THE COMPUTER MODEL 37 4.4 ASSUMPTIONS USED WITH THE MODEL 37 4.5 TEST OF THE MODEL 40 4.6 EVALUATION OF ELECTRODE CONFIGURATIONS 42 4.6.1 Configuration 3 44 4.6.2 Configuration 4 44 4.6.3 Dual-bladder cuff configuration 45 5. DEVELOPMENT, IMPLEMENTATION, AND EVALUATION OF THE INSTRUMENTATION USED IN THE STUDY 46 5.1 DESCRIPTION OF THE INSTRUMENTATION DEVELOPED FOR THE WORK 46 5.1.1 Explanation of system components 46 5.1.1.1 Analog c i r c u i t r y to measure impedance 46 5.1.1.2 D i g i t a l data a c q u i s i t i o n 48 5.1.1.3 Aspen A.T.S. 1000 Tourniquet System for cuff pressure control 49 v i 5.1.1.4 Stepping motor for cuff pressure control 49 5.1.1.5 Equipment to monitor the cuff j pressures 50 5.1.1.6 Equipment to measure the tissue pressure p r o f i l e 50 5.1.1.7 Equipment to detect blood flow past the cuff. 52 5.2 DESCRIPTION OF THE PLETHYSMOGRAPHS USED IN THE WORK 53 5.2.1 Specifications for the i n i t i a l plethysmograph 54 5.2.1.1 Electrode and tissue impedances 54 5.2.1.2 Safety requirements for the plethysmograph 54 5.2.1.3 Variable parameters used with the i n i t i a l plethysmograph 54 5.2.3 Description of the i n i t i a l plethysmograph 56 5.2.3.1 Current source 56 5.2.3.2 Instrumentation amplifiers 58 5.2.3.3 High-pass f i l t e r 58 5.2.3.4 Demodulator ^ 59 5.2.3.5 Band-pass f i l t e r 59 5.2.3.6 O s c i l l a t o r 59 5.2.3.7 Performance 60 5.2.4 F i n a l plethysmograph 60 5.2.4.1 Parameter optimization 61 5.3 DESCRIPTION OF THE ELECTRODES CONSIDERED FOR USE WITH THE IMPEDANCE PLETHYSMOGRAPH 63 6. INVESTIGATION OF THE POSSIBLE RELATIONSHIP BETWEEN TISSUE IMPEDANCE AND CUFF PRESSURE 65 6.1 INTRODUCTION TO THE EXPERIMENT 65 6.2 SIGNIFICANCE OF THE PRESSURE PROFILE BENEATH A DUAL-BLADDER TOURNIQUET CUFF 68 6.3 DESCRIPTION OF THE DUAL-BLADDER TOURNIQUET CUFF...70 6.4 EXPERIMENTAL PROCEDURE 70 6.5 ANALYSIS OF THE DATA FROM THE DUAL-BLADDER TOURNIQUET CUFF EXPERIMENT 76 6.6 RESULTS OF THE DUAL BLADDER-CUFF EXPERIMENT 80 6.7 SOURCES OF ERROR IN THE DUAL-BLADDER TOURNIQUET CUFF EXPERIMENT 81 6.7.1 Regulator nonlinearity 81 6.7.2 Limited resolution of the tissue pressure transducer 81 6.7.3 Variation i n the difference between peaks i n the tissue pressure p r o f i l e 82 6.7.4 Error i n the i n i t i a l bladder pressures 82 6.7.5 Time taken for the experiment 82 6.7.6 Presence of impedance pulses at r e l a t i v e l y high cuff pressures 83 6.7.7 Inter-subject v a r i a t i o n i n the difference between mean and s y s t o l i c blood pressure.... 83 7. APPROACH TO ADAPTIVE CONTROL OF THE TOURNIQUET CUFF PRESSURE 85 7.1 INTRODUCTION TO CONTROL 85 7.2 CHARACTERIZATION OF THE SYSTEM 86 v i i 7.3 DESCRIPTION OF THE CONTROL ALGORITHM 88 7.4 PROCEDURE FOR DEMONSTRATING THE FEASIBILITY OF THE CONTROL ALGORITHM 88 7.5 RESULTS OF THE DEMONSTRATION OF THE FEASIBILITY OF THE CONTROL ALGORITHM 92 8. CHARACTERIZATION AND HANDLING OF THE SIGNIFICANT NOISE AND ARTIFACT 95 8.1 OVERVIEW 95 8.2 CHARACTERISTICS AND HANDLING OF MOTION ARTIFACT... 97 9. RESULTS AND DISCUSSION 101 9.1 RESULTS 101 9.2 THE USE OF IMPEDANCE PLETHYSMOGRAPHY TO PREDICT THE ONSET OF BLOOD FLOW 103 9.2.1 The rel a t i o n s h i p i d e n t i f i e d between the tourniquet cuff pressure and the impedance of the underlying tissue 103 9.3 THE NOVEL PLETHYSMOGRAPHIC INSTRUMENT FOR ADAPTIVE TOURNIQUET SYSTEMS 106 9.4 CHARACTERIZATION AND HANDLING OF MOTION ARTIFACT.108 9.5 DISCUSSION OF THE FEASIBILITY OF USING AN ALGORITHM BASED ON IMPEDANCE PLETHYSMOGRAPHY TO CONTROL THE PRESSURE IN A TOURNIQUET CUFF I l l 9.6 THE CLINICAL IMPLICATIONS OF VASCULAR OSCILLATIONS 112 10. RECOMMENDATIONS FOR FURTHER WORK 114 10.1 RECOMMENDED CHANGES AND IMPROVEMENTS 114 10.2 THE ACCURACY OF THE RELATIONSHIP 114 10.3 HANDLING MOTION ARTIFACT 117 10.4 RECOMMENDED CHANGES TO MAKE THE SYSTEM MORE PRACTICAL,FUNCTIONAL, AND ACCEPTABLE IN A CLINICAL APPLICATION 118 11. REFERENCES 120 APPENDIX A: THE MEASUREMENT OF ELECTRODE, TISSUE, AND ARTIFACT RELATED IMPEDANCES AND THE DETERMINATION OF THE REQUIRED PLETHYSMOGRAPH SIGNAL RESOLUTION 125 A l . BASAL IMPEDANCE, PULSATILE IMPEDANCE, ELECTRODE AND MOTION ARTIFACT 125 A2. REQUIRED PLETHYSMOGRAPH SIGNAL RESOLUTION 129 A3. IMPEDANCE FROM THE SUBJECT TO GROUND 130 A4. ELECTRODE IMPEDANCES 132 APPENDIX B: IMPEDANCE PLETHYSMOGRAPH SPECIFICATIONS AND DESIGN 137 B l . GENERAL REQUIREMENTS 137 B2. CURRENT SOURCE 142 B3. INSTRUMENTATION AMPLIFIER 145 B3.1 Gain 145 B3.2 Common-mode voltage 145 B3.2.1 Limb tissue impedances, Zb i , Zb , and Zb 2 147 B3.2.2 Electrode impedances, Z e i , Z e 2 , Ze 3 , and Ze 4 147 B3.2.3 Current source output impedance, Zo.147 B3.2.4 Current source to ground impedance, Z c s g .147 B3.2.5 Subject to ground impedance, Z 0.^...148 B3.2.6 Common mode impedances, Z c m i , and Zc n>2 148 v i i i B3.2.7 D i f f e r e n t i a l impedance, Z<j 148 B3.2.8 Calculation of the common mode voltage 148 B3.3 Amplifier common-mode re j e c t i o n r a t i o 149 B3.4 Common mode input impedance 150 B3.5 Amplifier d i f f e r e n t i a l input impedance 152 B3.6 Noise performance 155 B3.7 Slew rate and frequency response 156 B3.8 Bias current and o f f s e t voltage 158 B3.9 Bias r e s i s t o r s 159 B4. DEMODULATOR 160 B5. FURTHER SIGNAL PROCESSING 160 B6 . DIFFERENTIATOR . 161 APPENDIX C: DISTURBANCE VARIABLES 162 CI. PATIENT RELATED VARIABLES 162 C l . l S y s t o l i c blood pressure 162 CI.2 Limb circumference 162 CI.3 Limb geometry 163 CI.4 Tissue compliance 163 CI.5 Condition of the ar t e r i e s 163 C2. EQUIPMENT RELATED VARIABLES 164 C2.1 Pressure sensor accuracy 164 C2.2 Regulator hysteresis 164 C2.3 Leaks and resultant pressure drops 164 C2.4 I n f l a t i o n flow rate 164 C3. CUFF RELATED DISTURBANCE VARIABLES 165 C3.1 Cuff width 165 C3.2 Cuff manufacture 165 C4. DISTURBANCES RELATED TO CONDITIONS DURING THE PROCEDURE AND TO THE APPLICATION OF THE CUFF TO THE LIMB 166 C4.1 Protective coverings 166 C4.2 Cuff p o s i t i o n 166 C4.3 Snugness of the cuff at application 166 C4.4 Lubrication 167 C4.5 Limb manipulation 167 APPENDIX D: THE LIMB MODEL 168 Dl. INTRODUCTION 168 D2. THE PROGRAMS 168 APPENDIX E: VALIDATION OF THE USE OF AN ULTRASOUND TRANSDUCER OVER THE BRACHIAL ARTERY TO DETECT BLOOD FLOW PAST THE CUFF 196 APPENDIX F: SOURCES OF NOISE AND INTERFERENCE ENCOUNTERED IN THE STUDY 199 F l . RESPIRATION ARTIFACT 199 F2. MUSCLE SPASM 200 F3. VASCULAR OSCILLATIONS 200 F4. PREMATURE VENTRICULAR CONTRACTIONS 205 F5. ALTERNATE CURRENT PATHWAYS 205 F6. VIBRATION 207 F7. CUFF MOTION FROM ARTERIAL EXPANSION 207 F8. BAND ELECTRODES 209 F9. ELECTRONIC INSTRUMENTATION 209 APPENDIX G: EQUIPMENT USED IN THE STUDY 210 APPENDIX H: FIRST AND SECOND DUAL-BLADDER CUFF EXPERIMENTS 213 HI. EXPERIMENTAL PROCEDURE 213 i x H2. DATA ANALYSIS 215 H3. RESULTS 216 APPENDIX I: SINGLE-BLADDER CUFF EXPERIMENTS 220 11. INTRODUCTION 220 12. CUFF TYPES 221 13. ELECTRODE CONFIGURATIONS 223 14. DATA COLLECTION AND ANALYSIS .225 15. RESULTS 231 15.1 F i r s t experiment 231 15.2 Second experiment 234 15.3 Third experiment 236 15.4 Reference 239 APPENDIX J: LEG TOURNIQUET EXPERIMENTS 241 J l . INTRODUCTION 241 J2. THE CUFF 241 J3. ELECTRODES 242 J4. CUFF MOTION ARTIFACT 244 J5. RESULTS 245 APPENDIX K: CONTROL SOFTWARE 246 K l . CONTROL SOFTWARE 246 K2. INITIALIZATION 246 K3. SAMPLING 256 K4. LOOKING FOR A PEAK 256 K5. LOOKING A BASE 257 K6. REFERENCE 257 K7. CONTROL 258 K8. REINFLATE THE PROXIMAL CUFF AND START CONTROL 258 K9. SUSPEND CONTROL AND WRITE THE SAMPLED DATA TO FILES 259 X LIST OF TABLES TABLE 3.1 TISSUE RESISTIVITIES 19 TABLE 4.1 MAGNITUDE OF POSITIVE PULSES WITH THE FOUR CONFIGURATIONS TESTED 42 \ 9 x i LIST OF FIGURES FIG. 2.1 C l a s s i f i c a t i o n of impedance plethysmography 12 FIG. 3.1 Constant pressure tourniquet system 17 FIG. 3.2 Tourniquet system with feedback 17 FIG. 3.3 Penetration of a r t e r i a l blood along the arm as the tourniquet cuff pressure i s varied 21 FIG. 3.4 An example of a p u l s a t i l e impedance waveform and i t s frequency spectrum 25 FIG. 3.5 A r t i f a c t from motion 29 FIG. 3.6 Frequency spectrum of a r t i f a c t from elbow f l e x i o n and extension 30 FIG. 4.1 Impedance sampling f i e l d with a transverse-o f f s e t electrode configuration 35 FIG. 4.2 The arm model g r i d showing the points at which the potentials were calculated 38 FIG. 4.3 Results of the model v e r i f i c a t i o n 41 FIG. 4.4 Evaluation of the e f f e c t i v e impedance s e n s i t i v i t y f i e l d for three electrode configurations 43 FIG. 5.1 Equipment used i n the study 47 FIG. 5.2 Tissue pressure transducer 51 FIG. 5.3 Block diagram of the plethysmograph 57 FIG. 6.1 Electrode configurations used with the dual-bladder cuff experiments 66 FIG. 6.2 Typical tissue pressure p r o f i l e beneath the dual-bladder tourniquet cuff 69 FIG. 6.3 Dual-bladder tourniquet cuff 71 FIG. 6.4 Algorithm for adjusting cuff pressures 73 FIG. 6.5 Variation of the bladder pressures with time 75 FIG. 6.6 Pressure versus pulse amplitude data for one subject showing the l i n e a r regression l i n e 78 FIG. 6.7 Relationship between the normalized impedance pulse amplitude and the r e l a t i v e cuff pressure 79 FIG. 7.1 Sampled control system 87 FIG. 7.2 System response with a change i n cuff pressure 87a FIG. 7.3 Sequence of events during control experiments 90 FIG. 7.4 E f f e c t of r a i s i n g the legs on the control system 93 FIG. 9.1 Test for electrode a r t i f a c t 107 FIG. 9.2 E f f e c t of f i l t e r i n g on motion a r t i f a c t 110 FIG. A l Experimental configuration to measure basal impedance, p u l s a t i l e impedance, and the a r t i f a c t from electrodes and motion 126 FIG. A2 Experimental configuration used to measure Zg, the impedance from the subject to power ground 131 FIG. A3 Experimental configuration used to measure the contact impedance of electrode, E3 133 FIG. Bl Current source 138 x i i FIG. B2 Instrumentation amplifier, demodulator, band-pass f i l t e r , and d i f f e r e n t i a t o r 139 FIG. B3 Carrier and demodulator signals 140 FIG. B4 C i r c u i t r y changes for f i n a l plethysmograph...141 FIG. B5 Equivalent c i r c u i t of subject, current source, and instrumentation amplifier 146 FIG. B6 C i r c u i t s used to determine Zd 154 FIG. B7 Equivalent c i r c u i t for noise calculations....157 FIG. F l additive noise from r e s p i r a t i o n 201 FIG. F2 M u l t i p l i c a t i v e modulation from r e s p i r a t i o n . . . 201 FIG. F3 A r t i f a c t from muscle spasm 202 FIG. F4 Modulation of impedance pulses from vascular o s c i l l a t i o n s 203 FIG. F5 A r t i f a c t from PVC's 206 FIG. F6 A r t i f a c t from an alternate current path 208 FIG. Hi Output from the reference electrodes 219 FIG. II The electrode configurations examined i n the single-bladder experiments 222 FIG. 12 Algorithm for determining the pulse amplitude from the r i s i n g and f a l l i n g edges 230 FIG. 13 Data from one of the subjects showing the res u l t s from a l l eight electrode configurations 233 FIG. 14 The re s u l t s of the second single-bladder cuff experiment 235 FIG. 15 The tissue pressure p r o f i l e s beneath the single-bladder cuff for 10 subjects 237 FIG. 16 The v a r i a t i o n i n impedance pulse amplitude with cuff pressure for the t h i r d s i n g l e -bladder cuff experiment 238 FIG. 17 The output from the reference electrodes for s i x of the subjects showing the v a r i a t i o n i n amplitude with cuff pressure 240 FIG. J l Comparison of the same electrode configuration on the arm and leg 243 FIG. Kl(a) I n i t i a l i z a t i o n 247 FIG. Kl(b) Sampling 248 FIG. Kl(c) Looking for a peak 249 FIG. Kl(d) Looking for the base 250 FIG. Kl (e) Reference 251 FIG. K l ( f ) Control 252 FIG. Kl(g) Reinflate the proximal cuff and s t a r t control 253 FIG. Kl(h) Suspend control and write sampled data to f i l e s 254 ACKNOWLEDGMENT x i i i I would l i k e to thank Dr. James McEwen for his help and encouragement throughout the course of this research, and Dr. M. Beddoes for his helpful suggestions to improve the text of the thesis. I would also l i k e to thank the following people for t h e i r technical assistance and advice: Mr. Glenn Anderson, Mr. Michael Jameson, Mr. Gary Kunz, and Mr. Glen Tracey. I am also thankful for the p a r t i c i p a t i o n of the s t a f f of the Biomedical Engineering Department at the Vancouver General Hospital i n the lab t r i a l s for my research and for t h e i r helpfulness i n general. F i n a l l y , I am gratef u l to the Natural Sciences and Engineering Research Council of Canada for their f i n a n c i a l support. 1. INTRODUCTION 1 1.1 TOURNIQUET RELATED HAZARDS Tourniquets have been used since antiquity to control the loss of blood from amputations. Today, they are s t i l l widely used, primarily i n hospitals to prevent the flow of blood into a limb or d i g i t and thus provide a bloodless s i t e for a wide variety of sur g i c a l procedures. They are usually pneumatic, consisting of an i n f l a t a b l e armlet or cuff which, when pressurized with a gas source, w i l l compress and collapse the underlying blood vessels and prevent the flow of blood through them. Tourniquet use i s not without r i s k . Injury to tissue can occur beneath the cuff from a high application pressure, and both beneath and d i s t a l to the cuff from an application of an excessive duration. Other means of injury also e x i s t ; for example, the sudden loss of pressure i n a pneumatic tourniquet cuff can even lead to death [1]. 1.2 THESIS OBJECTIVES Ideally, a tourniquet system w i l l r e l i a b l y keep the cuff pressure at the minimum value necessary to stop the flow of blood past the cuff under a l l the conditions encountered i n the procedure i n which i t i s used. To approach th i s minimum, there have been tourniquet systems which adaptively altered the cuff pressure i n accordance with the patient's changing 2 intra-operative s y s t o l i c blood pressure [7]. These implementations had the disadvantages of a slow response time, s u s c e p t i b i l i t y to motion a r t i f a c t , and a requirement for an extra cuff to measure the blood pressure. To overcome these disadvantages and f i n d a means of regulating the cuff pressure at a value approaching the minimum necessary, the impedance of the tissue beneath the cuff was investigated i n thi s study to determine i f i t could be used to predict the onset of blood flow and subsequently to control the cuff pressure. It was hypothesized i n thi s study that the t r a n s i t i o n from the condition i n which there i s no blood flow past the cuff to that of blood flow would be characterized by an increase i n the penetration of a r t e r i a l blood into the tissue beneath the cuff from the proximal edge of the cuff as the onset of flow i s approached. Because blood i s more e l e c t r i c a l l y conductive than the other limb tissues, the impedance of the tissue beneath the cuff would decrease as the volume of a r t e r i a l blood within the tissue increases. The increase i n the a r t e r i a l blood volume i n the tissue, which would be an ind i c a t i o n of the closeness of the onset of blood flow, could be determined with impedance plethysmography, a non-invasive technique that i s used to determine volume changes i n tissue by measuring the impedance of the tissue. In turn, as an ind i c a t i o n of the nearness of the onset of blood flow, the signal from the impedance plethysmograph could 3 be used to keep the cuff pressure near the minimum required to prevent blood flow past the cuff, thereby helping to reduce the incidence of tourniquet related i n j u r i e s . With the minimization of the cuff pressure i n mind, the objectives of the thesis were to: 1) investigate the use of impedance plethysmography to predict the cuff pressure at which blood flow w i l l resume i n a limb once the flow has been arrested by an occlusive cuff; 2) design, implement, and evaluate equipment for thi s purpose; 3) account for the expected a r t i f a c t from subject motion; and 4) implement closed loop control of the pressure i n a pneumatic tourniquet cuff using an algorithm based on impedance plethysmography. 1.3 CONTRIBUTIONS OF THE RESEARCH The major contributions of the research i n this thesis were as follows: 1) the development of a computer model into a useful tool for predicting the performance of electrode configurations used with impedance plethysmography, and the v a l i d a t i o n of the model through experimentation; 2) the development, design, implementation, and evaluation of a novel impedance plethysmograph with 4 i t s performance optimized to measure the changes i n the impedance of the tissue beneath an occlusive cuff; 5) the measurement of the tissue pressure p r o f i l e along the surface of the arm beneath various occlusive cuffs which, to the author's knowledge, had not been done before; 4) the discovery of a quantitative relationship between the pressure i n a tourniquet cuff and the impedance of the underlying tissue which can be used to predict the nearness of the onset of blood flow; 5) the i d e n t i f i c a t i o n and p r i o r i t i z a t i o n of the sources of a r t i f a c t and interference associated with the use of impedance plethysmography i n the context of this thesis together with the development of approaches for detecting and handling the most s i g n i f i c a n t types of a r t i f a c t ; and 6) the establishment of the f e a s i b i l i t y of c o n t r o l l i n g the pressure i n a tourniquet cuff on the basis of changes i n the e l e c t r i c a l impedance of the underlying tissue. 1.4 THESIS OUTLINE In Chapters 2 and 3, the background for the research i s summarized and the problem addressed i n the research i s defined. Subsequently i n Chapters 4 and 5, the computer model used to predict system performance and the instrumentation used throughout the study are described. In the next chapter, 5 the experiment from which the rela t i o n s h i p between the tissue impedance and cuff pressure was uncovered i s described. The implementation of a system to control the cuff pressure using an algorithm based on impedance plethysmography i s described i n Chapter 7, and noise and a r t i f a c t are discussed i n Chapter 8. F i n a l l y , i n Chapters 9 and 10, the re s u l t s are discussed and recommendations for improvement and further research are made. i 6 2. BACKGROUND 2.1 DESCRIPTION OF TOURNIQUETSYSTEMS ANDUSAGE A tourniquet i s an instrument for the compression of a blood vessel by application around an extremity to control the c i r c u l a t i o n and prevent the flow of blood to or from the d i s t a l area [2]. Pneumatic tourniquets have an i n f l a t a b l e bladder within a cuff that circumferentially envelops a portion of the extremity. When i n f l a t e d with a gas source, the pressurized bladder compresses the blood vessels within the extremity beneath the cuff. Pneumatic tourniquets have four components: a source of gas to pressurize the bladder, a device to regulate the pressure of the gas within the bladder, a v i s u a l display of the bladder pressure, and a tourniquet cuff (along with the interconnecting tubing) which contains the bladder. Tourniquets can also be non-pneumatic, usually an e l a s t i c or e l a s t i c i z e d c l o t h that i s wrapped t i g h t l y around the proximal end of the extremity to compress the underlying vessels. The Esmarch's tourniquet, a strong f l a t rubber bandage, has h i s t o r i c a l l y been used for the limbs but because of the d i f f i c u l t y i n c o n t r o l l i n g the pressure applied by i t to the tissue, the pneumatic tourniquet i s preferred [3,4]. However, i t continues to be used to exsanguinate the limb (that i s , remove the blood from the limb) prior to i n f l a t i o n of the pneumatic tourniquet. By s t a r t i n g at the d i s t a l end of 7 the limb and wrapping the Esmarch's bandage s p i r a l l y around the limb towards the trunk, the blood i s forced from the limb [5]. For the d i g i t s , an e l a s t i c band or finger from a surgical glove can be used as a tourniquet [6]. After cutting the t i p from the finger of the glove, i t i s r o l l e d towards the base of the d i g i t . This not only exsanguinates the d i g i t , but the r o l l e d up glove finger provides s u f f i c i e n t compression to prevent blood flow into the d i g i t . The primary use of the tourniquet i s to prevent the flow of blood to a limb or d i g i t thus providing a bloodless su r g i c a l s i t e . I t i s estimated that i n North America 10,000 tourniquets are used i n 1 m i l l i o n s u r g i c a l procedures annually [7]. By elevation, wrapping with an Esmarch's bandage, or some other means, the extremity i s exsanguinated. Application of a tourniquet at the proximal end of the extremity w i l l then prevent blood from returning, providing a bloodless surgical f i e l d . The limb or d i g i t can remain bloodless from 1 to 3 hours before i t i s necessary to release the tourniquet to minimize the r i s k of injury from ischemia [4,8,9,10,11,12]. For surgical procedures that take longer, the limb can again be exsanguinated and the tourniquet reapplied after restoring blood flow to the limb for 10 to 20 minutes [8,11]. As well as providing a bloodless operative s i t e , the tourniquet can simultaneously provide the additional function of confining a l o c a l l y infused anaesthetic to the limb, a procedure known as Bier's block [1]. 2.2 OUTLINEOF TYPES OF TOURNIQUET INJURIES [13] The use of tourniquets has long been known to cause injury. Injury can r e s u l t from: 1) over-pressurization, 2) under-pressurization, 3) excess time of i n f l a t i o n , and 4) cuff related causes. Over-pressurization can cause: a) post-operative paralysis of the limb; b) post-operative muscle weakness; c) post-operative pain at the cuff s i t e ; and d) compression i n j u r i e s to blood vessels, nerves, muscles, and skin. Under-pressurization can cause: a) blood i n the surgical s i t e ; b) passive congestion and edema; c) hemorrhagic i n f i l t r a t i o n of nerve tissue; and d) premature release of agents infused into the limb [1,5]. An excessive period of i n f l a t i o n can cause: a) post-operative paralysis of the limb; b) post-operative muscle weakness; c) ischemic injury to the tissue d i s t a l to the cuff; d) excessive post-operative reactive hyperemia; and e) pronounced and prolonged swelling of the limb after release of the tourniquet [14]. 9 Cuff application or peri-operative procedures can cause: a) bruising, b l i s t e r i n g , pinching, and necrosis of the skin; and b) chemical burns due to pooling of dis i n f e c t a n t s . From the extent of the i n j u r i e s due to over-pressurization and under-pressurization, and from the possible dependence of some of the time related i n j u r i e s on the magnitude of the cuff pressure [15], i t i s c l e a r l y desirable to use the lowest possible cuff pressure that w i l l arrest blood flow for the duration of the procedure i n which the tourniquet i s used [5]. Because the patient's s y s t o l i c pressure i s known to be d i r e c t l y related to the occlusion pressure, one means of reducing the cuff pressure has been to relate i t to the patient's pre-operative s y s t o l i c pressure. Where fix e d pressures of 300 mmHg and 500 mmHg for the arms and legs respectively have often been used [16], current recommendations are for a pressure twice the patient's s y s t o l i c pressure for leg cuffs, and 50 to 100 mmHg above s y s t o l i c pressure for arm cuffs [8,9,17]. Recognizing that the patient's s y s t o l i c pressure can vary as much as 70 mmHg i n t r a -operatively [8], tourniquet systems have been implemented that adaptively vary the cuff pressure according to the changing s y s t o l i c pressure [7] Although these methods reduce the cuff pressure and consequently the r i s k of injury, the pressures must s t i l l be subs t a n t i a l l y higher than the minimum necessary i n order to account for other variables. 10 2.3 DETECTION OF BLOOD FLOW BENEATH TOURNIQUETCUFFS To f i n d a way to r e l i a b l y , e f f i c i e n t l y , and quickly r e l a t e the cuff pressure to the onset of blood flow, the t r a d i t i o n a l methods of measuring blood flow [18] were investigated. Only three of the methods, o p t i c a l plethysmography, impedance plethysmography, and ultrasound, met the following necessary c r i t e r i a . 1) The technique must be noninvasive. Many of the techniques employed to measure blood flow require that a blood vessel be exposed to the instrumentation. 2) The technique must be capable of measuring a r t e r i a l blood flow. Because the i n t e r s t i t i a l pressure from a cuff occluding blood flow i s high enough to collapse a l l the vessels except perhaps the larger a r t e r i e s just beneath the edge of the cuff, only a r t e r i a l blood flow i s expected beyond the proximal edge of the cuff when the cuff i s i n f l a t e d to a pressure near the occlusion pressure. 3) I t must be feasi b l e and economical to adapt the method to the task of regulating the pressure i n a tourniquet cuff . Magnetic resonant imaging, for example, would be too expensive to consider. Impedance plethysmography was chosen for thi s study because of i t s v e r s a t i l i t y . By appropriate s e l e c t i o n of the electrode configuration, either large or small volumes can be 11 sampled. Also, deep as well as surface vessels can be ea s i l y monitored [19]. 2.4 PROPERTIES OF IMPEDANCE PLETHYSMOGRAPHY RELEVANT TO BLOOD FLOW MEASUREMENT Plethysmography i s the recording of the changes i n the size of a part of the body as modified by the c i r c u l a t i o n of the blood i n i t [20]. With impedance plethysmography, the change i n size i s determined by measuring the impedance of the part. Impedance plethysmographs can be two-electrode, four-electrode, or guard-ring devices (Fig. 2.1) [21]. A four-electrode device was used for the research i n thi s thesis. In each of these devices a constant amplitude, alternating, sinusoidal current i s passed through the tissue of in t e r e s t and the r e s u l t i n g voltage measured. The frequency of the current i s between 10 kHz and 250 kHz and the magnitude between 0.5 mA and 4 mA [22,23,24,25]. With the appropriate combination of frequency and magnitude, the current does not stimulate the tissue. With the two-electrode plethysmograph, the constant current i s introduced into the tissue through the same electrodes from which the r e s u l t i n g voltage i s measured. The r e s u l t i n g voltage i s : V = I * Z 12 Electrode on t i s s u e Constant current source A --High Input-Impedance d i f f e r e n t i a l a m p l i f i e r B --Unity gain voltage buffer F i g . 2.1 C l a s s i f i c a t i o n s of Impedance piethysinographs, (a) two electrode (b) four electrode (c) guard r i n g 13 where I i s the constant current and Z represents not only the impedance of the tissue under test, but the contact impedance of the electrodes as well. The contact impedance i s the sum of impedances dependent on the electrode, skin, and electrode to skin interface. Furthermore, the pulsations of blood cause changes i n the contact impedance which cannot be separated from the desired p u l s a t i l e impedance changes of the tissue [21]. F i n a l l y , the current concentrates beneath the electrodes, putting more emphasis on t h i s tissue, which may not be desired [21]. With the four-electrode device, the constant current i s introduced through one pair of electrodes, the source electrodes, while the r e s u l t i n g voltage i s sensed from a separate pair of electrodes, the sense electrodes, by a high input-impedance amplifier. Because the current drawn by the amplifier through the sense electrodes i s very small, the e f f e c t of t h e i r impedance i s reduced. Also, less emphasis i s placed on the tissue immediately beneath the source electrodes. The voltage measured i s now: V = I * Zm where Zm i s the mutual impedance shared by the two sets of electrodes and does not include the impedance of the electrodes through which the current i s introduced (the source ^ electrodes). 14 The guard-electrode technique i s similar to the two-electrode technique except that a t h i r d electrode i s placed around the perimeter of one of the other electrodes. By dri v i n g t h i s electrode at the same potential as the electrode i t surrounds, the constant current from the electrode i n the middle i s forced to flow more perpendicular to the electrode than i t otherwise would. This technique i s used to d i r e c t current into tissue i t would otherwise avoid. The impedance of the tissue can be resolved into i t s r e s i s t i v e and reactive components by synchronous demodulation. If the current through the tissue i s : I(t) = Ipeak * COS(w*t) (1) then the voltage at the sense electrodes can be represented by: V(t) = I Z m l * I p e a k * COS(W*t + tt) (2) where Z a i s the mutual impedance, I p e a k i s the maximum instantaneous value of the current, w i s the radian frequency of the current source, a i s the phase angle of the mutual impedance, and t i s time. Multiplying the sense electrode voltage by a reference voltage having a phase angle, B, with the current y i e l d s : V d e m o d ( t ) = K l * cos(w*t + a) * cos(w*t + 6) (3) where K i = I Zm I * I P e a k * C , and C i s the maximum instantaneous value of the reference voltage. Equation (3) can be equivalently expressed as: V d e m o d ( t ) = K2 * cos(a -B) + K2 * cos(2*w*t + a + B) where K2 = K i / 2 The second term can be removed by f i l t e r i n g leaving: V d e m o d ( t ) = K 3 * I Z m ! * C O S ( a - B) (5) Where K3 = I p e a k * C / 2 If K 3 • 1.0, then V d e m o d ( t ) = I Z m i = magnitude of the impedance i f P = a, V d e m o d (t) = | Z m l*cos(a) = r e s i s t i v e component i f B = 0.0, V d e m o d ( t ) = | Z m l * s i n ( a ) = reactive component i f B = n/2 . 16 3. PROBLEM DEFINITION 3.1 THESIGNIFICANCE OF ADAPTIVE CONTROL OF THE CUFF .PRESSURE IN TOURNIQUET CUFFS The optimum tourniquet system w i l l maintain the pressure i n the tourniquet cuff at a value equal to the occlusion pressure for the duration of the procedure. In a constant pressure pneumatic tourniquet system, the cuff pressure i s regulated about a constant cuff pressure setting which must of necessity be much higher than the occlusion pressure (Fig. 3.1). There are, however, a number of disturbance variables which can influence the minimum cuff pressure that i s necessary to collapse the underlying ar t e r i e s (the occlusion pressure) and prevent the flow of blood past the cuff . In order to provide a reasonable degree of confidence that i t w i l l be adequate to prevent the flow of blood past the cuff throughout the entire procedure, the cuff pressure must be high enough to account for each of th e i r e f f e c t s . This s i t u a t i o n i s shown i n F i g . 3.1. The disturbance variables are described i n Appendix C. In this study, an attempt has been made to i d e n t i f y a variable which can accurately predict the cuff pressure at which blood flow past the cuff w i l l resume once i t has been arrested. If the variable can be used as a feedback element i n a control system to modify the cuff pressure setting, then the disturbance variables can be included i n the feedback loop and t h e i r e f f e c t decreased [26] (Fig. 3.2). 1 disturbance v a r i a b l e * cuff pressure s e t t i n g cuff pressure pressure source tiss u e pressure r e l a t i v e to occlusion F i g . 3 .1 Constant pressure tourniquet system r e l a t i v e t i s s u e - i / ^V pressure vcp/ s e t t i n g cuff pressure s e t t i n g disturbance variables cuff pressure pressure source va r i a b l e representative of r e l a t i v e t i s s u e pressure tissue pressure r e l a t i v e to occlusion F i g . 3 .2 Tourniquet system with feedback 1 18 A l t e r n a t i v e l y , the variable can provide a warning of impending flow. This would allow a lower i n i t i a l cuff pressure s e t t i n g since the cuff pressure could be increased during the procedure before blood flow resumes should the i n i t i a l s e tting prove to be too low. 3.2 DEVELOPMENT OF A MODEL TO EXPLAIN IMPEDA^  ASSOCIATED WITH BLOOD FLOW UNDERTOURNIQUET CUFFS The variable chosen as part of thi s study to predict the cuff pressure at the resumption of blood flow past the cuff a f t e r occlusion was the e l e c t r i c a l impedance of the tissue beneath the cuff as measured with a four-electrode impedance plethysmograph. Although the relat i o n s h i p between the cuff pressure and tissue impedance was sought empirically, a simple model of the cause of the impedance changes which was developed i n the course of the work proved he l p f u l i n explaining experimental observations and predicting the ef f e c t of a l t e r i n g certain experimental parameters. The tissue beneath the i n f l a t e d cuff was considered to be an e l e c t r i c a l l y homogeneous and i s o t r o p i c conductive cylinder of tissue. As the cuff pressure decreases towards the occlusion pressure, a r t e r i a l blood penetrates farther into the tissue, reducing i t s impedance. Several assumptions were used with the model: a) The impedance changes experienced near the resumption of blood flow are due to the i n f l u x of a r t e r i a l blood into the tissue. Blood, and 19 espec i a l l y blood plasma, i s very e l e c t r i c a l l y conductive. Table 3.1 shows the impedance of several body tissues including blood. The impedance of blood also exhibits a decrease with an increase i n ve l o c i t y that i s attributed to the alignment of the red blood c e l l s i n the moving blood [27]. As blood enters the tissue beneath the cuff, not only i s less conductive tissue displaced by blood, but the volume of the tissue may increase and the geometric re l a t i o n s h i p between tissue structures and the plethysmograph electrodes changes. A l l these af f e c t the impedance of the tissue as measured with the plethysmograph although, because the tissue i s enclosed on the limb surface by the confining tourniquet cuff, the actual contribution of each i s not known. TISSUE blood plasma muscle, transverse muscle, longitudinal lung fa t RESISTIVITY (ohm-cm) 150 63 1600 300 1275 2500 SPECIES human mammals dog dog mammals mammals Table 3.1 TISSUE RESISTIVITIES [28] 20 b) The penetration of a r t e r i a l blood into the tissue increases as the cuff pressure decreases or some other relevant variable changes to bring the point of the resumption of blood flow nearer. This occurs because the pressure applied to the limb by the cuff i s not uniform; instead, the pressure d i s t r i b u t i o n t y p i c a l l y shows a maximum near the cuff center and minima at the ends. The e f f e c t i s i l l u s t r a t e d i n F i g . 3.3. As the cuff pressure i s lowered, the length of artery at the proximal edge of the cuff that i s exposed to a pressure lower than what i s necessary to collapse i t increases and a r t e r i a l blood i s able to penetrate farther into the tissue beneath the cuff. c) The volume of a r t e r i a l blood i n the tissue beneath the cuff increases as the cuff pressure decreases or some other relevant variable changes to bring the point of the resumption of blood flow nearer. The volume not only increases because the blood penetrates farther into the tissue, but the p u l s a t i l e translumenal pressure across the a r t e r i a l walls increases allowing the non-collapsed ar t e r i e s to expand further a n d . f i l l with more blood. d) There i s no blood i n the venous vessels, c a p i l l a r y bed, or smaller members of the a r t e r i a l tree i n the tissue beneath the cuff. Because the cuff pressure near the point of resumption of blood 21 TOURNIQUET APPLIED TO ARM Pressu r e Along C u f f O c c l u s i o n Pressure D i s t a n c e Along Arm P e n e t r a t i o n o f A r t e r i a l Blood Along Arm F i g . 3.3 Penetration of a r t e r i a l bipod along the arm as the tourniquet cuff pressure i s varied. 22 flow i s s u f f i c i e n t to collapse a l l but the larger a r t e r i e s near the cuff edges, i t i s s u f f i c i e n t to collapse and exsanguinate the smaller vessels throughout the tissue beneath the cuff. e) The penetration of a r t e r i a l blood into the tissue beneath the cuff i s p u l s a t i l e when the cuff i s i n f l a t e d above the occlusion pressure. Unlike i n the venous vessels, there are no valves i n the large a r t e r i e s i n the limbs. Thus as the p u l s a t i l e blood pressure proximal to an i n f l a t e d cuff f a l l s below the i n t e r s t i t i a l pressure transmitted by the cuff, the a r t e r i e s beneath the proximal edge of the cuff collapse and the a r t e r i a l blood i s forced back out the a r t e r i e s . 3.3 IMPEDANCE SIGNALSASSOCIATED WITH BLOOD FLOW UNDER TOURNIQUET CUFFS 3.3.1 P u l s a t i l e impedance The tissue impedance has two components: a large basal impedance which corresponds to the impedance of the exsanguinated limb tissue beneath the cuff, and a small p u l s a t i l e component which corresponds to the incursion of a r t e r i a l blood into the tissue. This study focussed on the p u l s a t i l e component since: 23 a) the dominant impedance change near the onset of flow was expected to be due to the p u l s a t i l e i n f l u x of a r t e r i a l blood, and b) the frequency spectrum of much of the noise encountered i n the study showed that the noise generally occupied a frequency band lower than the fundamental frequency of the p u l s a t i l e impedance. The r a t i o of p u l s a t i l e to basal impedance was found to be approximately 0.002 for the arm without a r t e r i a l occlusion. When a cuff was placed over the tissue and i n f l a t e d to near occlusion pressure, a three-fold increase i n amplitude was experienced. The increase i n amplitude i s a r e s u l t of the hyperemic response of the vascular system which may have more than one cause. Assuming the increase corresponds to an increase i n a r t e r i a l volume, i t may be due to the nonlinear c h a r a c t e r i s t i c of the volume e l a s t i c modulus of the a r t e r i e s . Because the a r t e r i a l translumenal pressure i s much lower with the cuff i n f l a t e d , the e l a s t i c i t y of the artery i s less and the artery i s capable of more expansion for a given p u l s a t i l e pressure [29]. Another cause might be an autoregulation c a p a b i l i t y of the a r t e r i e s involved. One theory postulates that t h i s c a p a b i l i t y i s due to a decrease i n the tension of smooth muscle i n the a r t e r i a l wall i n response to a decrease i n the mean pressure [30]. 24 The frequency content of the p u l s a t i l e impedance waveform varies with the subject as well as the electrode placement and configuration. It i s also affected by the degree of a r t e r i a l occlusion caused by the pressure i n the tourniquet cuff surrounding the tissue. Furthermore, i t depends on the subject's heart rate which can be considered quasi-periodic. The heart rate i s under the control of the autonomic nervous system and can change d r a s t i c a l l y i n response to stimuli [31]. It i s also affected by r e s p i r a t i o n . While bearing i n mind the factors that a f f e c t the frequency content, i t was found that the energy of the pulses was concentrated within the f i r s t f i v e harmonics. For the average heart rate of 72 beats per minute, the f i r s t f i v e harmonics occupy the narrow band of frequencies from 1.2 Hz to 6 Hz. The frequency spectrum of a t y p i c a l p u l s a t i l e impedance waveform i s shown i n F i g . 3.4. 3.3.2 Reference to normalize the impedance data In order for the impedance to be a variable independent of the subject and variations i n the electrode c h a r a c t e r i s t i c s , i t i s necessary to normalize the impedance derived from the sense electrode voltage with a reference. Three approaches were t r i e d . With the f i r s t approach, an attempt was made to normalize the impedance pulse height with an i n i t i a l spike which precedes each impedance pulse due to the negative s e n s i t i v i t y of the impedance s e n s i t i v i t y f i e l d of a s p e c i f i c electrode configuration. The impedance s e n s i t i v i t y f i e l d i s discussed i n (a) PULSATILE IMPEDANCE _| , , , , , 1 0 2 4 6 TIME (a«eand«) (b) PULSATILE IMPEDANCE FREQ. SPECTRUM 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 IB FREQUENCY (Hi) F i g . 3.4 An example of a p u l s a t i l e impedance waveform an# i t s frequency spectrum. The p u l s a t i l e impedance waveform shown i n (a) was f i l t e r e d from 0.5 Hz to 100 Hz and sampled at 360 Hz. The frequency spectrum shown i n (b) was determined with a Fast Hartley Transform from the 2048 samples shown i n (a). A window was not used. The approximate impedance change represented by each pulse i n (a) i s 50 milliohm with a basal impedance of 10 ohm. 26 the next chapter. Unfortunately, the electrode configuration had to be too exact to give reproducible r e s u l t s . In the second approach to f i n d a reference, a separate set of electrodes (reference electrodes) proximal to the sense electrodes was used to get a second impedance s i g n a l . By placing the electrodes far enough proximally, blood flows e n t i r e l y beneath these electrodes before flowing beneath the sense electrodes. The amplitude of the p u l s a t i l e impedance signal was then compared to the amplitude of the p u l s a t i l e reference s i g n a l . With th i s scheme, the reference signal often showed considerable v a r i a t i o n with cuff pressure and was not a r e l i a b l e standard. A t h i r d approach was more successful. Using a dual-bladder cuff, the proximal bladder pressure was l i n e a r l y decreased with time while the pressure i n the d i s t a l bladder remained high enough to prevent blood flow past the cuff. The maximum height of the impedance pulses achieved when the proximal bladder pressure was decreased was used to normalize the impedance data. 3.3.3 Sources of impedance related noise a n d a r t i f a c t One of the contributions of t h i s work was to i d e n t i f y and p r i o r i t i z e the various kinds of a r t i f a c t and interference associated with the use of impedance plethysmography i n the context of th i s thesis and to develop approaches for detecting and handling them. The sources of noise and a r t i f a c t 27 encountered i n th i s study include subject motion, res p i r a t i o n , muscle spasm, vascular o s c i l l a t i o n s , premature ventricular contractions, alternate current pathways, vibration, cuff motion from a r t e r i a l expansion, band electrodes, and the electronic instrumentation. Motion a r t i f a c t i s discussed below and the other sources of noise and interference are discussed i n Appendix F. Several other sources are expected i n the c l i n i c a l environment although they were not experienced i n the laboratory s e t t i n g . Like the premature ventricular contractions encountered during the lab t r i a l s , a variety of arrhythmias exi s t which can change the a r t e r i a l blood pressure for one or two beats and momentarily depress the p u l s a t i l e impedance amplitude. Other i n t e r f e r i n g sources would include the variety of energy sources within the c l i n i c a l environment, es p e c i a l l y those with a spectrum that overlaps the operating frequency of the impedance plethysmograph. For example, some electrocardiographs detect whether a lead i s unattached by measuring the impedance between leads by passing a high frequency current through the patient just l i k e the impedance plethysmograph. Current from electrosurgery can also a f f e c t impedance plethysmographs [32]. Local FM transmissions and fluorescent l i g h t s are other potential i n t e r f e r i n g sources. 3.3.4 Impedance related a r t i f a c t from motion Motion a r t i f a c t has been recognized as the major source of interference i n impedance plethysmography which would have 28 to be handled i n order to develop a tourniquet system capable of routine use i n r e a l i s t i c c l i n i c a l situations [24,33,34,35]. Any motion which involves the musculature beneath the plethysmograph electrodes w i l l generate a r t i f a c t by changing the volume of tissue and also the r e l a t i v e p osition of anatomical structures of varying e l e c t r i c a l impedances within the t issue. Motion can also cause a r t i f a c t by varying the impedance of alternate pathways for the current from the plethysmograph current source. The magnitude of the a r t i f a c t depends mostly on the degree of involvement of the musculature beneath the electrodes. For electrodes mounted to measure the impedance of the upper arm, elbow f l e x i o n and extension can produce peak deflections 20 times as large as those r e s u l t i n g from vascular events. Motion of the other arm such as r a i s i n g i t from below the waist to overhead, however, causes very l i t t l e a r t i f a c t . Flexion and extension of the wrist can cause a r t i f a c t larger than the magnitude as the desired s i g n a l . The r e l a t i v e magnitudes are i l l u s t r a t e d i n F i g . 3.5. The frequency content of motion a r t i f a c t tends to be low. Fi g . 3.6 shows the spectral analysis of the a r t i f a c t from elbow f l e x i o n and extension shown i n F i g . 3.5. 29 NOISE AND ARTIFACT P f I I I1 I I 1 I 1 "I I " ' T — l I 1 I 'I H "I I1 I1 I 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 TIME (seconds) F i g . 3.5 A r t i f a c t from motion. 1. Impedance pulses 2. The background noise with the cuff pressure s u f f i c i e n t to suppress the impedance pulses 3. A r t i f a c t from moving fingers 4. A r t i f a c t from wrist f l e x i o n and extension 5. A r t i f a c t from elbow fl e x i o n and extension 6. A r t i f a c t from gross movement of the opposite arm The v e r t i c a l scale for each tracing i s impedance and i s the same. A l l the^ tracings were obtained with the same equipment settings except that with trace (1) a lower cuff pressure was used. M O T I O N F R E Q U E N C Y S P E C T R U M ., L** 0 t 2 3 4 5 6 7 B 9 10 11 12 13 H 15 16 17 18 19 FREQUENCY (Hz) Fig. 3.6 Frequency spectrum of artifact from elbow flexion and extension. 31 3.5 CLINICAL CONSIDERATIONSFOREQUIP^USED I N I N A N ADAPTIVE TOURNIQUET SYSTEM Because the plethysmograph and associated equipment would be used for laboratory t r i a l s , i t was b u i l t to meet the hospital's safety standards of leakage current and grounding resistance. The leakage current requirements for a r i s k class 2 device are given i n CSA standard C22.2 no. 125, Electromedical Equipment [36] . C l i n i c a l usage would place other requirements on the equipment. It must not off e r a low resistance to el e c t r o s u r g i c a l current or present a hazard to the patient or s t a f f during d e f i b r i l l a t i o n . I t must also survive the harsh c l i n i c a l environment which includes electrosurgery, d e f i b r i l l a t i o n , and exposure to solvents. 3.6 O V E R V I E W P R O B L E M The variable used to predict the onset of blood flow was the p u l s a t i l e impedance measured with a four-electrode impedance plethysmograph and normalized with a reference. The p u l s a t i l e impedance was band lim i t e d from 1.5 Hz to 30 Hz with analog f i l t e r s . An analog 60 Hz n o t c h - f i l t e r was also used to reduce l i n e frequency interference. The r e l a t i v e l y high lower corner-frequency w i l l cause both amplitude and phase d i s t o r t i o n of the p u l s a t i l e impedance since 1.5 Hz correspondsp to a heart rate of 90 Hz. Much of the noise experienced during the study, however, was concentrated within the bandwidth of 0 32 Hz to 1.5 Hz. Since the object of t h i s study was to f i n d a variable to represent the onset of blood flow rather than to restore the impedance signal, the loss i n f i d e l i t y was acceptable to improve the signal to noise r a t i o . The reduction i n noise was found to outweigh the loss of signal information both from empirical experimentation and from analysis of the frequency content of the signal and noise sources. 4. COMPUTER MODEL OF THE IMPEDANCE CHANGE IN THE LIMB 33 4.1 THEORY A model was developed i n t h i s work to evaluate prospective electrode configurations for use with impedance plethysmography. This model was based on e a r l i e r work by Penny and others [19,27,37,38], which i n turn was based on expressions derived by Geselowitz [39], Lehr [40], and M o r t a r e l l i [41]. The Geselowitz-Lehr r e s u l t can be obtained from M o r t a r e l l i ' s r e s u l t i f the following assumptions are made: 1) The f i e l d s are qua s i - s t a t i c ; that i s , the magnetic and propagative e f f e c t s can be ignored. Some experimental v a l i d a t i o n exists for the implied frequency independence [19]. 2) The electrodes do not move when the tissue impedance changes. With the electrodes attached to the inner surface of an i n f l a t e d cuff, they w i l l not move lo n g i t u d i n a l l y or circumferentially. The i n f l a t e d cuff w i l l also r e s t r i c t , i f not prevent, r a d i a l movement. The model used i n this study was based on the Geselowitz-Lehr r e s u l t : 6Z = Z ( t i ) - Z(to) - J J J - ( g ( t i )-g(to ) ) * (La (ti )-Lb (to ) ) dv V 34 where: 5Z i s the change i n measured impedance between times t i and to ; Z(t) i s the measured impedance at time t; V i s the volume of tissue under consideration; g(t) i s the conductivity everywhere i n volume V at time t; La (ti) i s the lead vector associated with electrode set a at time ti ; Lb(to) i s the lead vector associated with electrode set b at time to; and La ( t i ) * L b ( t o ) i s the dot product between the two vectors. The lead vector i s the voltage gradient f i e l d that re s u l t s when a unit current i s passed between the indicated pair of electrodes. By considering the dot product of the lead vectors (that i s , the second term i n parentheses i n the integrand), an impedance s e n s i t i v i t y f i e l d can be established which gives the s p a t i a l s e n s i t i v i t y of the volume to conductance changes within the volume. Since the integrand i s non-zero only where a change of conductance occurs, r e s u l t i n g changes i n the^ measured impedance of the entire volume can be estimated for l o c a l i z e d changes i n conductance from the impedance s e n s i t i v i t y f i e l d . F i g . 4.1 gives an example of the voltage gradient f i e l d from a transverse-offset sense electrode configuration. The source electrodes are assumed to be far enough apart that the voltage gradient f i e l d from a unit current passing through these electrodes i s uniform along the limb. The s e n s i t i v i t y to a conductance change at any point region of negative _ sensitivity—^ Se (a) SAMPLING FIELD MODEL OF THE^RM So...Source E l e c t r o d e s Se...Sense E l e c t r o d e s outline of cuff—7 So So C Se ELECTRODE PLACEMENT above o c c l u s i o n pressure below o c c l u s i b r (c) PULSATILE IMPEDANCE CHANGES pressure F1g. 4.1 Impedance sampling f i e l d with a t r a n s v e r s e - o f f s e t e l e c t r o d e c o n f i g u r a t i o n . 36 i n the volume i s dependent on both the density of the two sets of flux l i n e s at that point and the angle between the two sets of f l u x l i n e s at that point. The integrand w i l l be non-zero only where conductance changes occur i n the volume which are confined to the a r t e r i e s within the volume. By following the p u l s a t i l e incursions of a r t e r i a l blood along the a r t e r i e s into the volume of tissue, the r e l a t i v e change i n impedance and thus the effectiveness of a s p e c i f i c electrode configuration can be estimated. The length of the incursions along the art e r i e s w i l l depend on the cuff pressure. The e f f e c t i v e impedance s e n s i t i v i t y f i e l d i s considered that part of the f i e l d where incursions of a r t e r i a l blood would contribute more than 5% to the magnitude of the impedance pulses at the occlusion pressure. 4.2 AN EXAMPLE OFMODELUTILIZATION With the transverse-offset sense electrode configuration using conductive adhesive electrodes, for example, blood i n the artery shown w i l l be l i m i t e d to a region of negative s e n s i t i v i t y at high cuff pressures (Fig. 4.1). Thus the penetration of blood into t h i s region which causes a decrease i n the tissue impedance w i l l actually cause an increase i n the measured impedance. As the cuff pressure i s lowered, the blood w i l l pass through a region of zero s e n s i t i v i t y into a region of p o s i t i v e s e n s i t i v i t y . The i n f l u x of blood into this region w i l l return the measured impedance change to zero and eventually cause a net decrease. As the p u l s a t i l e blood pressure f a l l s during diastole, the blood i s forced back out 37 the artery. The negative pulse then returns to i t s baseline, becomes posit i v e again, and f i n a l l y returns to i t s baseline. These events are i l l u s t r a t e d i n the actual recording using t h i s electrode configuration i n F i g . 4.1. The changes i n impedance at the beginning of each pulse are compressed i n time because the pressure waveform has a risetime much faster than i t s f a l l t i m e . 4 . 3 DESCRIPTION QF.JTHE A segment of the limb was modelled i n two dimensions only as an e l e c t r i c a l l y homogeneous and i s o t r o p i c medium by a 65 by 30 point g r i d with 0.1 inch (0.254 cm) spacing (Fig. 4.2). The model can represent a longitudinal section of the limb. To f i n d the voltage gradient f i e l d that r e s u l t s from passing a unit current through one set of electrodes, the potential at each point i n the array was determined when the potential at the each of the electrodes i n the set was fixed. An algorithm was then used to draw the potential gradient l i n e s between electrodes. After the r e s u l t i n g set of l i n e s for each set of electrodes was determined, they were plotted on the same sheet of paper or displayed on a computer video terminal at the same time. The model programs are included i n Appendix D. 4.4 ASSUMPTIONS USEDWITH THE MODEL The assumptions and s i m p l i f i c a t i o n s used with the model include the following: 38 Fig. 4.2 The arm model grid showing the points at which the potentials were calculated. 1) The arm i s modelled i n two dimensions only. 39 2) The limb tissue i s considered e l e c t r i c a l l y homogeneous. The various limb tissues include bone, ca r t i l a g e , muscle, f a s c i a , and perhaps some blood and lymph and the limb segment i s c l e a r l y not homogeneous. 3) The limb tissue i s considered e l e c t r i c a l l y i s o t r o p i c . This assumption i s not true. Muscle, the dominant tissue i n the exsanguinated limb, for example, has a conductivity i n the longitudinal d i r e c t i o n that i s 5 times that i n the transverse d i r e c t i o n [28]. 4) The f i e l d s i n the tissue are quas i - s t a t i c at the frequency used [19]. v 5) The lead vectors do not change with the incursion of blood into the tissue. To calculate the impedance s e n s i t i v i t y f i e l d , both lead vectors were determined at time to, before the conductance changes occurred at time t i . The actual dot product of the lead vectors i n the Geselowitz-Lehr r e s u l t requires that one lead vector be determined at to and the other at t i [19]. 6) The solution of the potential within the tissue at f i n i t e points accurately r e f l e c t s the continuous solution. 40 7) The voltage gradient f i e l d s that constitute the lead vectors are confined by the ends of the section of the limb representation. 8) The area outside the perimeter of the model i s a perfect insulator. 9) The volume of the limb segment i s s t a t i c and does not change with the incursion of blood into the tissue. 10) The electrodes do not move with the incursion of blood into the tissue. 4.5 TEST OF THEMODEL In spite of these many l i m i t a t i o n s , the model was found to be e f f e c t i v e i n subjectively predicting the performance of electrode configurations. To test the model further, four electrode configurations were chosen and the r e l a t i v e s e n s i t i v i t y and presence of attributes for each predicted from the model. Each configuration was then applied to four subjects and the res u l t s compared. F i g . 4.3 shows the results when the average pulse amplitudes from configurations l a b e l l e d 2, 3, and 4 were normalized with the average pulse amplitude from configuration 1 for each of the subjects. It also shows normalized p u l s a t i l e impedance predicted from the model. The measured re s u l t s agree quite well with the model predictions, es p e c i a l l y considering the e f f e c t of high-pass f i l t e r i n g the impedance pulses has not been taken into account. MODEL VERIFICATION 1.7 § o o n o 0) B o cu 6 o 1.6 -l.S -a.4 1.3 -1.2 -1.1 1 -0.9 -0.8 0.7 0.6 - \ 0.5 0.4 -0.3 0.2 H 0.1 0 M 3 A 2 M 4A 3 A 11 3B 4B 4A 3B 2 3A 4A l 4 B M C O N F I G U R A T I O N F i g . 4.3 Results of the model v e r i f i c a t i o n . The numbers represent the four subjects tested. The numbers with no s u f f i x and the "A" s u f f i x represent r e s u l t s taken with the cuff pressure 15 mmHg below the occlusion pressure. The numbers with the "B" s u f f i x represent r e s u l t s taken with the cuff pressure 30 mmHg below occlusion pressure. The model predictions" are indicated with the l e t t e r "M". 42 The presence of posi t i v e impedance pulses preceding the negative pulses were predicted from the model. Table 4.1 shows the predictions along with the experimental observations. The presence of high-pass f i l t e r i n g would aff e c t the model predictions. SUBJECT CONFIG. 1 CONFIG^ CONFIG. 3 CONFIG. 4 1 small none none large 2 large small small very large 3 none none none small 4 large small small very large model large small small very large TABLE 4..1......_ MAGNITUDE OF JPJMITJVE PULSES WITH THEFOURCONFIGURATIONS TESTED 4.6 EVALUATION OF ELECTRODE CONFIGURATIONS The electrode configurations used i n the experiments with both single-bladder and dual-bladder cuffs are shown i n F i g . 4.4. Because the computer model was for a longitudinal section of the arm, the transverse electrode configurations could not be evaluated. However, the use of lead vectors to determine the e f f e c t i v e impedance s e n s i t i v i t y f i e l d provided insight into the expected impedance signals from the transverse configurations. 43 (a) Electrode configuration 3 (b) Electrode configuration 4 (c) Electrode configuration used with the dual-bladder cuff experiments F i g . 4.4 Evaluation of the effective impedance sens i t iv i ty f i e l d for three electrode configurations. 4.6.1 Configuration 3 44 Electrode configuration 3 was used i n the f i r s t and second experiments with the single-bladder c u f f s . The voltage gradient f i e l d from the source electrodes i s assumed to be directed p a r a l l e l to the long axis of the arm and of uniform density. The approximate position of the brachial artery i s shown, the shoulder i s to the l e f t , and the inside of the arm i s down. Because of opposing horizontal components of the two f i e l d s there w i l l be a region of negative s e n s i t i v i t y i n the impedance s e n s i t i v i t y f i e l d and a r e s u l t i n g increase i n the measured impedance with the encroachment of blood into this region. After that, the s e n s i t i v i t y w i l l be p o s i t i v e . Because the voltage gradient f i e l d s do not have a small angle between them where the f i e l d s are the densest, the o v e r a l l s e n s i t i v i t y w i l l not be high. The configuration w i l l put emphasis on the tissue around the brachial artery but w i l l sample a large volume of tissue, including the other ar t e r i e s of the arm. Because i t samples a large volume, i t w i l l be sensitive to movement involving the musculature of the arm. 4.6.2 Configuration 4 Configuration 4 was also used i n the f i r s t and second experiments with the single-bladder c u f f s . There w i l l be a portion of the impedance s e n s i t i v i t y f i e l d with a negative S s e n s i t i v i t y along the artery, but i t w i l l be much smaller i n 45 proportion to the region of p o s i t i v e s e n s i t i v i t y than that of configuration 3. The largest s e n s i t i v i t y along the artery w i l l be about the midpoint between the sense electrodes where the angle between the voltage gradient f i e l d s i s small. The e f f e c t i v e impedance s e n s i t i v i t y f i e l d w i l l be small and w i l l emphasize the tissue around the brachial artery. 4.6.3 Dual-bladder cuff configur.ation The configuration shown i n F i g . 4.4 (c) was used i n the second and t h i r d experiments with the dual-bladder cuf f . The configuration used i n the f i r s t was similar but with the sense electrodes moved d i s t a l l y and a pair of reference electrodes between the sense electrodes and proximal source electrode. The configuration i s shown with outlines of the dual-bladder cuff and the two bladders. There i s no negative s e n s i t i v i t y region i n the e f f e c t i v e impedance sampling f i e l d , which as with configuration 4, i s small and emphasizes the tissue around the brachial artery. I t also does not extend beyond the midpoint of the proximal bladder where the peak of the pressure beneath the bladder i s expected to be, which i s an important consideration i n the experiments with the dual-bladder cuff. 46 5. DEVELOPMENT, IMPLEMENTATION, AND EVALUATION OF THE INSTRUMENTATION USED IN THE STUDY The majority of the equipment used i n this study was available from the Biomedical Engineering Department at the Vancouver General Hospital and i s shown i n F i g . 5.1 and described i n Appendix G. Because i t was necessary to design and b u i l d a novel impedance plethysmograph having features optimized for the prediction of the occlusion pressure, i t i s described i n d e t a i l i n this chapter. Also mentioned are the electrodes considered for use with the plethysmograph. 5.1 DESCRIPTION OF THE INSTRUMENTATION DEVELOPED FORTHE WORK 5.1.1 Explanation of system components The pieces of equipment that constitute the instrumentation shown of Fig . 5.1 are grouped into the functional components of the o v e r a l l system and described i n the following sections. 5.1.1.1 Analog c i r c u i t r y tomeasureimpedance The impedance plethysmograph provided a voltage output representative of the impedance, resistance, and reactance of the tissue being sampled. The signal was band-limited and represented only the time varying component. The output from the plethysmograph was connected to the input of the 8811A 47 o o o o o array of l i g h t emitting diodes mercury manometer sphygmomanometer bulb Arteriosonde r e c t l f l e chart recorder with four amplifiers remote box plethys-mograph pressure trans-ducer 50 pslg hospital a i r supply rotary regulating valves < e i t h e r data acqu i s i t i o n board stepping motor stepping motor c o n t r o l l e r AT-compatlble personal computer Aspen automatic tourni-quet system CONNECTIONS .... e l e c t r i c a l >••• pneumatic 1. tissue pressure transducer 2. ultrasound sensor 3. electrodes 4. c u f f F i g . 5.1 Equipment used 1n the study. 48 b i o e l e c t r i c amplifier plug-in module i n the chart recorder mainframe. The A.C. mode of the amplifier was used and a range of gains, low frequency half-power points, and high frequency half-power points were available. With chart speeds of 0.25 mm/s to 100 mm/s, the chart recorder provided a record of the channel input. Also available was a voltage output of the amplified and f i l t e r e d input s i g n a l . The output signal was a.c.-coupled to the data a c q u i s i t i o n board; the coupling added a lower frequency half-power point of 0.5 Hz. 5.1.1.2 D i g i t a l data a c q u i s i t i o n The analog signals from the chart recorder plug-in amplifiers were connected to the DT707 Screw Terminal Panel of the DT2801 Single Board Analog and D i g i t a l I/O System i n s t a l l e d i n the Copam PC 501-TURBO computer. Although having other c a p a b i l i t i e s , the I/O system was used only for multiplexing the analog signals and for analog to d i g i t a l conversion. The DT2801 was configured to accept single-ended bipolar signals. Up to four channels from the chart recorder amplifiers were attached for multiplexing and A/D conversion. The outputs from the 8805B c a r r i e r amplifiers were connected d i r e c t l y while the outputs from the 8811A b i o e l e c t r i c amplifiers were a.c.-coupled with an RC combination giving a low frequency half-power point of 0.5 Hz. The sync output from the Wavetek 182A Function Generator provided an external trigger for the conversions. A sampling rate of either 360 Hz or 180 Hz was used for frequency analysis of the signals; a rate of 90 Hz was used for general data c o l l e c t i o n and 49 control. E x i s t i n g C-language subroutines were used for data acqui s i t i o n , software reset, and software error analysis. The signals were represented d i g i t a l l y as off s e t binary. 5.1.1.3 Aspen"A.T.S. 1000 Tourniquet -System for c u f f p r e s s u r e control For many of the tests, the s t a t i c cuff pressures were adjusted and controlled with the Aspen A.T.S. 1000 Tourniquet System which can be adjusted i n steps of 1 mmHg. These provided the convenience of a d i g i t a l readout of the cuff pressure and accurate regulation of the cuff pressure to within ± 4 mmHg. Although the Tourniquet System manufacturer's accuracy and hystersis s p e c i f i c a t i o n s are as noted i n Appendix G, the actual hysteresis and accuracy were measured and both found to be ± 2 mmHg. 5.1.1.4 St.epping motor for cuff pressure control For reasons mentioned i n Appendix H, i t became necessary to vary the cuff pressure continuously and much more quickly than can be done with the Aspen A.T.S. 1000 Tourniquet Systems. To accomplish t h i s , stepping motors were attached to rotary pressure regulators intended for manual control of pressure from 0 mmHg to 300 mmHg. An input pressure of about 500 mmHg was provided with another pressure regulator attached to the 50 p s i (2590 mmHg) hospital a i r supply. The stepping motors were driven with 2N6045 darlington power trans i s t o r s connected to a Tecmar Inc. PC-Mate Dual Stepper Motor 50 Controller Board i n s t a l l e d i n the Copam PC 501-TURBO computer. The power to the stepper motor was supplied with a Standard Power Supplies Inc. SPS-60-5 f i v e v o l t power supply. The software to control the cuff pressures was written i n C-language and made use of existing subroutines to communicate with the motor c o n t r o l l e r board. 5.1.1.5 E g u i p m e n t t o m p n i t p r t h e c u f f p r e s s u r e s When the cuff pressures were controlled with the Aspen A.T.S. 1000 Tourniquet Systems, the pressures were displayed on th e i r d i g i t a l readouts. With stepping motor control, however, separate pressure transducers (HP Model 1280) were connected to cuff bladders. The transducers were attached to the model 8805B Carrier Amplifier plug-in modules i n the chart recorder. 5.1.1.6 Eguipmentto measure t h e t i s s u e pressure p r o f i l e The tissue pressure p r o f i l e along the arm beneath the cuff was measured with a transducer developed i n the Biomedical Engineering Department at VGH [42]. The transducer i s e s s e n t i a l l y a sealed normally-closed membrane switch array, 2.8 cm wide and 13.8 cm long (Fig. 5.2). The contacts are located along the long axis and are equally spaced 1.92 cm apart. By introducing a i r into the sealed chamber, each pair of normally closed contact w i l l separate when the pressure inside the chamber i s the same as the pressure applied II O O O © Q 1.92cm 13.8cm 0 0 pneumatic connector e l e c t r i c a l connector F i g . 5.2 Tissue pressure transducer. The f i v e normally-closed switches on the sealed membrane can be opened by pressurizing the chamber between the contact p a i r s . 52 externally to that membrane contacts. The a i r was introduced with either a sphygmomanometer bulb or a manually turned rotary pump from a Cobe Therapeutic Plasma Exchanger. The chamber pressure was measured with a W.A. Baum mercury manometer. The opening and closing of the contacts was indicated with l i g h t emitting diodes i n series with the contacts. Before the measurements were taken, the pressure i n the chamber was increased u n t i l a l l contacts were open, and then released i n order to remove any kinks i n the transducer. The chamber pressure was then increased u n t i l one pair of contacts (or the contacts of interest) opened. The pressure was recorded after i t was varied about the point at which the contact opened. I t was then increased u n t i l the next pair of contacts opened and the process repeated. 5.1.1.7 Equipment to detec: t _JbLood_.fJ.ow past the cuff Blood flow past the cuff was detected with a modified Hoffmann-LaRoche Inc. Arteriosonde 1010 Blood Pressure Instrument. This instrument detects a r t e r i a l wall motion caused by blood flow using ultrasound. The ultrasound transducer was placed with an ultrasonic coupling gel against the skin over the brachial artery d i s t a l to the tourniquet cuff. The voltage output normally provided for a headphone set was r e c t i f i e d with a precision half-wave r e c t i f i e r which i n turn was connected to an 8811A b i o e l e c t r i c amplifier plug-in module and band-limited to 1.5 Hz to 30 Hz. The r e c t i f i c a t i o n was necessary to display the signal on the chart recorder and to reduce the sampling frequency requirements for i t s analog to d i g i t a l conversion. The v a l i d i t y of using ultrasound to detect the onset of flow past the cuff was v e r i f i e d by comparison with impedance plethysmography, oscillometry, photoplethysmography, and a second ultrasound device monitoring the r a d i a l artery at the wrist. These r e s u l t s are reported i n Appendix E. 5.2 D E ^ IN THE WORK Two impedance plethysmographs were b u i l t i n the course of thi s work to investigate the tissue impedance changes at the onset of blood flow beneath a tourniquet cuff. A four-electrode design was used with both of them for the reasons mentioned i n 2.4. The i n i t i a l plethysmograph was b u i l t with the f a c i l i t y to allow evaluation of parameters relevant to impedance plethysmography. Based on results from tests using the i n i t i a l device, a second plethysmograph was b u i l t which was simpler and performed better. The design i s explained and complete schematics are given for both plethysmographs i n Appendix B. 54 5.2.1 S p e c i f i c a t i o n s f o r the initialplethysmograph 5.2.1.1 E1ec t^ode and tissue impedances A series of tests was performed on a single subject (Appendix A) to estimate the magnitude of the basal arm impedance, the p u l s a t i l e change i n arm impedance, the s t a t i c impedance of electrodes, the electrode motion a r t i f a c t , and arm motion a r t i f a c t . 5.2.1.2 Safety requirements for the plethysmograph The plethysmograph was designed to meet the relevant hospital standards. These included leakage current standards, measured as described i n CSA standard C22.2 No.125, Electromedical Equipment [36], and the hospital's grounding resistance standard. 5.2.1.3 Variable parameters used with the i n i t i a l plethysmograph A range of operating parameters and a variety of features for impedance plethysmography are mentioned i n published l i t e r a t u r e . In order to evaluate th e i r merit, the plethysmograph was designed with three possible operating frequencies, f i v e operating currents, and the f a c i l i t y for synchronous demodulation. 55 A range of frequencies between 10 kHz and 250 kHz.is reported i n sim i l a r applications [22,24]. An increase i n frequency o f f e r s the advantage of decreased s e n s i t i v i t y of b i o l o g i c a l tissue. This allows a larger current to be used with a r e s u l t i n g larger signal voltage. The impedance of electrodes also decreases with frequency while the impedance of b i o l o g i c a l tissue l a r g e l y remains unchanged over the frequency range mentioned above. An increase i n frequency also generally r e s u l t s i n a deterioration i n the performance of c i r c u i t components. The common mode rejecti o n , gain, and input impedance of d i f f e r e n t i a l amplifiers tend to decrease with increasing frequency, for example. Similarly, the e f f e c t of stray c i r c u i t capacitances becomes much more prominent and must be accounted f o r . A range of currents from 0.5 mA to 4 mA for plethysmographs used i n similar circumstances i s reported [22,23,24,25]. The advantage of larger currents i s an increase i n the signal l e v e l and hence an increase i n the signal to noise r a t i o . As the current density increases, however, so does the l i k e l i h o o d that the tissue w i l l be stimulated. Large currents could place a lower l i m i t on electrode area. The impedance of b i o l o g i c a l tissue also increases with increased current density. The impedance signal can be resolved into i t s reactive and r e s i s t i v e components by synchronous demodulation. In one study [43], the p u l s a t i l e component of the impedance of the forearm was found to be almost e n t i r e l y r e s i s t i v e . If the 56 impedance change due to motion i s r e f l e c t e d i n both the r e s i s t i v e and reactive components, then possibly the reactive component could be subtracted from the r e s i s t i v e component to remove the motion a r t i f a c t . 5.2.3 Description of the i n i t i a l pietjbj^smjsjgraph The i n i t i a l plethysmograph was b u i l t with a battery operated current source for i s o l a t i o n . Three frequencies, 10 kHz, 31.6 kHz, and 100 kHz, and 5 current l e v e l s , 1 mA, 2.5 mA, 5 mA, 7.5 mA, and 10 mA, could be selected. The plethysmograph provided voltage outputs representative of resistance, reactance, and impedance. Two instrumentation amplifiers and the current source were housed i n a small remote box that could be placed by the subject to minimize the electrode lead length. The remote box was connected to the remaining electronics with a cable. Figure 5.3 shows a block diagram of the plethysmograph (a complete schematic i s included i n Appendix B). The major components are the current source, instrumentation amplifier, high-pass f i l t e r , demodulator, band-pass f i l t e r , and o s c i l l a t o r . 5.2.3.1 Currentsource A sinusoidal voltage at each of the three frequencies i s produced by removing the higher harmonics from a square wave of the desired frequency with a four-pole low-pass f i l t e r . The 57 Instrumentation amplifier and f i l t e r current source remote box # demodulator and band-pass f i l t e r Impedance o s c i l l a t o r F1g. 5.3 Block diagram of the plethysmograph 58 square waves are generated off-board and coupled to the f i l t e r s with high frequency o p t i c a l i s o l a t o r s (General Instrument MCL2630) to preserve the current source i s o l a t i o n . The outputs from the three f i l t e r s are a.c.-coupled to a summing amplifier to remove the d.c. o f f s e t and low frequency noise. The summing amplifier output i s connected to a transconductance amplifier for voltage to current transformation. The current amplitude i s determined by a bank of switchable r e s i s t o r s i n the transconductance amplifier. 5.2.3.2 Instrumentation amplifiers The sense electrodes are a.c.-coupled to a monolithic instrumentation amplifier (Analog Devices AD521K) configured with a potentiometer to adjust the gain. A pair of common mode voltage drivers was connected to guard rings around the input terminals with f a c i l i t y to be connected to shields should shielding and stray lead capacitance compensation be necessary. Two instrumentation amplifiers are housed i n the remote box. 5.2.3.3 High-pass f i l t e r The 1 kHz two-pole high-pass f i l t e r at the output of the instrumentation amplifier removes any low frequency noise p r i o r to demodulation. Any low frequency noise or o f f s e t would be multiplied by the of f s e t or low frequency noise of the mu l t i p l i e r and contaminate the small p u l s a t i l e impedance si g n a l . 59 5.2.3.4 Demodulator Three commercial grade m u l t i p l i e r s (Motorola MC1494L) intended for communication applications were used to synchronously demodulate the impedance s i g n a l . Their outputs contained a large D.C. component corresponding to the s t a t i c basal resistance, reactance, or impedance; a low frequency component corresponding to the time varying part of the resistance, reactance, or impedance; and a large high frequency component due to the c a r r i e r voltage. 5.2.3.5 Band-pass f i l t e r The D.C. component was removed with a single-pole RC high-pass f i l t e r with a corner frequency of 0.03 Hz. The high frequency component was reduced i n amplitude with three single-pole low-pass f i l t e r s i n s e r i e s . Additionally, a 60 Hz n o t c h - f i l t e r removed unwanted l i n e frequency interference. 5.2.3.6 O s c i l l a t o r The o s c i l l a t o r section generated the necessary square waves for the current source. I t also provided the correct demodulation signals to synchronously demodulate the reactance and resistance. ^ 5.2.3.7 Performance 60 The plethysmograph captured the vascular events well. The l i n e a r i t y was within the error of the test procedure (5%) and the current lev e l s were measured to be within 5% of the expected values. The decay waveform displayed on the chart recordings from switching a 4.7 kilohm resistance i n and out of a p a r a l l e l connection with a 4.7 ohm load resistance showed no signs of d i s t o r t i o n . There was, however, excessive low frequency noise o r i g i n a t i n g from the m u l t i p l i e r s (peak amplitude -20 dB of the p u l s a t i l e impedance amplitude i n the frequency range of 1.5 Hz to 30 Hz). To test the v u l n e r a b i l i t y of the plethysmograph to electrode motion, a 10 ohm r e s i s t o r was alternately placed i n series with each of the electrode leads and a switch placed across the r e s i s t o r was r e p e t i t i v e l y opened and closed. The leads were connected to electrodes placed on the upper arm. Ten ohm i s more than the maximum impedance change measured with gross electrode movement. No a r t i f a c t could be noticed i n the background noise when the test was performed. 5.2.4 F i n a l plethysmograph After i n i t i a l tests with the plethysmograph described above, mostly with one subject, a second plethysmograph was b u i l t using an Analog Devices 630K Balanced Demodulator i n place of the Motorola MC1494L M u l t i p l i e r , a fixed operating frequency of 30 kHz, and no synchronous demodulation. The new 61 demodulator reduced the peak electronic noise to less than -32 dB of the p u l s a t i l e impedance signal amplitude i n the 1.5 Hz to 30 Hz bandwidth. 5.2.4.1 Parameter optimization There was no noticeable subjective difference i n signal f i d e l i t y between the three operating frequencies or f i v e current l e v e l s . Because the subject could f e e l the 5 mA and 10 mA current l e v e l s at a 10 kHz operating frequency using the disposable s i l v e r - s i l v e r chloride EKG electrodes, the operating frequency of 30 kHz was chosen for the f i n a l plethysmograph. The f a c i l i t y to change the current l e v e l was kept. The attempt to use the reactance as a noise channel was not successful. The amplitude of motion a r t i f a c t compared to the impedance pulses was about the same for both the resistance and the reactance. Subtraction of the reactance from the resistance only resulted i n a very attenuated version of the o r i g i n a l s i g n a l . Because of t h i s , the synchronous demodulation f a c i l i t y was not retained with the f i n a l plethysmograph. Rather, the demodulator was configured to demodulate the impedance signal, with the output e s s e n t i a l l y that of a precision full-wave r e c t i f i e r . Other changes included: 62 a) r a i s i n g the lower corner frequency of the band-pass f i l t e r to 0.15 Hz to reduce the recovery time of the f i l t e r to an overload; b) removal of the basal impedance buffers; and c) removal of the analog d i f f e r e n t i a t o r . The current source was l e f t unchanged although several improvements were possible. These include: a) the removal of the c i r c u i t r y for the 10 kHz and 100 kHz channels; b) the addition of the 30 kHz o s c i l l a t o r to the current source board rather than generating i t offboard and coupling i t to the current source with an o p t i c a l i s o l a t o r ; c) the use of an LC passive f i l t e r to remove the harmonics from the square wave—a passive f i l t e r was found to reduce the noise i n the output current and also reduced the drain on the batteries; d) the addition of protection from d e f i b r i l l a t i o n and electrosurgery—although unnecessary for laboratory t r i a l s , the protection would be needed for c l i n i c a l use; e) the addition of coupling capacitors to the current source output to prevent the p o s s i b i l i t y of d.c. current passing through the patient due to a component f a i l u r e or excessive o f f s e t voltage; a small d.c. current can cause tissue decomposition beneath the electrodes. 63 5.3 DESCRIPTION OF THE ELECTRODES CONSIDERED FOR USE WITH THE IMPEDANCE PLETHYSMOGRAPH Five electrode types were considered for use i n the study based on t h e i r use i n other studies, use i n the c l i n i c a l environment, and a v a i l a b i l i t y . They were: 1) disposable conductive adhesive EKG electrodes (3M-Littman 2322 Disposable EKG Electrodes, and Lec Tec Corporation SynCor Tracets); 2) s i l v e r - s i l v e r chloride disposable EKG electrodes (Hewlett Packard 14445A, and Marquette 9431-001); 3) disposable band electrodes (Contact Products Ltd. M6001 Electrode Tape); 4) reusable band electrodes (EDI Inc. Reusable Impedance Bands); and 5) conductive hook and loop fastener (Velcro HI-MEG Conductive Products) i n 1.3 cm wide bands. The conductive adhesive electrodes were chosen for studies on the arm, and the reusable band electrodes were chosen for studies on the leg. The c r i t e r i a for t h e i r s e l e c t i o n included the following. 1) P r o f i l e . In order to be mounted under the cuff, a low p r o f i l e was necessary to reduce the r i s k of a concentration of pressure beneath them. 2) Impedance. The impedance at the frequency of the current source must be low. 3) Comfort ^  They should not f e e l uncomfortable when applied beneath an i n f l a t e d cuff. 64 4) E a s e o f application. They should not be d i f f i c u l t to apply i n r e l a t i o n to both the limb and the cuff. 5) A b i l i t y to i s o l a t e the impedance changesnear the o n s e t o f b l o o d f l o w . This c r i t e r i o n i s a measure of the i r a b i l i t y to provide impedance pulses with a good rate of change i n amplitude with cuff pressure near occlusion and also a good signal to noise r a t i o . In general, the sense electrodes tend to emphasize the change i n impedance immediately below them. This was an advantage for the conductive adhesive electrodes for the arm studies. Since the brachial artery l i e s near the surface of the arm and i s usually easy to palpate, a conductive adhesive electrode could be placed accurately i n r e l a t i o n to the artery. Because i t was not as easy to palpate an artery i n the upper leg, the band electrodes were used i n the leg studies. They emphasize the impedance change i n an annular band of tissue beneath the electrodes about the leg. v 65 6. INVESTIGATION OF THE POSSIBLE RELATIONSHIP BETWEEN TISSUE IMPEDANCE AND CUFF PRESSURE 6 .1 INTRODUCTION TO THE EXJP_|RJ_fENT In i n i t i a l experiments using single-bladder cuffs (Appendix I ) , the rel a t i o n s h i p between the amplitude of the p u l s a t i l e impedance and the bladder pressure r e l a t i v e to the occlusion pressure was found to have a large v a r i a b i l i t y between subjects due to an inconsistent tissue pressure p r o f i l e along the skin beneath the cuff. A dual-bladder tourniquet cuff was subsequently adapted from a conventional cuff so that i t could be used on the arm to control the pressure p r o f i l e and produce a consistent r e l a t i o n s h i p between tissue impedance and cuff pressure. The p r o f i l e was controlled by keeping the pressure i n the proximal bladder fixed r e l a t i v e to the pressure i n the d i s t a l bladder. The i n - l i n e longitudinal electrode configuration was used with the sense electrodes cut i n half to measure 1.25 cm by 3.0 cm and placed as shown i n F i g . 6.1. Three experiments were performed to investigate the r e l a t i o n s h i p between the cuff pressure and tissue impedance. In the f i r s t and second experiments (reported i n Appendix H), using 10 and 20 subjects respectively, the bladder pressures were controlled with two Aspen Labs Tourniquet Systems. Sources of error inherent to c o n t r o l l i n g the pressure i n t h i s manner led to the t h i r d experiment i n which the pressures were 66 So...Source electrode Se...Sense electrode Sr...Reference electrode So Sr Sr Se Se (b) Fig;- 6.1 Electrode configurations used with the dual-bladder c u f f experiments. a) Configuration used -with the second and t h i r d experiments b) Configuration used with the f i r s t experiment 67 controlled with rotary pressure regulators driven with stepper motors. Twenty subjects were tested i n t h i s experiment. As well as the means of c o n t r o l l i n g the bladder pressures, a major difference between the experiments was the manner i n which the reference was derived. In the f i r s t experiment, the reference was determined from a separate pair of electrodes, the reference electrodes, i n a way sim i l a r to that i n the single bladder experiments. The reference electrodes were placed along the proximal edge of the tourniquet cuff and the amplitudes of the pulses from them were used to normalize the amplitudes of the corresponding impedance pulses from the sense electrodes placed more d i s t a l l y . In the second and t h i r d experiments, a reference was obtained by keeping the pressure i n the d i s t a l cuff at a value high enough to prevent flow, and then decreasing the pressure i n the proximal cuff u n t i l a maximum i n the impedance pulse amplitude from the sense electrodes was observed. The maximum was used as a reference. The other procedural differences were 1) attaching the electrodes f i r s t to the cuff i n the second and t h i r d experiments instead of d i r e c t l y to the arm , and 2) moving the sense electrodes s l i g h t l y proximally to l i e at the center of the cuff i n the second and t h i r d experiment. 68 The res u l t s from the f i r s t two experiments showed excessive inter-subject v a r i a t i o n i n the point of onset of the impedance pulses. By setting the i n i t i a l cuff bladder pressures i n the t h i r d experiment to produce the desired pressure difference between the peaks i n the tissue pressure p r o f i l e instead of between the bladder pressures, the results became much more consistent. Also, with t h i s improvement and the improvement i n the means of c o n t r o l l i n g the bladder pressures i n the t h i r d experiment, the reference was found to adequately compensate for inter-subject differences. 6.2 SIGNIFICANCE OF THE PRESSURE PROFILE BENEATH A DUAL-BLADDER TOURNIQUET CUFF A t y p i c a l pressure p r o f i l e beneath the dual-bladder cuff with the pressure i n the proximal bladder lower than that i n the d i s t a l bladder i s shown i n F i g . 6.2. The computer model of the impedance sampling f i e l d i s shown i n F i g . 4.4 (c). Before a r t e r i a l blood can enter the e f f e c t i v e impedance s e n s i t i v i t y f i e l d of the electrodes, the pressure i n the proximal cuff must be less than the pressure necessary to collapse the underlying a r t e r i e s . At this point, the pressure i n the d i s t a l bladder, which i s kept a fixed l e v e l above that i n the proximal bladder, i s s u f f i c i e n t to prevent blood flow past the cuff. As the bladder pressures become lower, the p u l s a t i l e volume of blood within the art e r i e s i n the e f f e c t i v e impedance s e n s i t i v i t y f i e l d w i l l increase, and the size of the impedance pulses w i l l increase. Eventually, when the i n t e r s t i t i a l pressure i n the e f f e c t i v e impedance sampling o> 240 X E 200 E. »*. 160 .3 u JC 120 o «J C 4> 80 CO © 40 w 3 */> . i n 07 1— a. 0 A \ 0 2 4 6 8 10 12 14 f Distance Along Arm (cm) \ Proximal end Distal end (shoulder) (elbow) F i g . 6.2 Typical t i s s u e pressure p r o f i l e beneath the dual-bladder tourniquet c u f f . 70 f i e l d becomes lower than the mean blood pressure, the amplitude of the impedance pulses w i l l begin to decrease due to the nonlinear rel a t i o n s h i p between the vessel compliance and translumenal pressure. When the pressure i n the d i s t a l bladder f a l l s below the occlusion pressure, blood w i l l flow past the cuff. The e f f e c t of changing the constant pressure difference between the bladders i s to change the point of onset, the bladder pressure r e l a t i v e to the occlusion pressure at which the impedance pulses begin to increase rapidly i n amplitude. ^ 6.3 DESCRIPTI^ OF THE DUAL-BLADDER TOURNIQUETCUFF The Aspen Labs dual-bladder tourniquet cuff used i n these experiments was shaped to f i t a tapered limb. The cuff contained a r i g i d backing for the bladders to expand against and the edges of the bladders were fixed to the cuff . While each bladder was 6.0 cm wide, the cuff i t s e l f was 14.7 cm wide. The cuff dimensions are shown i n F i g . 6.3. Each bladder had two Luer locking pneumatic connectors. As with the other cuffs, the dual-bladder cuff was secured around the limb with hook and loop fasteners. 6.4 EXPERIMENTAL PROCEDURE The electrodes were attached to the cuff before the cuff was applied thereby assuring proper alignment between the cuff and electrodes. The tissue pressure transducer was also attached to the cuff and placed just above the electrodes such Luer pneumatic connectors 60 cm -hook straps F i g . 6.3 Dual-bladder tourniquet cuff s t r i p s that the f i r s t and l a s t switches of the transducer were aligned with the centers of the two tourniquet cuff bladders. The subject was seated with her l e f t forearm re s t i n g comfortably on a desk. At the midpoint of the l e f t upper arm, the brachial artery was located by palpation and the position marked. The cuff was then aligned to the arm such that the marked a r t e r i a l p o s i t i o n was midway between the sense electrodes. After the cuff was pressed against the arm to adhere the electrodes to the arm, i t was wrapped around the arm and tightened enough to allow one finger to be slipped beneath the proximal edge of the cuff but not three. Coupling gel was applied to the face of the ultrasound sensor and the sensor was placed over the brachial artery d i s t a l to the cuff, secured with micropore tape, and connected to the Arteriosonde. The electrode leads were then attached to the leads from the remote box housing the current source and instrumentation amplifiers and the equipment turned on. The proper operation of the plethysmograph was v e r i f i e d by looking at the input to the demodulators with an oscilloscope. r The cuff pressures were controlled with stepping motors under computer control. The control software algorithm i s i l l u s t r a t e d i n F i g . 6.4. To adjust the pressures, the program c o n t r o l l i n g the bladder pressures was started causing the d i s t a l and proximal cuff bladders to i n f l a t e to 240 mmHg and 180 mmHg respectively. At thi s point, the d i s t a l peak i n the pressure p r o f i l e was manually measured with the tissue pressure transducer and associated instrumentation, and the adjust proximal- and distal-bladder pressures to 240 mmHg and 180 mmHg respectively ramp down both ramp down proximal-bladder pressures bladder pressure sample impedance, blood Clow, sample impedance, and and bladder pressures proximal-bladder pressure Q stop ^ Fig. 6.4 Algorithm for adjusting cuff pressures 74 required compensation i n the d i s t a l cuff pressure entered into the computer from the console to bring the d i s t a l peak pressure to within 2 mmHg of 220 mmHg. The process of measuring the d i s t a l peak pressure and adjusting the d i s t a l cuff pressure was repeated u n t i l the d i s t a l peak pressure was within the desired accuracy. The procedure was then repeated to bring the proximal peak of the pressure p r o f i l e to within 2 mmHg of 160 mmHg by adjusting the pressure i n the proximal bladder. These adjustments were then repeated i f necessary u n t i l both p r o f i l e peaks were within 2 mmHg of the desired values. The bladder pressures were then decreased 100 mmHg over 80 s with the impedance and blood flow signal sampled at 90 Hz and recorded d i g i t a l l y , and the two bladder pressures sampled at 7.2 Hz and recorded d i g i t a l l y (Fig. 6.5). To record the data necessary to establish a reference, the process of i n f l a t i n g the bladders to i n i t i a l pressures and then adjusting the pressures to produce the desired peak tissue pressures was repeated. Instead of then decreasing both bladder pressures with time, only the proximal bladder pressure was decreased. The proximal bladder pressure and the impedance signal were sampled and recorded d i g i t a l l y . After removing the cuff, the arm was examined for injury. Also, the imprint of the cuff l e f t on the arm was checked for proper alignment of the cuff and electrodes to the marked position of the artery. 75 llillllllililllli, | Increasing impedance Time (sec) F i g . 6.5 Variation of the bladder pressures with time (a) and the resulting Impedance signal (b) and blood flow signal ( c ) . 76 6.5 ANALYSIS OF^EpkTk FROM THE DUAL-BLADDER TpURNIQUET CUFF EXPERIMENT The data from the experiment was f i r s t f i l t e r e d using an i n t e g e r - c o e f f i c i e n t low-pass d i g i t a l f i l t e r with i t s f i r s t zero at 7.5 Hz. A program then measured the amplitude of both the r i s i n g edges and f a l l i n g edges of the peaks within the 80 s epochs using the same algorithms mentioned i n Appendix H but with an adaptive threshold. The maximum sample was found i n the f i r s t 450 samples and 1/2 of t h i s value was used as the i n i t i a l threshold. The programs then checked to make sure at least 3 peaks could be found i n the f i r s t 450 samples with t h i s threshold. If not, the threshold was decreased by a factor of two and the f i r s t 450 samples were again checked for 3 peaks. The process was repeated u n t i l 3 peaks were found. The program would scan the samples i n the epoch sequentially u n t i l the threshold was exceeded. The scan continued and the maximum value encountered was recorded u n t i l the sample values again f e l l below the threshold. The maximum value recorded was an impedance pulse peak. After then finding the bases of the r i s i n g and f a l l i n g edges, the pulse amplitude was calculated for both the r i s i n g and f a l l i n g edges by subtracting the value of the samples at the bases from that of the peak. Instead of continuing the scan using the same threshold to f i n d the next peak, the threshold became 1/2 the median value of the l a s t three peak amplitudes. Also, i f no peaks were found within 60 samples (2/3 s ) , the threshold was decreased by a factor of two. The envelope of peak amplitudes was displayed along with 77 the f i l t e r e d impedance data using the Lotus 123 program (Release 2, Lotus Development Corporation, Cambridge, Ma) to v i s u a l l y check the accuracy of the process. The pulse amplitudes and corresponding r e l a t i v e cuff pressures were then stored i n a f i l e for processing by Lotus 123. Lotus 123 was used to do a l i n e a r regression from the pressure at which the maximum pulse amplitude occurred to a pressure 60 mmHg higher (Fig. 6.6). The standard error of estimate [44] for each of the l i n e a r regressions was multiplied by the absolute value of the slope of the regression l i n e to get a pressure equivalent to the standard error of the estimate. An average regression l i n e for a l l the subjects was determined from the data that used the amplitude of the f a l l i n g edge of the impedance pulses. F i g . 6.7 shows the average l i n e with the pulse . amplitude now the independent variable and two, means of normalization. Also shown are the l o c i of points displaced from the average l i n e by one standard error of the cuff pressure estimate. The occlusion pressure was found from the blood flow samples. The appearance of the f i r s t pulse from the Arteriosonde was found using a threshold value of the maximum value divided by 10. The accuracy was checked v i s u a l l y using the Lotus 123 program. The nearest pulse was then found from the array of pulse heights and addresses and the d i s t a l cuff pressure at that address determined from the stored f i l e of the d i s t a l cuff pressure samples. 78 Relative Cuff Pressure (mm Hg) F i g . 6.6 Pressure versus pulse amplitude data f o r one subject showing the l i n e a r r e g r e s s i o n l i n e . 79 90 F i g . 6.7 R e l a t i o n s h i p between the normalized impedance p u l s e amplitude and the r e l a t i v e c u f f p r e s s u r e . A l s o shown are the l o d of points d i s p l a c e d by one standard e r r o r o f e s t i m a t i o n 80 The maximum amplitude of the reference impedance pulses was found from the impedance data obtained when the pressure i n the proximal bladder was ramped down 100 mmHg over 80 s while the pressure i n the d i s t a l bladder was kept constant. The f i l e was sequentially scanned for a maximum after a 9-point running average was applied to the data. The maximum was used as a reference to normalize the impedance pulse amplitudes. 6.6 RESULTS Op ^ JHE DUAL BLADDER-CUFF J^PJBRJiMENT As mentioned above, a pressure equivalent to the standard error of estimation was calculated for the l i n e a r regression of impedance pulse amplitude versus the r e l a t i v e cuff pressure (Fig. 6.6). For the r i s i n g edge amplitude, the pressure equivalent was 6.5 mmHg averaged over the 20 subjects tested with a maximum of 25.8 mmHg. For the f a l l i n g edge amplitude, i t was 5.2 mmHg averaged over the subjects with a maximum of 14.1 mmHg. As well as using as a reference the maximum impedance pulse amplitude from the data generated when only the proximal cuff pressure was ramped down with time, the maximum impedance pulse amplitude found when both bladder pressures were ramped down was also used. These situations are referred to respectively as normalization with the reference and normalization with s e l f . An average regression l i n e was determined for for a l l the subjects from the data that used the amplitude of the f a l l i n g edge of the impedance pulses for both normalization with the reference and with s e l f . The f a l l i n g edge data had a larger amplitude than the corresponding r i s i n g edge data yet had a smaller standard error of estimation. F i g . 6.7 shows these average l i n e s with the pulse amplitude now the independent variable. Also shown are the l o c i of points displaced from the average l i n e by one standard error of the "cuff pressure estimate. 6-7 SOURCES OF ERRORIN THE DUAL-BLADDERTOURNIQUETCUFF EXPERIMENT Several sources of error existed i n the t h i r d experiment which affected the accuracy of the rel a t i o n s h i p between the bladder pressure and the tissue impedance and are mentioned below. 6.7.1 Regulator nonlinearity Because one of the regulators has a nonlinearity, the actual difference between cuff pressures varied by ±3 mmHg. 6.7.2 Limited resolution of the tissue pre^su?!.® transducer The pressure difference between the peaks i n the pressure p r o f i l e was determined with the tissue pressure sensor. Unless 82 the discrete points on the transducer at which the pressure . can be determined l a i d exactly at the points of maximum pressure, an error arose when the cuff pressures were i n i t i a l l y set. The rate at which the pressure f a l l s off with distance along the arm can be large, and a 0.5 cm of f s e t of the transducer could cause an error as large as 17 mmHg. 6.7.3 Variation i n t h e difference betweenpeaks_in the... tissue pressure p r o f i l e It has also been assumed that for a l l subjects the i n i t i a l pressure difference i n the p r o f i l e peaks did not change as the bladder pressures are ramped down i n unison. Although the v a r i a t i o n i n the difference was found to be small (2 mmHg) when tested using a single subject, i t i s possible i t may have changed s i g n i f i c a n t with other subjects. 6.7.4 Error i n the i n i t i a l bladder pressures When the i n i t i a l bladder pressures were adjusted to produce the desired maximums i n the pressure p r o f i l e , an error of ± 2 mmHg was accepted for each maximum. This could cause a combined error of ± 4 mmHg i n the pressure settings. 6.7.5 Time taken f o r t h e e x p e r i m e n t Another source of error would be the time taken to complete the experiment which was about 5 min. Although the time i s much shorter than i n the previous dual- and sing l e -83 bladder cuff experiments, i t i s possible that some parameter which affects the tissue impedance versus cuff pressure re l a t i o n s h i p could change s i g n i f i c a n t l y . 6.7.6 Presence of impedance pulses at r e l a t i v e l y high cuff pressure g Another source of error i s the presence of impedance pulses at r e l a t i v e l y high cuff pressures (more than 80 mmHg higher than the occlusion pressure) which i s probably caused by either cuff motion from a r t e r i a l expansion at the proximal edge of the cuff or a larger e f f e c t i v e impedance s e n s i t i v i t y f i e l d than expected. In either case, the eff e c t i s to add an of f s e t to the amplitude of the impedance pulses which w i l l cause error i n the impedance versus cuff pressure re l a t i o n s h i p . The e f f e c t of both of the possible causes i s dependent on the electrode configuration. 6.7.7 Inter-subject v a r i a t i o n i n t h e d i f f e r e n c e between mean and s y s t o l i c b l o o d pressure The reference used to normalize the impedance data was found by maintaining the pressure i n the d i s t a l bladder at a value s u f f i c i e n t to prevent blood flow past the cuff while the pressure i n the proximal cuff was gradually decreased. The maximum amplitude of the impedance pulses during t h i s period was used as the reference. Other studies have found a similar r e l a t i o n s h i p between the sensed p u l s a t i l e volume of blood i n a limb segment and the pressure i n a surrounding cuff. The 84 tissue pressure at which there i s a sudden increase i n the amplitude of the p u l s a t i l e blood volume as the cuff pressure i s dropped towards the occlusion pressure has been considered i n d i c a t i v e of the subject's s y s t o l i c pressure. The maximum amplitude experienced as the cuff pressure was further reduced i s i n d i c a t i v e of the mean pressure. If th i s r e l a t i o n holds with the impedance pulses sensed as the pressure i n the proximal bladder i s reduced, then the impedance pulse amplitudes are normalized with the impedance value at the subject's mean pressure. Because the difference between the mean pressure and s y s t o l i c pressure i s variable, normalization with the maximum pulse height w i l l r e s u l t i n a v a r i a t i o n between subjects i n the slope of the relat i o n s h i p between the normalized pulse height and the r e l a t i v e cuff pressure. 85 7. APPROACH TO ADAPTIVE CONTROL OF THE TOURNIQUET CUFF PRESSURE 7.1 INTRODUCTIONTOCONTROL To meet one of the thesis objectives, closed loop control of the pressure i n a tourniquet cuff was implemented based on an algorithm using impedance plethysmography. The successful implementation helped to esta b l i s h the f e a s i b i l i t y of integrating impedance plethysmography into a p r a c t i c a l tourniquet system to minimize the cuff pressure used to arrest blood flow and thereby reduce the r i s k of injury from the pressure i n the cuff. In the implementation, the bladder pressures i n a dual-bladder tourniquet cuff placed on the arm were regulated to maintain a constant impedance pulse height, using both the procedure and the rel a t i o n s h i p between the impedance pulse amplitude and the r e l a t i v e bladder pressure established i n Chapter 6. The tests of the control of the bladder pressures were done with only one subject and using the i n - l i n e longitudinal electrode configuration. The software previously used to measure the pulse amplitude, establish a reference, and control the bladder pressures with the stepping motors was adapted to real-time control. 7.2 CHARACTERIZATION OF THE SYSTEM 86 Because the impedance pulses occur only once for every heart beat, the pulse amplitude i s e f f e c t i v e l y sampled at the subject's heart rate and the system can be considered a sampled system (Fig. 7.1) with a quasi-periodic sampling rate (the subject's heart r a t e ) . As shown i n F i g . 7.2, both the time needed to change the bladder pressures and the response time of the impedance pulse amplitude to changes i n cuff pressure are less than the e f f e c t i v e sampling rate. The rate of change of the pressures of the bladders i s lim i t e d by the speed of the stepping motors and the rate at which a i r can be supplied to and taken from the bladders by the regulators. In F i g . 7.2, the rise-time of the pressure pulse i s limited by the speed of the stepping motor, and the f a l l - t i m e by the rate at which a i r can be removed from the bladder. It was necessary to l i m i t the motor speed to reduce the mechanical demands put on the regulators, which were intended only to be manually adjusted, and on the coupling devices between the motors and regulators. To arrive at the stepping motor speed used i n the tests, the motor speed was increased u n t i l a noticeable lag between the movement of the regulator and the cuff pressure was evident. The speed was then decreased marginally. disturbance variables Thr G c u f f \ pressure —*~ Pt tissue pressure r e l a t i v e to occlusion Thr H P t . . . t i s s u e pressure r e l a t i v e to the t i s s u e o c c l u s i o n pressure T h r . . . p e r i o d between the q u a s i - p e r i o d i c heart beats K...the r e l a t i o n s h i p between the t i s s u e pressure r e l a t i v e to the t i s s u e o c c l u s i o n pressure and the measured, f i l t e r e d , and normalized p u l s a t i l e impedance Set...the p r o p o r t i o n of the r e f e r e n c e amplitude which would give the d e s i r e d d i f f e r e n c e between the t i s s u e pressure and the t i s s u e o c c l u s i o n pressure G...the t r a n s f e r f u n c t i o n of the product of the c o n t r o l a l g o r ithm, a sample and hold, and r e a l i z a t i o n of a pressure change i n the c u f f from a software command F i g . 7.1 Sampled c o n t r o l system. 7.3 DESCRIPTION OF THE CONTROL ALGORITHM 88 The basic algorithm involved continually searching for impedance pulse peaks. When one was found, i t s amplitude was measured and compared to the set point which was one-half the reference amplitude. The difference between the two, the error, was used to calculate new bladder pressures according to: P(n+l)=E(n)*F/K where n i s the current heart rate sample number; P(n+1) i s the bladder pressure that w i l l be applied u n t i l the next heart rate sample; E(n) i s the current calculated value of the error; F i s the conversion factor between pulse amplitude and cuff pressure taken from the slope of the average regression l i n e of the pulse amplitude versus r e l a t i v e cuff pressure re l a t i o n s h i p ; and K i s an integer d i v i s o r used to adjust the gain of the control loop. 7.4 PROC^ THE FEASI^ THE CONTROL A J L t G O R J_ nIM The same protocol was used for applying the electrodes and cuff, and adjusting the gains of the b i o e l e c t r i c amplifiers as that used i n the experiment i n Chapter 6 . The electrodes were applied to the cuff which was then placed 89 around the arm such that the electrodes were properly aligned with the brachial artery. The amplifier gains were adjusted with the d i s t a l bladder i n f l a t e d to approximately the occlusion pressure. With a l l the tests, a 60 mmHg pressure difference between the peaks i n the pressure p r o f i l e was again used. I n i t i a l l y , the tests were done with the subject seated with hi s arm resting on a desk. To invoke a disturbance to examine the response of the control system, the legs were raised to trans i e n t l y change the subject's blood pressure. The subject was supine for these t e s t s . After the required equipment adjustments, the control program was started, i n i t i a t i n g the following sequence of events. 1) After prompting from the control program, the pressure transducers were manually exposed to atmospheric pressure and the b i o e l e c t r i c amplifier outputs switched off to establ i s h zero baseline values for use by the program (point 1 on Fi g . 7.3). Again after prompting, the transducers and amplifiers were returned to thei r operational mode (point 2). 2) Under program control, the d i s t a l and proximal bladders were i n f l a t e d to 240 mmHg and 180 mmHg respectively (point 3). 91 3) The peaks i n the pressure p r o f i l e along the arm beneath the cuff were measured manually and the bladder pressures adjusted to bring them to within + 2 mmHg of 220 mmHg and 160mmHg respectively by entering estimations of the required bladder pressures on the computer console (from point 3 to point 4). 4) The proximal bladder pressure was then decreased by 100 mmHg i n 25 seconds (point 4 to point 5). The maximum impedance pulse amplitude encountered during t h i s period became the reference. 5) The proximal cuff pressure was then increased to i t s value p r i o r to event (4) (point 6). 6) The program then began control of the bladder pressures using the impedance pulse amplitude and the control algorithm (from point 6 to point 7). The legs were raised at point x. 7) Control was suspended'after 72 seconds and the bladders l e f t i n f l a t e d for 40 seconds (point 7 to point 8) . This period without control provided a means of comparison for the previous period i n which control was used. 8) The bladder pressures were then decreased 60 mmHg i n 40 seconds (point 8 to point 9). The slow decrease i n pressures at th i s point was to establish the occlusion 92 pressure, the pressure at which the Arteriosonde indicated that blood flow past the cuff had resumed. 7.5 RESULTS OF THE DEMONSTRATION OF THEFEASIBILITY 0 F THE CONTROL From F i g . 7.4, the response of the control system to the i n i t i a l error step when the program switched on the control algorithm shows the system to be underdamped. When a three-point median f i l t e r or a running average was used to smooth the impedance pulse data, the response actually became unstable. Because of the i n s t a b i l i t y , no further attempt was made to remove noise from the impedance pulse amplitude data i n t h i s i n i t i a l control algorithm. With the slow ramp at the end of the period without control (point 8 to point 9 i n F i g . 7.3), the occlusion pressure was found to be 50 mmHg below the average pressure during the period with control. From the transfer function regression l i n e , a set point of 0.5 would r e s u l t i n a pressure of 30 mmHg above occlusion with a standard error of estimation of 12 mmHg. Raising the legs forces much of the venous blood pooled i n the legs into the general c i r c u l a t i o n and temporarily changes the blood pressure. The ef f e c t of thi s on the impedance pulses without control i s indicated by point 1 i n Fig . 7.4 whereas with control, the e f f e c t i s reduced to that shown at point 2. e6 94 The subject used i n these tests exhibited a large vascular o s c i l l a t i o n a r t i f a c t (Fig. 7.3). Without removing the a r t i f a c t from the impedance pulses, the ef f e c t i s to modulate the cuff pressure i n an attempt to correct for the o s c i l l a t i o n s . This e f f e c t i s shown i n F i g . 7.3 i n the period with and without control. 95 8. CHARACTERIZATION AND HANDLING OF THE SIGNIFICANT NOISE AND ARTIFACT 8.1 OVERVIEW To eventually achieve the goal of integrating impedance plethysmography into a p r a c t i c a l c l i n i c a l tourniquet system, the c l i n i c a l l y s i g n i f i c a n t sources of noise and a r t i f a c t must be r e a l i s t i c a l l y and e f f e c t i v e l y handled. The o r i g i n and c h a r a c t e r i s t i c s of such a r t i f a c t and noise encountered during th i s study are discussed i n 3.4 and Appendix F. In general, the disturbances that af f e c t the e f f i c i e n c y with which the cuff pressure i s transmitted to the underlying tissue, or a f f e c t the magnitude of the i n t e r s t i t i a l pressure necessary to collapse the a r t e r i e s i n the tissue, w i l l be reduced i n the proposed system employing feedback. Disturbances that a f f e c t the normalized amplitude of the measured impedance pulses for a constant r e l a t i v e cuff pressure, however, can be a source of noise and a r t i f a c t because they are not attenuated and, therefore, must be dealt with i n other ways. Because the energy of the p u l s a t i l e impedance signal i s concentrated within i t s f i r s t f i v e harmonics, f i l t e r i n g i s e f f e c t i v e i n removing any interference with most of i t s energy outside t h i s narrow band. This would include a r t i f a c t and noise from motion, res p i r a t i o n , vibration, and the electronic instrumentation. 96 Much of the other noise experienced during t h i s study was of short duration and appeared much l i k e an impulse without any c o r r e l a t i o n to the p u l s a t i l e impedance s i g n a l . The source for such noise can include muscle spasm, alternate current pathways, and band electrode motion. By referencing the time of occurrence of the detected peaks i n the impedance signal to the subject's electrocardiogram (ECG), the number of detected peaks due to impulse noise could be reduced s u b s t a n t i a l l y . The patient's cardiovascular status i s normally monitored with an electrocardiograph during an operation and the signal i s r e a d i l y available. For example, any peaks detected that did not occur from 150 to 300 ms following the main peak of the the subject's ECG could be considered i n v a l i d . Since the time between heart beats i s t y p i c a l l y 830 ms, impulse noise not correlated with the subject's heart rate w i l l stand a good chance of being rejected. A running median of the peak amplitudes could also remove the e f f e c t of noise of a short duration provided the noise occurs at a rate much lower than the subject's heart rate. Such noise would appear as an impulse-like contamination of the impedance pulse amplitudes and median f i l t e r i n g i s e f f e c t i v e i n removing impulse noise. The a r t i f a c t from PVC's can s i m i l a r l y be removed by using the median since PVC's change the amplitude of one of two consecutive impedance peaks. Median f i l t e r i n g would decrease the response time of the control system. 97 8.2 CHARACTERISTICS AND HjANpLJNG OF MOTION ARTIFACT Motion a r t i f a c t , the major source of interference deserving of spe c i a l attention, was found to be a low frequency phenomenon and high-pass f i l t e r i n g was e f f e c t i v e i n reducing the energy of the disturbing s i g n a l . Due to the magnitude of the a r t i f a c t , which could be 20 times the desired impedance signal, f i l t e r i n g alone was not always s u f f i c i e n t to remove the interference. Two attempts were made to f i n d a noise channel that r e f l e c t e d the a r t i f a c t from motion while having l i t t l e contribution from the desired p u l s a t i l e impedance s i g n a l . A portion of the signal from the noise channel was subtracted from the p u l s a t i l e impedance signal contaminated with motion a r t i f a c t to try to remove the a r t i f a c t . Neither attempt was successful. In the f i r s t attempt, s i l v e r - s i l v e r chloride disposable EKG electrodes were used i n an i n - l i n e longitudinal configuration beneath a tourniquet cuff with an additional set of electrodes, the motion electrodes, placed d i s t a l l y to the sense electrodes. With s u f f i c i e n t pressure i n the cuff to occlude blood flow, vascular a c t i v i t y was not evident i n the impedance signal from the motion electrodes but was i n the signal from the sense electrodes. The signal from both sets of electrodes did respond to motion. Subtraction of a portion of the motion electrodes output from the sense electrodes output did not remove the motion a r t i f a c t . When the proportion was adjusted to minimize a r t i f a c t from a p a r t i c u l a r motion, 98 a r t i f a c t from other motion was not removed, and i n some cases, increased r e l a t i v e to the impedance pulse amplitude. When located d i s t a l l y to the sense electrodes, the motion electrodes may not give a signal that adequately r e f l e c t s the c h a r a c t e r i s t i c s of the equivalent signal from the sense electrodes. Supination of the forearm, for example, was found to decrease the circumference of the upper arm near the elbow where the motion electrodes were located while increasing i t at the middle of the upper arm where the sense electrodes where located. Hence the impedance signal from the motion electrodes would increase while that from the sense would decrease. For f l e x i o n at the elbow, though, both outputs would decrease because the arm circumference at middle and near the elbow increased. By changing the configuration of the sense and motion electrodes, i t may be possible to have the desired response to vascular events from the sense electrodes as well as a good representation from the motion electrodes of the motion a r t i f a c t found i n the sense electrode output. Also, an adaptive algorithm, one better than simple subtraction, may be less affected by the the phase and amplitude differences between the sense and motion electrodes outputs for d i f f e r e n t motions. In the second attempt, the signal from the sense electrodes, again i n an i n - l i n e longitudinal configuration, was synchronously demodulated into i t s r e s i s t i v e and reactive components. Since blood i s primarily r e s i s t i v e at the frequency used by the plethysmograph, the r e l a t i v e 99 contribution to an impedance change by the r e s i s t i v e and reactive components may be quite d i f f e r e n t for a change caused by motion and one caused by an i n f l u x of blood. If true, subtracting a portion of one component from the other may reduce the amount of signal i n the difference due to motion r e l a t i v e to that due to vascular a c t i v i t y . However, when the portion of the reactance signal subtracted from the resistance channel was adjusted to minimize the motion a r t i f a c t , the magnitude of the impedance signal i n the difference due to vascular a c t i v i t y was s i m i l a r l y attenuated. The admittance signal was s i m i l a r l y demodulated into i t s conductance and susceptance to investigate the r e l a t i v e contribution of motion and vascular a c t i v i t y to each of the admittance components. To derive the admittance s i g n a l , the plethysmograph was modified to hold the voltage at the sense electrodes constant while a voltage representative of the re s u l t i n g current was taken to the input of the demodulators. With the voltage across the section of tissue held constant, the current becomes representative of the admittance of that section of tissue. As with impedance, however, the subtraction of a portion of the susceptance from the conductance to minimize the motion a r t i f a c t also attenuated the desired signal by a s i m i l a r amount. The i n - l i n e longitudinal electrode configuration used i n a l l the dual-bladder cuff experiments i s not the configuration most immune to motion a r t i f a c t . I t was used primarily because more than a single set of electrodes could e a s i l y be placed 100 between a single set of source electrodes. For the purpose of minimizing interference due to motion, other configurations are better. The i n - l i n e transverse configuration, for example, was one with much less s e n s i t i v i t y to motion. One other approach which could be used to reduce the e f f e c t of motion a r t i f a c t i n determining a r e l a t i v e cuff pressure from the measured impedance pulse amplitude, i s to remove any data used i n the determination of the pulse amplitudes that i s corrupted with motion a r t i f a c t . Excessive noise could be detected and processing suspended u n t i l the l e v e l of noise f a l l s within desired accuracy l i m i t s . 101 9. RESULTS AND DISCUSSION 9.1 RESULTS As noted i n 1.2, the thesis objectives were to: 1) investigate the use of impedance plethysmography to predict the onset of blood flow i n a limb once the flow has been arrested by an occlusive cuff; 2) design, implement, and evaluate equipment for th i s purpose; 3) account for the expected a r t i f a c t from subject motion; and 4) implement closed loop control of the pressure i n a pneumatic tourniquet cuff using an algorithm based on impedance plethysmography. The major r e s u l t s and contributions of the research i n t h i s thesis can be summarized as follows: 1) the development, design, implementation, and evaluation of an impedance plethysmograph with i t s performance optimized to measure the changes i n the impedance of the tissue beneath an occlusive cuff; 2) the development of a computer model into a useful tool for predicting the performance of electrode configurations used with impedance plethysmography, and the v a l i d a t i o n of the model through experimentation; 102 3) the discovery of a quantitative relationship between the pressure i n a tourniquet cuff and the impedance of the underlying tissue which can be used to predict the nearness of the onset of blood flow; 4) the i d e n t i f i c a t i o n and p r i o r i t i z a t i o n of the sources of a r t i f a c t and interference associated with the use of impedance plethysmography i n the context of this thesis together with the development of approaches for detecting and handling the most s i g n i f i c a n t types of a r t i f a c t ; 5) the measurement of the tissue pressure p r o f i l e along the surface of the arm beneath various occlusive cuffs which, to the author's knowledge, had not been done before; and 6) the establishment of the f e a s i b i l i t y of c o n t r o l l i n g the pressure i n a tourniquet cuff on the basis of changes i n the e l e c t r i c a l impedance of the underlying tissue. Many of the key r e s u l t s with respect to item (3) were described i n a paper submitted for publication i n the IEEE Transactions on Biomedical Engineering. Progress made towards achievement of the objectives i s discussed below and the detailed experimental results summarized. 103 9.2 THEJU^EOF IMPEPM fMTM3W9M?3I TO PREDICT THE ONSET OFBLOODFLOW 9.2.1 The relati o n s h i p M.eB.tiI.i..e.|d. be_tween..._the tourniquet c u f f p r e s s u r e a n d theimpedanceof the underlying tissue A major contribution of the research was to achieve the f i r s t objective of the thesis and f i n d the r e l a t i o n s h i p shown i n F i g . 6.7 between the amplitude of the p u l s a t i l e component of the tissue impedance and the cuff pressure r e l a t i v e to the occlusion pressure. I t i s the average of the l i n e a r regressions of the p u l s a t i l e impedance amplitude versus r e l a t i v e cuff pressure data from 18 subjects with the pulse amplitude shown as the independent variable. The data has been normalized with a reference as mentioned below. Also shown are the l o c i of points displaced by one standard error of estimation from the average. Although the error d i s t r i b u t i o n about the average l i n e was not normal, i t was centered about an error of 0.0. The standard error of estimation from the data for each of the 18 subjects was calculated as well. An average of the pressure equivalent of these was 5.2 mmHg with a maximum of 14.1 mmHg. To f i n d the relationship, i t was necessary to control the pressure p r o f i l e along the arm. If the pressure p r o f i l e was not uniform between subjects, the distance of the incursions of a r t e r i a l blood along the arm beyond the proximal edge of the cuff with a change i n cuff pressure was not predictable. 104 Since the electrodes and hence the e f f e c t i v e impedance s e n s i t i v i t y f i e l d were fixed r e l a t i v e to the cuff, the measured change i n impedance would not be well correlated with the cuff pressure. To control the p r o f i l e , a dual-bladder tourniquet cuff was used with the bladder pressures adjusted i n i t i a l l y to make the peak tissue pressure beneath the d i s t a l bladder (220 mmHg) 60 mmHg greater than the peak tissue pressure beneath the proximal bladder (160 mmHg). The peak tissue pressures were measured with a tissue pressure transducer recently designed and developed i n the Biomedical Engineering Department. Once the 60 mmHg difference was established, the bladder pressures were adjusted i n unison, maintaining the i n i t i a l difference between them. It was necessary to set the bladder pressures by measuring the peaks i n the pressure p r o f i l e with the tissue pressure transducer because the difference i n the peak pressures was found to vary considerably with a set 60 mmHg difference between the bladder pressures. For 20 subjects, the difference ranged from 20 mmHg to 90 mmHg. These extremes i n the difference were not expected. Only by measuring the tissue pressure p r o f i l e s , another s i g n i f i c a n t contribution of the research, was i t possible to i d e n t i f y the v a r i a b i l i t y i n the difference as a cause of the inconsistency i n the relat i o n s h i p between tissue impedance and cuff pressure among subjects. The subsequent additional lab t r i a l did necessarily extended the length of the study. 105 Using a dual-bladder tourniquet cuff was a convenient way to control the pressure p r o f i l e . Because the peak tissue pressure from the dual bladder cuff i s more d i s t a l than the peak tissue pressure beneath a single bladder cuff, the pressure gradient i n the tissue beneath the dual-bladder cuff may be greater. The pressure gradient may be a s i g n i f i c a n t factor i n causing tissue damage and consideration should be given to other possible p r o f i l e s . Both the signal from which the reference was derived and the impedance signal used i n the rel a t i o n s h i p shown i n F i g . 6.7, came from the same pair of electrodes. By using the same electrodes, there are no differences between the signals due to electrode dependent factors such as the position of the electrodes on the arm, placement of one electrode r e l a t i v e to the other, and electrode dimensions. Because the reference i s expected to change slowly with time, the control of the cuff pressure based on the impedance signal would have to be interrupted occasionally to remeasure the reference. A more desirable scheme would be to derive the reference continuously through the same set of electrodes. This was t r i e d unsuccessfully with the transverse-offset electrode configuration using a single bladder cuff. With the dual-bladder cuff and a controlled pressure p r o f i l e , t h i s technique may be f e a s i b l e . By using the computer model of the arm to evaluate prospective electrode configurations, another contribution of the thesis, i t may be possible to i d e n t i f y other promising configurations. 106 The population sample i n thi s study consisted of 18 subjects, 10 male and 8 female. The ages ranged from 20 to 38 years and the arm circumferences from 24.1 cm to 35.5 cm. None of the subjects were obese and none had any known cardiovascular problems. In the future, the population sample w i l l have to be extended to include subjects with c h a r a c t e r i s t i c s outside the ranges just mentioned. 9.3 THE NOVELPLETHYSMOGRAPHY INSTRUMENTFORADAPTIVE TCURNXfiUETSYSTEMS Another contribution of the study was the development of the plethysmograph to measure the p u l s a t i l e tissue impedance i n a limb using a four-electrode technique. The main requirements were to capture the vascular events with a resolution of 28 dB (25:1) and to re j e c t a r t i f a c t from a variety of sources, primarily electrode movement and alternate current pathways. The peak instrumentation noise was -32 dB r e l a t i v e to the amplitude of the impedance pulses measured with an i n - l i n e l ongitudinal electrode arrangement, no hyperemic response, and a bandwidth of 1.5 Hz to 30 Hz. This noise was much less than the background noise normally present due to vib r a t i o n , muscle tremor, alternate current pathways, and other possible sources (Fig. 9.1(b) ) 107 (a) Impedance p u l s e s ; Bandwidth 1.5 Hz to 30 Hz (b) The same impedance s i g n a l as i n (a) with the c u f f bladder pressure high enough to suppress the impedance pulses F i g . 9.1 Test f o r e l e c t r o d e a r t i f a c t . A 10 ohm r e s i s t o r was switched i n (0) and out (C) of a s e r i e s connection with one of the e l e c t r o d e leads. The two r e c o r d i n g s have the same s c a l e . 108 The l i n e a r i t y was within the accuracy of the method used to test i t (5%). No d i s t o r t i o n was noticeable i n the exponential decay of the step input r e s u l t i n g from switching a 4.7 kQ c a l i b r a t i o n r e s i s t o r i n and out of a p a r a l l e l connection with a 4.7 0 r e s i s t o r connected to the current source and instrumentation amplifier. The leakage current from the plethysmograph and chart recorder was within the CSA requirements for a r i s k class 2 device and the chassis grounding resistance was less than the hospital's standard of 0.1 Q. The s e n s i t i v i t y to electrode movement was tested by switching a 10 ohm r e s i s t o r i n series with each of the electrodes connected to the subject i n an i n - l i n e longitudinal configuration. Ten ohm i s larger than the measured change i n electrode impedance due to gross electrode movement. No a r t i f a c t could be seen i n the background noise that was present due to the sources mentioned above (Fig. 9.1). In a subjective test, the plethysmograph worked as well with the conductive adhesive removed from the electrodes, which would increase the electrode impedance su b s t a n t i a l l y . 9.4 CHARACTERIZATION AND HANDLING OF MOTION ARTIFACT Because the energy i n the impedance signal from motion tended to occupy a lower frequency band than that from vascular events, f i l t e r i n g was very successful i n removing a r t i f a c t from motion. F i g . 9.2 shows examples of the interference both before and after f i l t e r i n g . Motion involving the musculature beneath the occlusive cuff caused an a r t i f a c t as much as 20 times the amplitude of the impedance signal and f i l t e r i n g alone was not s u f f i c i e n t . Attempts to f i n d a channel with a signal due to motion that correlated well with the motion a r t i f a c t contaminating the desired impedance signal were not successful. The separation of the reactance and resistance components of the impedance signal showed equal contribution i n each component from both motion and vascular events. The same re s u l t was found from separating the admittance signal into i t s conductance and susceptance components. A channel representing only motion-related impedance changes was found using a separate pair of electrodes placed d i s t a l to the sense electrodes, but the ch a r a c t e r i s t i c s of the signal were s u f f i c i e n t l y d i f f e r e n t to make simple subtraction unviable. It may be possible to f i n d an adequate noise channel for simple subtraction using a d i f f e r e n t electrode configuration. A separate channel could also be used to detection motion as well. When excessive a r t i f a c t from motion i s detected, processing of the impedance data could be suspended and, i f used to control the pressure i n a cuff, the pressures could be temporarily increased. The best means of reducing the e f f e c t of motion on the impedance signal i s to f i n d an electrode configuration that i s less sensitive to motion but retains the desired c h a r a c t e r i s t i c s of the p u l s a t i l e impedance related to vascular events. no motion «4-« wrist and finger movement (a) Impedance signal at occlusion; bandwidth is 1.5 Hz to 30 Hz wrist and finger movement (b) Impedance signal at occlusion; bandwidth is 0.05 Hz to 100 Hz F i g . 9.2 Effect of f i l ter ing on motion art i fact . I l l 9.5 DISCUSSION OF THE FEASIBILITY OF USING AN ALGORITHM BASED ON IMPEDANCE PLETHYSMOGRAPHYTOCOUTROLTHE The f e a s i b i l i t y of using the impedance of the underlying tissue to control the cuff bladder pressures was demonstrated by implementing real-time control of the bladder pressures using a single subject with the cuff placed on the arm. The control held the impedance pulse amplitude stable with a cuff pressure of about 50 mmHg above occlusion. By changing the set point or the pressure difference between the tissue pressure peaks, the operating pressure r e l a t i v e to occlusion could be changed. When the system was disturbed with a transient change i n blood pressure by r a i s i n g the legs, the magnitude of the re s u l t i n g disturbance was greatly reduced (Fig. 7.4). Control was implemented by continuously scanning the impedance pulse signal for peaks. When one was found, i t s amplitude was calculated and compared to the set point, which was 1/2 the reference amplitude. The difference between the two, the error, was used to calculate the change i n cuff pressure needed to bring the impedance pulse amplitude nearer to the set point. The change i n pressure required was calculated from: P(n+l)=[Z(n)-S]*F/K where Z(n) i s the current peak amplitude; S i s the set point; 112 P(n+1) i s the cuff pressure that w i l l be applied u n t i l the next peak amplitude i s found; F i s the conversion factor between pulse amplitude and cuff pressure taken from the slope of the average regression l i n e of the pulse amplitude versus r e l a t i v e cuff pressure rel a t i o n s h i p ; and K i s an integer d i v i s o r used to adjust the gain of the control loop. A value of 2 was used. Because the control system became unstable when averaging or median f i l t e r i n g of the pulse amplitudes was t r i e d , the control algorithm was used without these features and remained sensitive to motion and other sources of a r t i f a c t . It did, however, demonstrate the f e a s i b i l i t y of c o n t r o l l i n g the pressure of a pneumatic tourniquet cuff based on the impedance of the underlying tissue, and was one of the contributions of the thesis research. 9.6 THE CL^ICAL IMPLICATIONS OF VASCUL.AR OSCJ^iLATJK^ One further contribution of the thesis research was the i d e n t i f i c a t i o n and p r i o r i t i z a t i o n of the sources of a r t i f a c t and interference encountered i n the study together with the development of approaches for detecting and handling the most s i g n i f i c a n t types. One of the sources of interference was vascular o s c i l l a t i o n s . The low frequency modulation of the amplitude of the impedance pulses caused by the o s c i l l a t i o n s showed a large inter-subject v a r i a b i l i t y i n amplitude. Possible causes include the o s c i l l a t o r y synchronous 113 contraction of the smooth muscle i n the a r t e r i a l wall, and an i n s t a b i l i t y i n the blood flow versus a r t e r i a l pressure re l a t i o n s h i p due to the increase i n a r t e r i a l d i s t e n s i b i l i t y with the reduced translumenal pressure. The o s c i l l a t i o n s have implications i n the i n d i r e c t measurement of blood pressure by using a blood pressure cuff and sensing the flow of blood past the cuff. As the pressure i n the cuff i s being slowly decreased, vascular o s c i l l a t i o n s may be present near the point of blood flow resumption. If the o s c i l l a t i o n s give a maximum at th i s point, blood flow may occur prematurely, giving an a r t i f i c i a l l y high blood pressure reading. Alternately, i f the o s c i l l a t i o n are at a minimum, the resumption of blood flow w i l l be delayed, giving an a r t i f i c i a l l y low reading. This phenomenon may be a factor i n the auscultatory gap noticed when taking the blood pressure with some subjects i n which the Korotkoff sounds disappear for several pulses [45]. The investigation of this phenomenon was outside the scope of t h i s thesis, but considering the s i g n i f i c a n t implications with respect to the increasingly popular technique of non-invasive blood pressure measurement, i t does warrant further investigation. 114 10. RECOMMENDATIONS FOR FURTHER WORK 10.1 RECOMME^ ^ I)IIP™XMPRJPX?M?NT^  To bring the adaptive control of the tourniquet cuff pressure using impedance plethysmography closer to actual c l i n i c a l implementation, changes and improvement are required i n three main areas: 1) Improve the accuracy of the rel a t i o n s h i p between the normalized impedance pulse amplitude and the r e l a t i v e cuff pressure. 2) Account for the a r t i f a c t from motion involving the musculature beneath the cuff. 3) Change the equipment and technique to be more p r a c t i c a l and acceptable i n the c l i n i c a l environment. 10.2 THEACCURACYOFTHERELATION Before taking steps to improve the accuracy of the impedance pulse amplitude versus r e l a t i v e cuff pressure relati o n s h i p , the pressure p r o f i l e along the limb used to control the rel a t i o n s h i p should be investigated for i t s pote n t i a l to injure tissue beneath the cuff, and the method used to produce the p r o f i l e should be examined. 115 As mentioned i n 9.2, the peak pressure beneath a dual-bladder cuff i s d i s t a l to the middle of the cuff and the pressure gradient i n the tissue beneath the cuff may actually be greater than that beneath a single bladder cuff i n f l a t e d to the same pressure. Since a large pressure gradient i s a potential cause of injury to the tissue beneath the cuff, the p o s s i b i l i t y of the p r o f i l e increasing the r i s k of injury should be investigated. The p r o f i l e under the dual-bladder cuff was e f f e c t i v e ~ i n generating the rel a t i o n s h i p between cuff pressure and impedance pulse amplitude but required that the tissue pressure p r o f i l e peaks be measured. The instrumentation and procedure for measuring the peaks may not be adaptable to a c l i n i c a l application and other p r o f i l e s along with the means of generating them should be considered. A p r o f i l e similar to the one beneath the dual-bladder cuff was produced, for instance, by s t i t c h i n g the bladder i n a blood pressure cuff to e f f e c t i v e l y divide i t into two smaller bladders, one twice as wide as the other. A channel was l e f t between the two bladders so that they could both be i n f l a t e d to the same pressure from one source, and the s t i t c h i n g was sealed to prevent leaks. Although a v a r i a t i o n i n the p r o f i l e between subjects existed when the cuff was applied and i n f l a t e d , the technique i s promising and possibly could be used to produce a controlled pressure p r o f i l e that i s consistent among subjects. The f i r s t step i n improving the accuracy of the, rela t i o n s h i p i s to remove the the sources of error mentioned 116 i n 6.7, thereby reducing the standard error of estimation. The standard error of estimation for a normalized pulse amplitude used during the control was 12.5 mmHg. A misalignment of the tissue pressure transducer with the pressure maximum by only 0.5 cm could cause an error as large as 17 mmHg by i t s e l f . Once the errors are reduced, i t may be possible to i d e n t i f y and account for other relevant variables a f f e c t i n g the relationship, such as arm circumference, inter-subject v a r i a t i o n i n the r e l a t i o n between mean and s y s t o l i c pressures, and limb tissue composition. The problem with transducer misalignment can be solved with redesign of the transducer. By clus t e r i n g the points at which the transducer can determine the tissue pressure around the expected location of the pressure peaks, the s p a t i a l resolution would be improved. Other improvements are l i s t e d below. 1) Replace the regulators with ones that are more li n e a r , or a l t e r n a t i v e l y regulate the bladder pressures using pressure transducers for feedback to remove fluctuations i n the difference between the bladder pressures as the pressures are varied i n unison. 2) Determine whether the difference between the peaks of the tissue pressure p r o f i l e remains constant when the difference between cuff bladder pressures i s held constant as the bladder pressures are varied. Any 117 s i g n i f i c a n t change i n the difference would a f f e c t the relati o n s h i p between the height of the impedance pulses and the r e l a t i v e cuff pressure. 3) Determine the cause of and remove the a r t i f a c t at high cuff pressures. The probable cause i s either cuff motion or a larger e f f e c t i v e sampling f i e l d than expected. The magnitude of the e f f e c t from either would depend on the electrode configuration. 10.3 HANDLINGMOTIONARTIFACT The i n t i a l step i n dealing with motion a r t i f a c t would be experimentation to f i n d the electrode type and configuration which: 1) retains the desirable features of the p u l s a t i l e component of the impedance si g n a l ; 2) gives the lea s t a r t i f a c t from motion; and 3) has an additional pair of electrodes from which an impedance signal can be taken that accurately r e f l e c t s the motion a r t i f a c t corrupting the p u l s a t i l e impedance signal from the sense electrodes. It would be desirable for the signal from the additional channel to be well correlated with the noise corrupting the p u l s a t i l e impedance signal such that i t could be subtracted from the channel to remove the corrupting a r t i f a c t from motion. If not fea s i b l e , the signal must at least adequately r e f l e c t the occurrence of the corrupting 118 signal to allow suspension of the processing of the impedance signal u n t i l the l e v e l of the corrupting signal returns to an acceptable value. 10.4 RECOMMEN CHANGES TO MAKE THE g ,^ g M O R E PRACTICAL, FUNCTIONAL, AND ACCEPTABLE IN A (^jIimSAl! Changes to the equipment and technique to make them more p r a c t i c a l , functional, and acceptable i n the c l i n i c a l environment would include the following: 1) derive the reference continuously; 2) simplify the method of positioning the cuff; 3) change the electrode type and improve the electrode configuration; and 4) assure the equipment w i l l be safe and capable of surviving the c l i n i c a l environment. Because parameters such as the electrode c h a r a c t e r i s t i c s can change during a procedure, i t would be necessary to measure the reference throughout the procedure. The technique that was used i n t h i s study to measure the reference would require suspending control of the bladder pressures b r i e f l y during a procedure, and ramping down the pressure i n the proximal cuf f . A better method of establishing a reference would be to take the necessary signal from the same electrodes as the p u l s a t i l e impedance signal and update the reference every sample. The only technique t r i e d that was capable of 119 t h i s r e l i e d on the negative s e n s i t i v i t y of the e f f e c t i v e sampling f i e l d and was very dependent on electrode placement when used with the single-bladder c u f f s . With a controlled tissue pressure p r o f i l e , the technique may be e f f e c t i v e . The cuff must be simple to apply to be acceptable i n the hurried environment of the operating room. Accordingly, alignment of the cuff to the arm must be simpler than the technique used i n t h i s study of f i r s t locating the brachial artery by palpation and then positioning the cuff r e l a t i v e to the artery. Although the temptation i s to use band electrodes because they are less position s e n s i t i v e than spot electrodes, spot electrodes can increase the signal to noise r a t i o , reduce a r t i f a c t from sources such as motion and alternate current paths, and of f e r the means of establishing a continuous reference. Because of these advantages, an investigation should be directed at finding a method to automatically select the electrodes with the best output from an array attached to the tourniquet cuff which would could then be placed i n p r e c i s e l y over the brachial artery. The equipment must also survive i n the harsh environment of an operating room without putting the patient of s t a f f at r i s k of injury. This would not only require meeting the CSA standard G22.2 #125, but would also require that the high energy from e l e c t r o s u r g i c a l units and d e f i b r i l l a t o r s not damage the instrument or f i n d an alternate current path that could put the s t a f f of patient at r i s k of injury. 120 11. REFERENCES [I] ECRI, "Hazard: pneumatic tourniquets used for regional anesthesia," Health Devices, Vol. 13, No.12, pp.48-49, 1984. [2] Dorland's Illustrated Medical Dictionary, 25th Ed. Philadelphia: W.B.Saunders, 1974, p.1622. [3] C.F.Bolton, R.M.McFarlane, "Human pneumatic tourniquet p a r a l y s i s , " Neurology, Vol. 28, pp.787-793, 1978. [4] Athol Parkes, "Ischaemic e f f e c t s of external and int e r n a l pressure on the upper limb," Hand, Vol. 5, No.2, pp.105-112, 1973. [5] ECRI, "Evaluation: pneumatic tourniquets," Health Devices, Vol. 13, No. 12, p.299-315, 1984. [6] James A. Shaw et a l , "Guidelines for the use of d i g i t a l tourniquets based on physiological pressure measurements," J. Bone and Joint Surg., Vol. 67-A, No.7, pp.1086-1090, 1985. [7] James A. McEwen and Robert W. McGraw, "An adaptive tourniquet for improved safety i n surgery," IEEE Trans. Biomed. Eng., Vol. BME-29, No.2, pp.122-128, 1982. [8] R.Sanders, " The tourniquet: instrument of weapon?" Hand, Vol. 5, No.2, pp.119-123, 1973. [9] L.Klenerman, "Tourniquet time-how long?" Hand, Vol. 12, No.3, pp.231-234, 1980. [10] L.Klenerman et a l , "Systemic and l o c a l e f f e cts of the application of a tourniquet," J. Bone and Joint Surg., Vol. 62-B, No.3, pp.385-388, 1980. [II] "Recommended practices," AORN J. , Vol. 39, No.5, pp.808-812, 1984. [12] L.Klenerman, "The tourniquet i n surgery," J. Bone and Joint Surg., Vol. 44-B, No.4, pp.937-943, 1962. [13] James A. McEwen and G.F.Auchinleck, "Advances i n surgical tourniquets," AORN Jv, Vol. 36, No.5, pp.889-896, 1982. [14] J u l i a n M. Bruner, "Time, pressure, and temperature factors i n the safe use of the tourniquet," Hand, Vol. 2, pp.39-42, 1970. 121 [15] J.Ochoa et a l , "Anatomical changes i n peripheral nerves compressed by a pneumatic tourniquet," J.Anat., Vol. 113, No.3, pp.433-455, 1972. [16] " E d i t o r i a l : The tourniquet, instrument or weapon?" Can. Med. Assoc. J. , Vol. 109, p.827, 1973. [17] L.N.Hurst et a l , "The pneumatic tourniquet: a biomechanical and electrophysiological study" Plastic and Reconstructive Surgery, Vol. 67, No.5, pp.648-652, 1981. [18] John P. Woodcock, Theory and Practice of Blood Flow Measurement. London: Butterworth & Co. Ltd., 1975, pp.238-239. [19] B i l l C. Penny et a l , "The impedance plethysmographic sampling f i e l d i n the human c a l f , " IEEE Trans. Biomed. Eng., Vol. BME-26, No.4, pp.193-198, 1979. [20] Dorland's Illustrated Medical Dictionary. 25th Ed. Philadelphia: W.B.Saunders, 1974, p.1210. [21] John G. Webster, Medical Instrumentation: Application and Design. Boston: Houghton M i f f l i n Company, 1978, pp.420-427. [22] J.P.Woodcock, "Plethysmography," Biomed.Eng., pp.406-409, 1974. [23] V.C.Roberts, "A review of non-invasive measurement of blood flow," Biomed.Eng., pp.332-335, 1974. [24] H.Herscovici and D.H.Roller,"Noninvasive determination of central blood pressure by impedance plethysmography," IEEE Trans. Biomed. Eng., Vol. BME-33, No.6, 1986. [25] L.E.Baker et a l , "Simple v e r s a t i l e instrument for measuring impedance changes accompanying physiological events," Med. Biol. Eng., pp.220-229, 1973. [26] Benjamin C. Kuo, Automatic Control Systems. 4th Ed. Englewood c l i f f s : Prentice-Hall, 1982, pp.10-11. [27] B i l l C. Penny, "Theory and cardiac applications of e l e c t r i c a l impedance measurements," CRC Critical Reviews in Biomedical Engineering, Vol. 13, No.3, pp.227-281. [28] L.A.Geddes and L.E.Baker, Principles of Applied Biomedical Instrumentation. New York: John Wiley and Sons, 1975, p.229. [29] G.W.Mauck et a l , "The meaning of the point of maximum o s c i l l a t i o n s i n cuff pressure i n the i n d i r e c t measurement of blood pressure-part 2," Journal of Biomechanical Engineering, Vol. 102, pp.28-33, 1980. 122 [30] James J. Smith and John P. Kampine, Circulatory Physiology. 2nd Ed. Baltimore: Williams and Wilkins, 1984, pp.154-156. [31] Arthur C. Guyton, Textbook of Medical Physiology. 5th Ed. Philadelphia: W.B.Saunders Company, 1976, pp.177-189. [32] F.F.Jorritsma et a l , "Peak-to-peak detector for the a r t e r i a l pulsations i n the plethysmogram 1: technical description and simulation," Med. & Biol. Eng. & Comp., Vol. 19, pp.262-266, 1981. [33] L.A.Marks, " D i g i t a l enhancement of the peripheral admittance plethysmogram," IEEE Trans. Biomed. Eng., Vol. BME-34, No.3, pp.192-198, 1987. [34] D.Murphy et a l , "Impedance imaging i n the newborn," Clin. Phya. Physiol. Meas., pp.301-140, 1987. [35] L.Tarassenko et a l , "Use of d i g i t a l techniques to process cerebral e l e c t r i c a l impedance signals i n the newborn," Med. & Biol. Eng. & Comp., Vol. 22, pp.55-62, 1984. [36] C.S.A., "Electromedical Equipment," C22.2 No. 125-M1984, Canadian Standards Association, Ottawa, 1984. [37] F.A.Anderson et a l , "Impedance plethysmography: the o r i g i n of e l e c t r i c a l impedance changes measured i n the human c a l f , " Med. & Biol. Eng. & Comput., Vol. 18, pp.234-240, 1980 [38] B i l l C. Penny, "Development of a two channel impedance plethysmograph, signal analysis techniques, and electrode arrays: with application to caroti d stenosis detection," Ph.D. thesis, 'Worcester Polytechnic I n s t i t u t e , University Microfilms, Ann Arbor, Mich., 1979 [39] D.B.Geselowitz, "An application of electrocardiographic lead theory to impedance plethysmography," IEEE Trans. Bio-Med. Eng. Vol. 18, pp.38-41, 1971. [40] J. Lehr, "A vector derivation useful i n impedance plethysmographic f i e l d c a l c u l a t i o n s , " IEEE Trans. Bio-Med. Eng., Vol. 19, pp.156-157, 1972 [41] J.R.Mortarelli, "A generalization of the Geselowitz re l a t i o n s h i p useful i n impedance plethysmographic f i e l d c a l c u l a t i o n s , " IEEE Trans. Bio-Med. Eng., Vol.27, pp.665, 1980 [42] J.A.McEwen, "Biomedical Pressure Transducer," U.S. Patent Application f i l e d March, 1987. [43] Jan Nyboer, Electrical Impedance Plethysmography. S p r i n g f i e l d : Charles C. Thomas, 1970, pp.110-112. 123 [44] Paul G. Hoel. Elementary Statistics. 2nd Ed. New York: John Wiley & Sons, Inc., 1967, p.218. [45] L.A.Geddes, The Direct and Indirect Measurement of Blood Pressure. Chicago: Year Book Medical Publishers, Inc., 1970, pp.111. [46] Surya N. Mohapatra, Non-invasive Cardiovascular Monitoring by Electrical Impedance Technique. London: Pitman Publishing Pty Ltd., 1981, pp.20-22. [47] John G. Webster, Medical Instrumentation: Application and Design. Boston: Houghton M i f f l i n Company, 1978, p.7. [48] Don Lancaster, Active-Filter Cookbook. Indianapolis: Howard W. Sams & Co., Inc., 1975, p.142. [49] Jacob Millman, Microelectronics. New York: McGraw-Hill Book Company, 1979, pp.409-444. [50] Lewis Smith and D.H.Sheingold, "Noise and operational amplifier c i r c u i t s , " Analog Dialogue, Vol.3, No.l, 1969. [51] Donn Soderquist, "Minimization of noise i n operational amplifier applications," PMI Application Notes, AN-15, 1975 [52] James A. McEwen, "Complications of and improvements i n pneumatic tourniquets used i n surgery, Med. Instrum., Vol. 15, No.4, pp.253-257, 1981. [53] Harold Alexander et a l , " C r i t e r i a i n the choice of an occluding cuff for the i n d i r e c t measurement of blood pressure," Med. & Biol. Eng. & Comput., Vol. 15, pp.2-10, 1977. [54] Arthur C. Guyton, Textbook of Medical Physiology. 5th Ed. Philadelphia: W.B.Saunders Company, 1976, p.275. [55] M.Y.Jaffrin and C.Vanhoutte, "Quantitative interpretation of a r t e r i a l impedance plethysmographic signals," Med. & Biol. Eng. & Comp., pp.2-10, 1979. [56] M.Intaglietta, "Capillary transport phenomena due to a r t e r i o l a r vasomotion," The Winter Annual Meeting of the American Society of Mechanical Engineers, Anaheim, C a l i f o r n i a , December 7-12, 1986. [57] Daniel J . Haumschild, "Microvascular blood flow measurement by laser Doppler flowmetry," TSI Application Note, TSI Inc., St. Paul, MN, 1986. [58] Ned C. Huang, David R. Gross, D a l i J. Patel, Quantitative Cardiovascular Studies. Boston: University Park Press, 1979, p.101. [59] Arthur C. Guyton, Textbook of Medical Physiology. 5th Ed. Philadelphia: W.B.Saunders Company, 1976, p.217. [60] P.A.Lynn, "Online d i g i t a l f i l t e r s for b i o l o g i c a l signals: some fast designs for a small computer," Med. & Biol. Eng. & Comput., Vol.15, pp.534-540, 1977. [61] Paul G. Hoel. Elementary Statistics. 2nd Ed. New York: John Wiley & Sons, Inc., 1967, pp.27-28. [62] W.J.Tomkins and J.G.Webster, Design of Microcomputer-Based Medical Instrumentation. Englewood C l i f f s : Prentice-Hall Co., 1981, pp.185-186. 125 APPENDIX A: THE MEASUREMENT OF ELECTRODE, TISSUE, AND ARTIFACT RELATED IMPEDANCES AND THE DETERMINATION OF THE REQUIRED PLETHYSMOGRAPH SIGNAL RESOLUTION In order to determine the impedance plethysmograph performance s p e c i f i c a t i o n s , several impedances were measured. These included the basal impedance of a section of an arm, the p u l s a t i l e impedance change i n the section due to blood c i r c u l a t i o n , the impedances of a variety of electrodes, the impedance change that accompanies electrode movement, the impedance change caused by motion, and the impedance between a subject and ground. From these measurements, the re s o l u t i o n requirement of the impedance plethysmograph was determined. A l . BASAL IMPEDANCE, PULSATILE IMPEDANCE, ELECTRODE AND MOTION ARTIFACT An experiment was devised to measure the magnitude of the following impedances: 1) basal tissue impedance of a section of the arm, 2) impedance change from electrode movement, 3) impedance change from finger, wrist, and elbow f l e x i o n and extension, and 4) p u l s a t i l e impedance change from blood c i r c u l a t i o n i n the limb. The impedance was determined at a frequency of 30 kHz and a current of 5 mA, rms. The experiment was performed with disposable s i l v e r - s i l v e r chloride EKG electrodes i n the configuration shown i n F i g . Al using a single subject. Wavetek Inc. Model 188 Function Generator 100 ohm Al Experimental c o n f i g u r a t i o n to measure basal impedance, p u l s a t i l e impedance, and the a r t i f a c t from e l e c t r o d e s and motion. 127 The equipment with connections i s also shown i n F i g . A l . The voltage V< was adjusted to 0.71 V peak to assure an i n i t i a l current of 5 mA, rms. D i f f e r e n t i a l voltages were measured using the d i f f e r e n t i a l mode of the 7A26 Dual Trace Amplifier plug-in of the Tektronix 7601 Oscilloscope with times 10 probes. Unipolar voltages were measured with the 7A15A V e r t i c a l Amplifier, again using a times 10 probe. The impedance between the sense electrodes ( i . e . , the basal impedance of the section of tissue between the sense electrodes) was calculated from Z s e n s e = (V2 " V3 ) / I . The t o t a l impedance was calculated from Zt o t a l = V i / I . For the remaining impedances, the change i n current due to the relevant change i n impedance was determined by measuring the change i n voltage across the 100 ohm r e s i s t o r ( 6 V < ) . The change i n current i s then given by: 6 V 4 / 100 ohm . To measure the change i n voltage (5V«), the r e s i s t o r voltage, V«, was buffered with an instrumentation amplifier 128 (Analog Devices AD521K). The output was connected to a diode detector which i n turn was connected to an 8811A B i o e l e c t r i c Amplifier i n the Hewlett-Packard 7754A Chart Recorder. The changes i n the peak value of V« could then be measured from the chart recordings. The change i n impedance, 6 Z , would cause a change i n current, 5 1 , such that: Vi = ( I - 61) * ( Z t o t a l + SZ) = I * Z t o t a l + 5Z * ( I - 61 ) - 51 * Z t o t a l If the in t e r n a l impedance of the generator i s ignored, and 51 i s assumed to be i n phase with I, then V i = I * Z t o t a l and ! 5 Z i = ( 1 6 1 1 * i Z t o t a i l ) / ( III - 1811 ) (1) The above assumptions allow a f i r s t approximation of the impedances; they could r e s u l t i n an error as large as 15%. Because the diode detector gives a voltage output i n d i c a t i v e of the peak value, the peak value of I must be used i n equation (1). 129 From the experiment: IMPEDANCE VALUE (ohm) Basal impedance 49.5 Total impedance 651 Electrode movement, sideways 0.78 Electrode movement, up and down 5.4 Finger extension and f l e x i o n 0.14 Wrist extension and f l e x i o n 5.6 Elbow extension and f l e x i o n 6.0 P u l s a t i l e impedance change 0.22 In a separate test with the electrodes beneath a cuff, the maximum value of electrode motion a r t i f a c t was calculated to be 0.38 ohm for band electrodes without conductive electrode paste, and 2.4 ohm for conductive adhesive electrodes. The motion consisted of proximal and d i s t a l movement of the cuff, rotation of the cuff, and squeezing the electrode beneath the cuff. A2. REQUIRED PLETHYSMOGRAPH SIGNAL RESOLUTION The calculated impedance changes from movement are for the entire section of the arm between the outer two electrodes and would be less for smaller sections of the arm such as the tissue beneath a cuff placed around the upper arm. This also applies to the p u l s a t i l e impedance change due to the c i r c u l a t i o n of blood within the arm. Because the distance between the middle two electrodes i s about 1/2 the distance 130 between the outer two electrodes, the r a t i o , W, of p u l s a t i l e impedance to basal impedance would be:. W = ( 0.22 * 1/2 ) / 49.5 = 0.0022 Using a desired resolution of the p u l s a t i l e impedance change of 25:1, the resolution required of the basal impedance pr i o r to extraction of the p u l s a t i l e component i s : R = 0.0022 / 25 = 88 * 10"6 A value of 1/2 of th i s (R = 44 * 10 - 6) was used as a target when designing the impedance plethysmograph. A3. IMPEDANCE FROM THE SUBJECT TO GROUND The c i r c u i t of F i g . A2 was used to determine the impedance of the subject to ground at a frequency of 100 kHz. The voltages were measured with the 7A26 Dual Trace Amplifier plug-in i n the Tektronix 7601 Oscilloscope using times 10 probes. A conductive adhesive electrode placed on the shoulder was used to provide the connection to the subject. The impedance was determined with no apparent connection to ground, with the ri g h t hand grasping a grounded connector, with the l e f t hand grasping a grounded connector, and with both hands grasping a ground connector. The impedance i s given by: 131 © Wavetek Inc. Model 188 F u n c t i o n Generator 10 ki.lohm z g g. A2 Experimental c o n f i g u r a t i o n used t o measure Zg, the impedance from the s u b j e c t t o the power ground. 132 Z„ = V2 / I where I = ( Vi - V 2 ) / 10 kilohm The following l i s t shows the measured impedances. It also shows the impedances when adjusted for a frequency of 30 kHz assuming the tissue capacitances and resistances are independent of frequency. The electrode contact impedance i s included i n these values. IMPEDANCE VALUE AT100kHZ (ohm) (degrees) No apparent path to ground 9.02k -86.5 Right hand grounded 663 -23.4 Left hand grounded 607.2 -23.3 Both hands grounded 415 -22.9 YAkUE A T 3 0 k H Z (ohm) (degrees) 30.0k -88.9 1069 -55.3 976 -55.1 661 -54.6 A4. ELECTRODE IMPEDANCES [46] The c i r c u i t of F i g . A3 was used to measure the electrode impedance of four electrode types at 5 frequencies and two current l e v e l s . D i f f e r e n t i a l voltages were measured using the d i f f e r e n t i a l mode of the 7A26 Dual Trace Amplifier plug-in of the Tektronix 7601 Oscilloscope with times 10 probes. Unipolar voltages were measured with the 7A15A V e r t i c a l Amplifier, again using a times 10 probe. Because there i s i n s i g n i f i c a n t current flow into the dual trace amplifier configured i n a d i f f e r e n t i a l mode, the d i f f e r e n t i a l voltage, Vi - V2 , i s approximately the voltage 133 F i g . A3 Experimental c o n f i g u r a t i o n used t o measure the c o n t a c t impedance of e l e c t r o d e , E3. 134 drop across the electrode, E 3 . The current, I, through the electode i s given by: T I = V 2 / 100 ohm The impedance of electrode E 3 i s given by: Z . l e c t r o d e = ( V i - V 2 ) / I The measured impedances of the d i f f e r e n t types are given i n the table below. ELECTRODE TYP conductive hook and loop, Velcro HI-MEG Conductive Products FREQUENCY CURRENT (kHz) (mA) 5 10 30 50 100 1.25 IMPEDANCE PHASE (ohm) (degrees) 962 -78 735 -70 283 -70 152 -68 79.2 -68 135 conductive adhesive, 5 1.0 601 -68 3M-Littman 2322 Disposable 10 389 -58 EKG Electrodes 30 205 -43 50 177 -38 100 141 -32 5 2.5 764 -68 10 452 -60 30 226 -44 50 181 -38 100 147 -32 disposable band, 5 1.0 255 -74 Contact Products Ltd. 10 120 -72 M6001 Electrode Tape 30 53 -66 50 39 -60 100 25 -52 5 2.5 198 -74 10 113 -71 30 49 -61 50 36 -58 100 32 -51 s i l v e r - s i l v e r chloride disposable EKG, Hewlett-Packard 14445A 136 5 1.0 530 -62 10 304 -52 30 190 -58 50 177 -24 100 156 -16 5 2.5 495 -62 10 311 -50 30 198 -28 50 175 -24 100 164 -16 137 APPENDIX B: IMPEDANCE PLETHYSMOGRAPH SPECIFICATIONS AND DESIGN In t h i s appendix, the s p e c i f i c a t i o n s for the impedance plethysmograph used for the detection of the onset of blood flow are discussed and d e t a i l s are given of the design of the i n i t i a l general purpose impedance plethysmograph and the second simpler unit with better performance c h a r a c t e r i s t i c s . B l . GENERAL REQUIREMENTS The i n i t i a l plethysmograph was designed to resolve to -28dB (25:1) the time-varying impedance signal which r e f l e c t s vascular a c t i v i t y i n the sampled tissue, without considering the hyperemic response. To allow inves t i g a t i o n of parameters relevant to the study, the design incorporated the f a c i l i t y to vary the operating frequency and the magnitude of the current passed through the tissue, and provided for synchronous demodulation to separate the r e s i s t i v e and reactive components of the tissue impedance. The plethysmograph systems shown.in Figs. Bl to B4 were chosen because they did not need to be manually balancing and would be simple to use. The disadvantage of a system without balancing i s that the signal corresponding to the large basal impedance cannot be removed from the desired s i g n a l , corresponding to the small impedance change due to blood flow, pri o r to demodulation as i t can with a bridge c i r c u i t or other configuration that can be balanced. This r e s t r i c t s the amount * selected f o r 4.2v peak at A 510 3 1 . 6 k H z ^ ^ ^ H ^ MCL2630 10 kHz 9IU j 10 kHz Channel MCL2630 +12 © p*y3S7 rj^CURRE 333 — 1 0 . 0 mA 461 - \ y — 7 . 5 mA 3k0 CU RENT OUT -o MC1723C MC1723C 11 10 12 2 6 3 4 5 7 13 „ 10 Capacitors without multipliers are 1n pF 750 — S - 0 ^ 1.6nF 2lc0 Aft 2.5 mA 506pF FREQUENCY 10 kHz 31.6 kHz 100 kHz RESISTOR VALUE 1 mA W F1g. Bl Current source. RF1 RF2 R01 RD2 19.1 kilohm 10.5 kilohm 28.7 kilohm 66.5 kilohm Sense Electrodes j ^ Buffered undemodulated outputs OImpedance L \ V U 50k ] identical I channel for i reactance I F1g. B2 Instrumentation amplifier, demodulator, band-pass f i l t e r , and d i f f e r e n t i a t o r . 330k On current source-Rl 100k 74LS123 14 4 15 2 1 13 1) Values shown for the 10 kHz channel 2) Rl i s adjusted for the proper phase for resistance demodulation 3) R3 1s adjusted for the proper phase for reactance demodulation 180k / + 5 H ^ \ A / — — 330p resi stance demodulation c a r r i e r reactance demodulation channel Fig. B3 Carrier and demodulation signals. 10k Input from demodulator f i l t e r o - 16 8 5 6 11 17 100k M AD630K 13 1 10 19 2 15 N I i I 10k y ±15 O.JA4= LF356 750kl 0.1* • I ' LF356J 3k9 •Ik ±15 60 Hz Notch xrTo5&* F i l t e r 4k7 F1g. B4 Ci r c u i t r y changes for f i n a l plethysmograph 142 of amplification possible p r i o r to demodulation and worsens the signal to noise r a t i o . As discussed i n Appendix A, the required resolution of the basal impedance i s 44 ppm. The bandwidth requirement for the measurement of blood flow i s DC to 20Hz [47] . Since only the time varying component of the impedance was to be used i n th i s study, a bandwidth of 0.05Hz to 20Hz was chosen. B2. CURRENT SOURCE The s p e c i f i c a t i o n s for the current source were: 1) regulation of the current magnitude to within 44 ppm with a maximal change i n impedance of 6.0 ohm (corresponding to gross arm movement); 2) peak noise of less than 44 ppm; 3) i s o l a t i o n from ground for e l e c t r i c a l safety, to minimize alternate current pathways, and to minimize the common mode voltage presented to the instrumentation amplifier. The current source (Fig. Bl) selected used feedback to provide the regulation and was battery-operated to assure the i s o l a t i o n . It accepted square wave inputs coupled through o p t i c a l i s o l a t o r s at three d i f f e r e n t frequencies (10 kHz, 31.6 kHz, and 100 kHz). The fundamental frequency component was extracted form the square waves by using low-pass active f i l t e r s [] to remove the higher harmonics. Any combination of the frequencies could be used with some r e s t r i c t i o n s on the 143 maximum current and load impedance. The output of the low-pass f i l t e r s were summed and applied to a transconductance amlpifier (current-series topology) [] with f i v e switch selectable current settings (1 mA, 2.5 mA, 5.0 mA, 7.5 mA, and 10 mA). By assuming the input impedance of the voltage amplifier within a transconductance amplifier i s much greater than the feedback resistance, Rf , the gain of a transconductance amplifier with a current-series topology i s given by [49]: Af = l o u t / V i n = A / ( Z i o a d + Ro + Rf * (A + 1)) (1). and the output impedance by [49]: Zout = A * Rf + Ro (2) where l o u t i s the output current; V i D i s the input voltage; A i s the gain of the voltage amplifier; Z i o a d i s the load impedance; Ro i s the output impedance of the voltage amplifier; and Rf i s the feedback resistance. D i f f e r e n t i a t i n g the output current i n (1) with respect to the load impedance, substituting S l o u t and S Z i o a d for the d i f f e r e n t i a l s d l 0 u t and d Z i o a d , and rearranging y i e l d s : 5Iout / l o u t = - S Z l o a d / ( Z i o a d + Ro + Rf * (A + 1)) (3) For a required resolution of 44 * 10 - 6 ( i . e . , S l o u t / l o u t ) and an impedance change of 6 ohm corresponding to gross arm movement: Rf * (A + 1) * (-8Z / ( S l o u t / l o u t ) ) - ( Z i o a d + Ro ) * 136 * 103 - ( Z i o a d + Ro ) a 136 * 103 (4) Then for Rf = 1000 ohm, A £ 135 500 ohm, A £ 272 300 ohm, A £ 453. Assuming the voltage amplifier has a frequency response r o l l - o f f of 20 dB per decade, the amplifier unity gain cross-over frequency for an Rf of 300Q and an operating frequency of 100 kHz must be approximately 45 MHz (or about 14 MHz for an Rf of 1000Q). Only by relaxing t h i s s p e c i f i c a t i o n under the worst case of high frequency and current was i t possible to use the monolithic amplifier, LF357AN. The best noise performance i n a current source was found with switching between a stable reference and ground using a 50% duty cycle and removing the higher harmonics with a 4-pole active low-pass f i l t e r . The MC1723C regulator was used as the reference because of i t s good noise s p e c i f i c a t i o n . A high speed d i g i t a l o p t i c a l i s o l a t o r ( G l MCL2630) was used to couple the switching signal to the current source. Because the 145 battery voltages were often found to be excessively noisy, they were regulated with MC1723C regulators. B3. INSTRUMENTATION AMPLIFIER The main function of the instrumentation amplifier shown in F i g . B2 (Analog Devices AD521K) was to amplify the d i f f e r e n t i a l voltage across the sense electrodes while r e j e c t i n g any common mode signals. A high input impedance was required to prevent a r t i f a c t from sense electrode movement. The required gain, common-mode r e j e c t i o n r a t i o , d i f f e r e n t i a l input impedance, common mode input impedance, noise performance, slew rate, bias current, and o f f s e t current were determined. B3.1 Gain The required gain was calculated assuming the output of the amplifier would be adjusted i n i t i a l l y for an 8 to 10 v o l t peak output. Assuming a minimum current of 1 mA and an impedance of 50 ohms, the required gain i s 141. B3.2 Common-modevoltage The c i r c u i t of F i g . B5 was used to determine the common mode voltage at the frequency of the current source. The impedances on the figure are explained below. 146 F i g . B5 Equivalent c i r c u i t of subject, current source, and Instrumentation a m p l i f i e r . 1 4 7 B 3 . 2 . 1 Limb tissue impedances, Z b L . , Z b,.and Z > 2 These impedances are dependent on the electrode configuration, frequency, and current. Z b was measured to be 5 0 ohm ( A 2 . ) at 3 0 kHz and with an i n - l i n e l o n g i t u d i n a l electrode configuration. B 3 . 2 . 2 Electrode impedances, Z B 1 , Z e 2 , Z e 3 , and Z e 4 The electrode impedances were found to vary from 1 0 to 3 9 0 ohm over the frequency range of 1 0 kHz to 1 0 0 kHz. A value of 2 5 0 ohm was used i n the calculations below. B 3 . 2 . 3 C u r ^ sourceputput....„ Z© The current source output resistance was calculated to be ;> 1 3 6 kilohm ( 2 ) . B 3 . 2 . 4 Current source to ground impedance, Z c « 0 The current source to ground impedance was assumed to be much higher than Z 0 and Z e 2 and was ignored. It was eventually measured to be 1 5 5 kilohm with a phase angle of - 8 1 degrees at 3 0 kHz. The assumption i s v a l i d for the li m i t e d accuracy required for the calculations below. 148 B3.2.5 Subjecttogroundimpedance, The value of Z 0 was found at 30 kHz to vary from 980 ohm to 30.0 kilohm with phase angles of -55 and -89 degrees respectively (A4.). B3.2.6 Commonmodeimpedances,Z c « 1,andZ c a? The common mode impedances include the input common mode impedance of the amplifier, the bias r e s i s t o r s , and any stray capacitance to ground from the sense electrode leads. B3.2.7 D i f f e r e n t i m p e d a n c e , Z a The d i f f e r e n t i a l impedance i s the d i f f e r e n t i a l input impedance of the instrumentation amplifier. B3.2.8 Cal c u l a t i o n o f t h e c o m m o n m o d e v o l t a g e Because Z 0 was expected to be much less than Zcmi , Vs was assumed to be close to 0.0 v o l t s . Since I i and I 2 are very much less than I , then: V 4 = I * Zb 2 and V 3 ~ I * ( Z b 2 + Z b ) Defining the common mode voltage as: Vcm = (V< + V 3 ) / 2.0 then V c « = I * ( Zb 2 + Zb/2.0) Defining the d i f f e r e n t i a l voltage to be: Vd = V 3 - V< » I * Zb , then the r a t i o of these voltages i s Vcm / Vd = ( Zb2 / Zb ) + 1/2 This r a t i o i s determined by the electrode configuration. For i n - l i n e configurations, i t approximately corresponds to the distance from the proximal source electrode to the midpoint between the sense electrodes divided by the separation between the sense electrodes. A value of 3 for thi s r a t i o i s a reasonable maximum for the electrode configurations considered. B 3 . 3 Amplifier common-mode re j..e c t i on r a t i o The common-mode re j e c t i o n r a t i o (CMRR) must be s u f f i c i e n t to keep the ef f e c t of changes i n the common-mode voltage to less than one-half of the minimum desired signal r e s o l u t i o n . The equivalent impedance change due to gross arm movement was 6.0 ohm while the p u l s a t i l e impedance change due to blood c i r c u l a t i o n was 0.11 ohm. Thus: 150 (I * 6.0 ohm) * Acm £ (I * 0.11 ohm) * Ad /50 or CMRR = 20 * log (Ad / Ac» ) J> 20 * log (6.0 * 50 / 0.11)) £ 68 dB where Ad i s the d i f f e r e n t i a l gain; and Acn i s the common mode gain. The CMRR of the AD521K instrumentation amplifier was measured to be 72 dB at 100 kHz with the input impedances balanced. B3.4 Common mode input impedance Another consideration i s that Z e n and Z c m 2 must be large enough to prevent the voltage d i v i s i o n of the common mode voltage by the electrode impedances Z e 3 and Z e 4 from causing any change i n common mode voltage to appear as a d i f f e r e n t i a l voltage. The impedance Zd can be ignored because V* approximately equals V - and Zd i s i n the same order of magnitude as Zcmi and Z c m 2 . The impedances, Zb and Zb2 have been ignored i n the cal c u l a t i o n s . The voltage d i v i s i o n would cause a d i f f e r e n t i a l voltage, V d d , across the instrumentation amplifier terminals of magnitude: 151 Vdd = V c « * ( ( Z c m l / (Zcml + Ze3>. - (Zc • 2 / ( Z c m 2 + Ze 4 ) ) = Vcm * ( ( Z c m l * Ze4> ~ ( Z c m 2 * Ze3>) / ( ( Z c m l + Ze 3 ) * ( Z c m 2 + Ze4 ) ) (1) or: 6Vdd = 5 V c m * ( ( Z c m l * Z e 4 ) - ( Z c m 2 * Ze 3 ) ) / ('(Zc • 1 + Ze 3 ) * (Zcm2 + Ze 4) ) (2) Since the electrode impedances are very much less the the common mode impedances, and assuming the common mode impedances are reasonably well matched: 5Vd d ~ 6V C m * (Ze 4 - Ze 3 ) / Zc m 1 This d i f f e r e n t i a l voltage must be less than the desired resolution of the d i f f e r e n t i a l voltage from vascular events within the sampled region. Using the same values as above with an electrode impedance imbalance of 100 ohm: 6Vdd £ 0 . 1 1 Q * I / 50 (3) 6.0Q. * I * 100Q / Zcmi £ 0 . 1 1 Q * I / 50 Z c o i 2 273 kQ Zcmi and Z c m 2 must also be large enough to prevent the same voltage d i v i s i o n of the common mode voltage by the electrode impedances Ze3 and Ze4 from causing a d i f f e r e n t i a l voltage when electrode motion causes the electrode impedances to change. 152 D i f f e r e n t i a t i n g equation (1) with respect to Ze3 y i e l d s : d V d d / d Z e 3 = -Vcm * Zc m 1 / (Ze3 + Zc m 1 ) 2 * ~Vc m / Zc a t Substituting 6 values for the d i f f e r e n t i a l s and rearranging gives: 5Vdd = "8Ze3 * Vcm / Zen. 1 (4) From A3. : | 8 V d d l S V d * 44 * 10"6 (5) Combining (4) and (5), using 5Z e 3 of 5 . 4Q (A2.), and rearranging: Z o i * 123 * 103 * Vco / Va and with Vcm / Vd = 3 Z c B I £ 368 * 103 ohm B3.5 Amplifier d i f f e r e n t i a 1 input impedan« A minimum value of the p a r a l l e l combination of Z e n and Zd can be estimated i n a manner sim i l a r to the procedure used i n B3 . 4 . This p a r a l l e l combination, Z P , must be large enough to prevent a voltage d i v i s i o n of the large d i f f e r e n t i a l mode voltage (corresponding to the large basal resistance) by the 153 electrode impedance Z e 3 from causing a time varying d i f f e r e n t i a l voltage when electrode motion causes the electrode impedance to change. F i r s t replace V 3 - V 4 and V 4 i n F i g . B5 with t h e i r approximate Thevenin equivalents of voltage sources I*Zb and I * Z b 2 and respective equivalent series impedances of Zb and Zb2 as shown i n F i g . B 6 ( a ) . According to the p r i n c i p l e of superposition, the e f f e c t of only the d i f f e r e n t i a l voltage, V3 - V 4, can be determined by replacing the other voltage source with a short. By doing t h i s along with a pie-to-tee conversion y i e l d s the c i r c u i t of F i g . B 6 ( b ) . To a f i r s t approximation: Z 3 = Z e 4 , Z 2 = Zo , and Z i =0.0 .Then as shown i n Fig.B6(c): Vd d 2: Vd * Z P / ( Z P + Ze 3 + Zb ) and 5Vdd » - 5 Z e 3 * Vd * Z P / ( Z P + Zes + Z b ) 2 or: 5Vdd a - 5 Z e 3 * Vd / Z P (6) Combining ( 6 ) and (5), using 6 Z e 3 of 5.4Q, and rearranging: Z P * 123 * 103 Ze3 Ze4 Zd Vdd Zcml (a) C i r c u i t after replacing the sources of the d i f f e r e n t i a l and common-mode voltages with t h e i r Thevenln equivalents Zg Zcm2 Ze3 Zd Vdd (b) C i r c u i t after shorting Vcm and a p1e-to-tee conversion Zl Z3 Z2 lorn) Ze3 zp Vdd (c) C i r c u i t after approximating Impedance values F1g. B6 C i r c u i t s used to determine Zd. 155 B3.6 Noise per• formance [50,51] Assuming an output from the Instrumentation Amplifier of 8 V peak, and a resolution of 44 ppm, then the peak noise referred to the output must be less than: E 0 v £ 8 volts * 44 * 10"6 = 350 uvolts, peak or 120 uvolts rms. The bandwidth i s approximately two times the bandwidth of the signal of i n t e r e s t , or 40 Hz, centered about the c a r r i e r frequency of 10 kHz, 31.6 kHz, or 100 kHz. The noise referred to output for the AD521K instrumentation amplifier for the bandwidth, 10 Hz to 10 kHz i s given by: N o u t p u t = /((1.2 * G) 2 + 2500)) uvolts rms Where G i s the maximum required gain (141). For thi s large bandwidth, the noise i s then: N o u t p u t = 176 uvolts rms. Assuming the noise i n thi s bandwidth i s white, the noise for the reduced bandwidth i s only: Eov = (40 Hz / 10 kHz) * 176 uvolts = 0.7 uvolts. 156 The equivalent input current of the Johnson noise from the bias r e s i s t o r s i s given by I n j = (1/8) * /(B * R) picoamps where B i s the bandwidth i n Hertz (40 Hz), R i s the resistance i n Mohm, and In j i s the white rms noise current i n picoamps. For R = 240 kohm, I n j =0.39 picoamps rms. The amplifier input noise i s given as 15 pA rms for a bandwidth of 10 Hz to 10 kHz. Assuming the noise i s white, the amplifier noise current i n the bandwidth of int e r e s t i s Ina = (15 pA) * ( 40 / 10,000) = 60 fA. This current i s much less than the Johnson noise current from the bias r e s i s t o r s and can be ignored. From the equivalent c i r c u i t of F i g . B7, the voltage, Vdd, from the noise current sources assuming that Zb i s much less than Z e 3 i s aproximately: IVdd I = 1 Inj * Z e 3 1 = 97 pV where Z e 3 i s the electrode impedance assumed to be 250 Q. The eqivalent noise at the output i s E©j = 97 pV * 141 = 0.013 uV. The t o t a l noise at the output i s then: (Eo j 2 + E o v 2 ) 1 / 2 = (0.7 2 + 0.0132) uV = 0.70 uV rms. This i s much less than the maximum desired noise of 120 uV rms. B3.7 Slew rate and fr.eguency__..response The AD521K instrumentation amplifier has a smalll signal bandwidth of 200 kHz at a gain of 100 and 40 kHz at a gain of 157 158 1000. This should be adequate for the maximum gain of 141 and maximum frequency of 100 kHz. The slew rate, SR i s given by: SR = Vnax * w where Vmax i s the maximum expected voltage amplitude (10 V) at the output of the amplifier, and w i s the maximum radian frequency (628 * 103 radians per second). These give a required slew rate of 6.28 V/us which i s less than the 10 V/us slew rate of the AD521K. B3.8 Bias current and o f f s e t voltage Assuming well matched bias r e s i s t o r s , the o f f s e t at the amplifier output due to of f s e t current and o f f s e t voltage i s given by: Vo = (Io * Rb + Vo ) * G B B X + Voo where I 0 the maximum expected input o f f s e t current (12.5 nA); Rb i s the bias resistance; Vo i s the maximum expected input o f f s e t voltage (1.6 mV); Goax i s the maximum expected gain (141); and Voo i s the maximum expected output o f f s e t voltage (203 mV). 159 The maximum expected values of l o , V 0 , and Voo are the maximums s p e c i f i e d i n the AD521K data sheet assuming a 20 degree centigrade r i s e i n temperature. Using a maximum desired o f f s e t at the output of 1.0 V so that the dynamic range of the signal i s not seriously compromised, a maximum value of the bias r e s i s t o r s becomes: Rt £ 324 kQ B3.9 Bias r e s i s t o r s The following r e s t r i c t i o n s apply to the common-mode input impedance of the instrumentation amplifier (including the bias r e s i s t o r s ) , the d i f f e r e n t i a l input impedance of the instrumentation amplifier, and the bias r e s i s t o r s : ! Z c m i £ 273 kQ ; I Z c m I £ 368 kQ ; | Z P I * 123 kQ ; Rb £ 324 kQ ; These conditions produce c o n f l i c t i n g requirements, and cannot be met i f the AD521K instrumentation amplifier i s used. The AD521K was used for reasons of a v a i l a b i l i t y and price, and bias r e s i s t o r s of 220 kQ ± 5% were matched to within 1% and used. At a frequency of 100 kHz, the equivalent common-mode input impedance became 203 kQ at an angle of -22.5 degrees. 160 The d i f f e r e n t i a l impedance became 884 kQ at an angle of -90 degrees. B4. DEMODULATOR By switching or demodulating the high frequency signal from the instrumentation amplifier with signals having the appropriate phase, low frequency signals arise that corresponds to the instantaneous values impedance, resistance, or reactance. The dynamic response of most demodulators w i l l be adequate since many demodulators and m u l t i p l i e r s are available that are s p e c i f i e d for performance well beyond 100kHz. However, low-frequency noise i s r a r e l y s p e c i f i e d . The bandwidth of the the demodulators' low frequency f l i c k e r noise inherent to many l i n e a r semiconductors overlaps the desired impedance signal bandwidth. Since t h i s parameter i s not often s p e c i f i e d , i t ' s e f f e c t must be determined experimentally. As i t turns out, the dominant souce of noise i n the f i r s t v e r s a t i l e impedance plethysmograph was from the MC1494L used as a demodulator. B5. FURTHER SIGNAL PROCESSING The main purpose of the signal conditioning af t e r the demodulators i s to remove the high frequency components of the demodulation process, remove the DC component corresponding to 161 the basal impedance, and amplify the the small time varying signal that represents the vascular a c t i v i t y . To record the output on the HP chart recorder without a f f e c t i n g the b i o e l e c t r i c amplifiers with large high frequency excursions, the high frequencies needed to be attenuated to about the l e v e l of the small time varying signal (about -60 dB). This was done with three single-pole low-pass f i l t e r s i n s e r i e s . A.c.-coupling with a low frequency cutoff of 0.05 Hz removed the d.c. component. B6. DIFFERENTIATOR A time constant of 50 ms was used. A double pole i n the transfer function at 100 Hz attenuated the high frequency components. 162 APPENDIX C: DISTURBANCE VARIABLES The disturbance variables which can af f e c t the cuff pressure required to collapse the a r t e r i e s beneath the cuff are patient related, equipment related, cuff related, and application related. CI PATIENT RELATED VARIABLES The patient related variables are l i s t e d below. CI.1 S y s t o l i c b l o o d p r e s s u r e The cuff pressure required for occlusion i s d i r e c t l y related to the patient's s y s t o l i c blood pressure. In fa c t , the patient's preoperative s y s t o l i c blood pressure has been recommended as a guide for setting the tourniquet cuff pressure [5]. The s y s t o l i c blood pressure can change intraoperatively by as much as 70 mmHg [8]. CI.2 Limbcircumference It has been established that a minimum r a t i o of cuff width to limb circumference i s necessary before the cuff pressure accurately r e f l e c t s the s y s t o l i c pressure i n the limb a r t e r i e s . Therefore, for a given cuff width, i f the limb circumference i s too large, the cuff pressure necessary to collapse the underlying a r t e r i e s w i l l be elevated [5]. 163 C l . 3 Limb geometry; The shape of the limb can a f f e c t the transmission of pressure from the cuff to the underlying tissue. The pressure d i s t r i b u t i o n beneath a cuff on a limb with a large taper can be quite d i f f e r e n t from that on a limb with a r e l a t i v e l y uniform circumference [5]. Also, the geometric r e l a t i o n of anatomical structures, such as the nearness of an artery to a bone, can possibly a f f e c t the cuff pressure necessary to collapse the artery. Cl.4 Tissuecpmpliance The compliance of the tissue beneath the cuff can also a f f e c t the transmission of pressure to underlying tissue [5]. A limb with a thick layer of subcutaneous f a t would be much more compliant than a limb with l i t t l e f a t and well toned muscles. C l . 5 C o n d i t i o of the....arteries. The compliance of art e r i e s can be reduced from atherosclerosis and c a l c i f i c a t i o n requiring s u b s t a n t i a l l y pressure to collapse them [5]. more C2. EQUIPMENT RELATED VARIABLES 164 The equipment related variables are l i s t e d below. C2.1 Pressuresensor_a Any inaccuracy i n the sensed cuff pressure w i l l cause the cuff pressure to be o f f s e t by the amount of the inaccuracy. C2 .2 Regu 1 ator _h eresis Tourniquet systems can allow the cuff pressure to vary within l i m i t s about the set point without correction [52]. C2.3 Leaks and resultant pressure d r„°P s Leaks can cause pressure drops along tubing connecting the components of a tourniquet system. If the pressure drop occurs between the cuff and the pressure sensing element, the cuff pressure can be o f f s e t by the amount of the pressure drop. C2.4 I n f l a t i o n f l o w r a t e The a b i l i t y to maintain the cuff pressure during periods of rapid cuff bladder volume change may be inadequate i f the tourniquet gas flow rate i s i n s u f f i c i e n t . This may happen during manipulation of the limb or the sudden removal of a 1 165 co n s t r i c t i o n around the cuff (such as someone leaning against the c u f f ) . C3. CUFF RELATED DISTURBANCE VARIABLES The cuff related disturbance variables are l i s t e d below. C3.1 Cuff width This was discussed i n C1.2. C3.2 Cuffmanufacture In t h i s study, i t was found that the pressure transmission by the tourniquet cuff was much d i f f e r e n t than that of a blood pressure cuff . The tourniquet cuff had a bladder with a smaller cross-section when i n f l a t e d but was longer. It also had a r i g i d backing against which the bladder could expand. The bladder was also secured along i t s edges whereas the bladder i n the blood pressure cuff was free to move and could actually be removed from i t s cl o t h containment. These differences appear to a f f e c t the transmission of pressure from within the cuff bladder to the limb. 166 C4. DISTURBANCES RELATED TO CONDITIONS DURING THE PROCEDURE AND TO THE APPLICATION OF THE CUFF TO THE LIMB Other disturbances are related to conditions during the procedure and to the application of the cuff to the limb., They are l i s t e d below. C4.1 Prqtectiye c Protective coverings are sometimes applied over the limb beneath the tourniquet cuff. The covering could hinder the transmission of pressure to the underlying t i s s u e . C4 . 2 Cuff p o s i t i ^ The po s i t i o n of the cuff on the limb can change the cuff width to limb circumference r a t i o and the limb geometry beneath the cuff. The e f f e c t of these i s mentioned i n CI.2 and CI.3. C4.3 S n u g n e s s o f t h e c u f f a t a p p l i c a t i o n In t h i s study, the snugness with which the cuff was applied was found to a l t e r the d i s t r i b u t i o n of the pressure beneath the cuff. 167 C4.4 Lubrication The presence of l u b r i c a t i o n beneath the cuff can a f f e c t the transmission of pressure from the cuff [53]. C4.5 Limbmanipulat By manipulating the limb aft e r application and i n f l a t i o n of the cuff, the d i s t r i b u t i o n of pressure beneath the cuff can be altered. 168 APPENDIX D: THE LIMB MODEL Dl. INTROpyCTIO The arm model i s based on a minimum energy solu t i o n for a region of space given that c e r t a i n parts of the region have fixed potentials. The solution requires that the pot e n t i a l at every point (except those with fixed potentials) i s the average of the poten t i a l of the surrounding region. The model i s a rectangular array of nodes (hence two dimensional) representing a longitudinal section of the arm. The program (CURRENT) that solves the nodal potentials requires only the number of rows and columns and the coordinates and potentials of the nodes with fixed p o t e n t i a l s . The program (HPFLUXLINES) that draws the flux l i n e s further requires that the fix e d potentials are along an edge of the model, although the nodal potentials can be solved for any combination or location of nodes with fixed p o t e n t i a l s . It also requires that the nodes from which the flux l i n e s are to eminate are the f i r s t nodes with fixed potentials i n the input f i l e (INPUT.DAT) and that they have the same p o t e n t i a l . D2. THEPROGRAMS There are s i x programs used with the model: INPUT, CURRENT, FLUX1, HPFLUXLINES, BASICFLUXLINES, and NUMBER. The 169 programs were separated because together they require more than 64k of memory, the maximum allowed for the compiler that was used. The program INPUT i s used to create a f i l e (INPUT.DAT) with the number of rows, columns and nodes with f i x e d potentials; and also the coordinates (row and column) and pote n t i a l of the nodes with fi x e d p o t e n t i a l s . The program CURRENT then i t e r a t i v e l y calculates the nodal potentials and writes them to a f i l e (POTEN.TXT). It requires the data from INPUT.DAT although t h i s data can be entered from the console. An acceleration factor to speed convergence, the maximum number of i t e r a t i o n s before the program aborts, the minimum difference i n potential at a node for successive i t e r a t i o n s , and the maximum potential are entered from the console. An acceleration factor of 1.7 to 1.9 seems to work well. A minimum difference ( i . e . , maximum error) of .00005 allows convergence within 300 i t e r a t i o n s for a 30 rows by 65 columns model. The program loads the array of nodal potentials with 1/2 of the maximum potential read from the console. I t was found empirically that t h i s sped up the convergence. The program FLUX1 calculates the x and y components of the nodal e l e c t r i c a l flux density. I t requires the data i n INPUT.DAT although this may be entered from the console. The flu x densities are written to DXDY.TXT. The x component i s calculated by taking the difference between the potentials of 170 the nodes on either side of the current node and d i v i d i n g by two. S i m i l a r l y , the c a l c u l a t i o n of the y component uses the nodes above and below the current node. The program HPFLUXLINES calculates the coordinates of the points that make up the flux l i n e s . It requires the data from INPUT.DAT although t h i s may be entered from the console. It also requires the nodal flux densities i n DXDY.TXT. The number of f l u x l i n e s and the d i r e c t i o n of the flux l i n e s ( i . e . , whether they go from a p o s i t i v e to negative p o t e n t i a l or v i s a versa) are entered from the console. After the number of f l u x l i n e s (N) are entered from the console, the program looks i n INPUT.DAT for the f i r s t N nodes with fixed p o t e n t i a l s . The program proceeds by moving along one flux l i n e at a time, finding the coordinates of where the f l u x l i n e intercepts the boundary of each c e l l . A c e l l i s the region bounded by consecutive rows and consecutive columns (thus there are NROWS-1 by NCOLS-1 c e l l s i n the model). At each interception point, the d i r e c t i o n of the f l u x l i n e i s determined by i n t e r p o l a t i o n of the f l u x densities at the surrounding nodes. When the coordinates of the l a s t point l i e outside the model boudaries, the current l i n e i s terminated and another one started. "The program asks whether the nodal points are to be plotted on the graph or not. p l o t t i n g the points s u b s t a n t i a l l y increases the p l o t t i n g time. 171 The program BASICFLUXLINES i s the same as HPFLUXLINES except the basic language graphic commands are generated instead of the HPGL commands. The commands are written to FLXLN.BAS without l i n e numbers. The program NUMBER generates the l i n e numbers for the basic program i n FLXLN.BAS and writes the new program with numbers to FLXLNl.BAS. 03-20-88 02:27:12 INPUT.FOR Thu 04-07-88 05:12:20 Pg 1 of 24 1-40 PR OGRAM INPUT C T h i s p r o g r a i s e t s up a file of i n p u t d a t a f o r use by the p r o g r a i C c a l l e d CURRBHT t o c a l c u l a t e nodal p o t e n t i a l s f o r a model of the C a n . I t r e q u e s t s the f o l l o w i n g i n p u t f r o i the c o n s o l e : the n u i b e r C of c o l u i n s and rows, and the address (row and c o l u i n ) and C p o t e n t i a l f o r each node w i t h a f i x e d p o t e n t i a l . There i s no e r r o r C t r a p p i n g so data t u s t be ente r e d c a r e f u l l y . The data i s w r i t t e n C t o INPUT.DAT. C RE AL 1ODPOTIS0) II TBGER RROIS,HCOLS,NPXPOT,RNODB(50),CIODE(50) C HODPOT()...an a r r a y of the p o t e n t i a l s a t the nodes w i t h f i x e d C p o t e n i a l s C HRO¥S...the n u i b e r of rows i n the a r r a y C BCOLS...the n u i b e r of c o l u i n s i n the a r r a y C RIODEO...an a r r a y of the row addresses of the nodes w i t h f i x e d C p o t e n t i a l s C CIODEl)...an a r r a y of the c o l u i n addresses of the nodes w i t h C f i x e d p o t e n t i a l s C C B R I T B ( V ( A \ ) ' ) ' DUMBER OF ROUS?-' R B A D ( V I B H , I 6 ) ' ) NROBS »RITE (VU\ ) ' ) ' HUMBBR OF COLUMNS?-1 READ(','IBH,I6)') HCOLS « R I T B ( * , ' U \ n ' NUMBER OF MODES WITH FIXED POTENTIALS?-' READ!','IBM,16)') NFXPOT 8RITE(*,'(/A/)') ' EITER RODE COORD AND FIXED POTENTIALS' DO 20 N=l,NFXPOT »RITEIVIA\)') ' ROB AND COLUMN?-1 READ!*,*) RNODE(N),CNODE(N) IRITBt«, ,(A\)') ' NODE POTENTIAL?-' READ C M NOD POT IN) POTEN(RNODE(N),CNODE(NI)=NODPOT(N) 20 CONTINUB OPEN 14,FILE='INPUT.DAT1,STATUS?'NEB') BRITE(4,1U) NROVSrNCOLS,NFXPOT BRITE(4,112) (RNODE(N),CNODE(N),NODPOT(N),N=l,NFXPOT) CLOSE(4} 111 FO RNATI1XJI5) 173 03-20-88 02:27:12 IHPUT.FOR Pg 2 Thu 04-07-88 05:12:20 of 24 41-43 112 FO mT(5(2I4,F6.3)/) C EH D 04-04-88 12:40:06 CUREEST.FOR Tbu 04-07-88 05:12:20 P« 3 of 24 1-38 PR OGRAH CURRBHT C T h i s p r o g r a i s o l v e s the n o d a l p o t e n t i a l s of a Hrows by R c o l s C a r r a y g i v e n the f i x e d p o t e n t i a l s a t c e r t a i n nodes. The s o l u t i o n C i s i t e r a t i v e and t e r m i n a t e s when e i t h e r the e r r o r i s l e s s than C a s p e c i f i e d a i n i i u i or the n u i b e r of i t e r a t i o n s exceeds a C l a x i i u i . The i n p u t data can be read f r o i a f i l e IIHPUT.DAT) or C e n t e r e d f r o i the c o n s o l e . R E A L POTEI (30,65), RODPOT (50 ) , ERTOL, ERROR, WAXPOT I I TEGER HROIS,HCOLS,HPXPOT,RHODE|50),CHODE|50),ITER, BAXIT, R,C , CH ARACTBR ARS C POTEH...the n o d a l p o t e n t i a l s C RODPOT...the f i x e d n odal p o t e n t i a l s C E R T O L . . . l i n i i u i e r r o r to t e n i n a t e p r o g r a i C ERROR...the l a x i i u i d i f f e r e n c e i n p o t e n t i a l a t an; node between C i t e r a t i o n s C KAXPOT...the l a x i i u i f i x e d p o t e n t i a l C RCOLS...the n u i b e r of c o l u i n s i n the a r r a j C HROIS... the n u i b e r of rows i n the a r r a y C HFXPOT...the n u i b e r of nodes w i t h f i x e d p o t e n t i a l s C RHODE...an a r r a y of the row addresses of nodes w i t h f i x e d poten C CIODE...an a r r a y of the c o l u i n addresses of the nodes w i t h f i x e d C p o t e n t i a l s C ITER...the c u r r e n t i t e r a t i o n C HAXIT...the l a x i i u i n u i b e r of i t e r a t i o n s b e f o r e t e r i i n a t i o n C R...the row C C . t b e c o l u i n C ARS...character v a r i a b l e f o r s c r e e n i n t e r a c t i o n C IR TEGER CHECK C CHECK...integer f u n c t i o n which checks whether c u r r e n t node i s one C w i t h a f i x e d p o t e n t i a l C CO NMOH R,C,RHODE,CRODE,HFXPOT C s i n g l e c o u o n area f o r g l o b a l v a r i a b l e s C C C see whether d a t a i s i n a f i l e or t o be entered f r o i the c o n s o l e ; 04-04-88 12:40:06 CURRENT.FOR Thu 04-07-88 05:12:20 P« 4 of 24 39-83 C answer i u s t be uppercase. 801 IR ITE|»,'|A\)') ' KRTER DATA BY COHSOLE?-' RB ADC,'(A)') ADS IF |(ARS.RR . ' r i.ARD.(ARS.HE.'R'))GOTO 801 IF |ARS.EQ.'Y')TBBR C read the data f r o i the c o n s o l e B R I T E I V U \ ) ' ) ' NUMBER OF ROBS?-' READ(»,'(BH,I6)') RROBS RRITRI*,'(A\)') ' RUBBER OF COLUMRS?-' READ(*,'(BM,16)') RCOLS WRITBI*.*(A\)'} ' RUBBER OF RODES WITH FIXED POTENTIALS?-' RBAD(»,'IBR,I6)') RFXPOT I R I T E C ' I / A / ! ' ) 1 EHTER RODE COORD ARD FIXED POTENTIALS' DO 20 N-1,NFXPOT BRITE(»,'IA\)') ' ROB ARD COLUBH?-' READI 1, 1) RRODE(R),CRODE(R) B R I T E I V U\)') ' RODE POTERTIAL?-' READ(*,*) RODPOT(R) POTER(ERODE(R),CRODE(I))=RODPOT(H) 20 CORTIRUE BL SE C read the d a t a f r o i a f i l e OPENU,FILE='INPUT.DAT') R E A D ( U l l ) RRORS,HCOLS,RFXPOT READ(4 r112) (RRODE(N),CNODE(N),NODPOT(N),H=l,HPXPOT) ER DIF 111 FO RNAT(1X,3I5) • 112 FO RKAT(5(2I4,F6.3)/) C C g e t the l a x i i u i f i x e d p o t e n t i a l BR ITEI»,'(A\)') ' MAXIMUM POTERTIAL?-' RE A D ( V ) HAXPOT IR I T E C M MAXPOT C get an a c c e l e r a t i o n f a c t o r (1.7 t o 1.9 s e e i s to be a good va l u e ) BR I T E ( I , ' ( A \ ) ' ) ' ACCBLBRATIOR FACTOR?-' RB ADI»,M FACTOR C C s e t a l l nodes to h a l f the t a x p o t e n t i a l ; i t speeds up s o l u t i o n DO 10 R-l,RROBS DO 10 0=1,RCOLS POTRHlR,C)=MAXPOT/2. 10 CO RTIHUE C C put the f i x e d p o t e n t i a l s i n t o the a r r a y DO 21 RFXPOT 04-04-88 12:40:06 CURRENT.FOR Thu 04-07-88 05:12:20 Pg 5 of 24 84-127 POTEH(RHODE(H),CIODE(H))=HODPOT(S) 21 CO 8TIHUE C C C get the u i i i m no. of i t e r a t i o n s ; v a l u e depends on j o u r p a t i e n c e IR I T E C M A U ' l ' MAXIMUM RUBBER NUMBER OF ITERATIONS?-' IE A D I V (BH,I6)') MAXIT C g e t the l a x i i u i e r r o r ; f o r 30 by 65 a r r a y , .00005 worked • e l l »R ITBI»,'IA\I') ' MAXIMUM ERROR?-' RE ADC,*) ERTOL C C s e t i t e r a t i o n counter to z e r o and be g i n IT ER=0 999 CO HTIHOE IT ER=ITER+1 ER ROI--0. C take the average a t the f o u r c o r n e r s R= 1 C= 1 IF (CHECH).EQ.0ITHEH C the node p o t e n t i a l i s not f i x e d TEMPl=POTEHIl,l) TEMP2«(POTEH|l,2)+POTEH(2,l))/2. C t e i p 2 i s the average of the s u r r o u n d i n g nodal p o t e n t i a l s TEMP3=TEMP1-TBMP2 C t e i p 3 i s the d i f f e r e n c e between the nodal poten and the averag POTEH(1,1)=P0TEH(1,1)-FACTOR 4TEMP3 C the a c c e l e r a t i o n f a c t o r g r e a t l y speeds up convergence TEHP4 sABS(TEHP3) TEMP5=ERROR C i f t h i s d i f f e r e n c e i s the l a r g e s t s o f a r , i t b e c o i e s ERROR ERR0R=AMAX1(TEMP4,TEMP5) EH DIF R= 1 C= HCOLS IF (CBBCKl).EQ.0)TBEH TENPl :POTEH(l,HCOLS) TEMP2=(POTEHI1,HCOLS-1)+POTEH(2,HCOLS))/2. TEHP3 STEMP1-TEHP2 POTBH(1,HCOLS)=POTEHI1,HCOLS)-FACTOR'TEMP3 TEMP4=ABS(TEMP3) TEMP5--ERROR ERR0R=AMAX1(TEMP4,TENP5) EH DIF . 04-04-88 12:40:06 CURRENT.FOR Tho 04-07-88 05:12:20 Pg 6 of 24 128-172 R; IROMS C= 1 IF (CHECRO .EQ.01TBES ?RNP1-P07EN(RR0«S,1> TBMP2=(POTER(HRORS-1,1I+POTEH(HROIS,2))/2. TERP3-TEHP1-TENP2 POTER (HROVS, 1) =POTBH (HROIS, 1) -FACTOR 'TEMP 3 TBMP4«ABS|TEMP3) TEMP5--ERROR ERR0R=AMAX1(TEMP4,TEMP5) ER DIP R= RROBS C= RCOLS IF !CHECK().EQ.OTHER TBMPl=POTBH(RROBS,RCOLS) TBHP2=(POTER(RROBS-1,HCOLS)+POTEN(RROBS,HCOLS-l))/2. TEMP3=TEMP1-TEMP2 POTBH(RROBS,RCOLS)=POTEM(RROBS,HCOLS)-FACTOR *TEMP3 TBMP4=ABS(TEMP3) TBMP5=ERROR ERR0R-AMAX1(TEMP4,TENP5) BR DIF C C now the f o u r c o r n e r s are done, do the o u t s i d e rows and c o l u i n s R= 1 DO 910 C=2,RCOLS-l IF (CHECK().EQ.0)THEN TEMPl=POTBH(l,C) TEMP2=(POTER I1,C-l)+POTEB(1,C+l)+POTEH(2,C))/3. TBNP3=TEMP1-TEMP2 POTER(1,0=POTER11,CI-FACTOR »TBMP3 TRHP4=ABSITBNP3) TEMP5 SERR0R BRR0R-AKAX1(TEMP4,TEMP5) ER DIF 910 CO RTIRUE R-- RROBS DO 920 C=2,HCOLS-l IF (CHECK11.EQ.0)THEN TEMPl-POTEH(NROBS,C) TEMP2=(POTEH(RROBS,C-l)+POTEH(HROBS,C+l)+POTBR(HROBS-1, CD/3. TEMP3=TEMP1-TEMP2 POTEB1HROWS,C)=POTBM(HROWS,C)-FACTOR »TEMP3 TEBP4=ABS(TEMP3) TEMP5=ERROR 04-01-88 12:40:06 CURRENT.FOR Thu 04-07-88 05:12:20 P8 7 of 24 173-216 ERR0R=AMAX1(TEMP4,TEMP5) BR DIF. 920 CO RTIHUE C * l DO 930 R=2,HROIS-l IF (CHECK) .EQ.01THEH TEHP1=P0TBH|R,1) TBHP2*(POTEHIR-l,1)+POTER(R+l,1)+POTEH(R,2))/3. TEHP3 STBRP1-TEHP2 P07BRIR,1}=POTEH(R,1}-FACTOR«TEHP3 TEKP4=ABS(TEMP3) TBMP5=BRROI ERR0R=AMAX1(TEMPI,TEKP5) ER DIF. 930 CO RTIHUE C=RCOLS DO 940 R=2,HROHS-l IF (CHBCKO.EQ.O)THEH TBHP1=P0TBR(R,HCOLS) TBHP2=(POTEH(R-l,HCOLS)+POTBH(R+l,HCOLS)+POTEH(R, HCOLS-1M/3. TEMP3=TEMP1-TEMP2 POTEH(R,HCOLS)=POTEH(R,HCOLS)-PACTOR*TEMP3 TEHP4 :ABS(TEMP3) TEMP5=ERROR BRR0R=AMAX1(TEKP4,TEHP5) ER DIF 940 CO RTIHUE C C now the o u t s i d e rows are done, do the r e s t of the nodes DO 950 C=2,RCOLS-l DO 950 R=2,HROIS-l IF(CHECK().EQ.0)TBEH TBKPl=POTEH(R,C) TBHP2*(POTBH(R,C-l)+POTEH(R,C+l)+POTEH(R-l, C)+POTEH(R+l,C))/4. TEMP3=TEMP1-TEMP2 POTEH(R,C)=POTEH(R,C)-FACTOR*TEMP3 TEKP4=ABS(TEXP3) TEHP5-ERROR ERROR=AMAXllTEKP4,TEHP5) BR DIF $50 CO RTIHUE IR I!EI», ,llI,A9 lI5,M fP12.8| ,| ' ITER-',ITBR, ' ERROR-'.ERROR C check l a x i i u i e r r o r and i t e r a t i o n and t e r i i n a t e i f r e q u i r e d IF , 11 ERROR. GT. ERTOL). AHD. I ITER. LT. MAXIT)) THER 01-01-88 12:40:06 CUREEHT.FOR Thu 04-07-88 05:12:20 CHECK Pg 8 of 24 217-242 GOTO 999 C end of l o o p EH OIF C t r i t e the p o t e n t i a l s t o POTRH.TXT OP EH(3,FILE='POTEH.TXT',STATUS*'HEI'I BR ITE(3,150) ((R,C,POTER(R,C),C=l,HCOLS),R=l,HROBS) BR ITE(3,») ' ITER'',ITER,' ERROR=',ERROR 150 FO RMAT15|2I3,F10.6|) ST OP EH D r / t t t t t t t t t t i t t U t l t t i t t t t t t t t t t t t t t l t i t t t t t U t t t t t U t t t t t t t t t i t t t t t t t IB TEGER FUHCTIOR CHECK 0 C T h i s f u n c t i o n determines whether the c u r r e n t node has a f i x e d C p o t e n t i a l . A l l v a r i a b l e s are g l o b a l . IH TEGER R,C,RHODE(50),CHODE(50),RFXPOT CO MKOH R,C,RHODE,CHODE,RFXPOT CH ECK=0 DO 900 R=l,HFXPOT IF I (RHODE (H).EQ.R). AND. (CHODE (H).BQ.O) THEN CHBCK=1 EHDIF 900 CO HTINDE ER D 01-15-88 04:56:52 FMJX.FOR Thu 04-07-88 05:12:20 Pg 9 of 24 1-44 PR OGRAM PLUX1 C T h i s p r o g r a i c a l c u l a t e s the x and ; c o i p o n e n t s of the. e l e c t r i c C f l u x d e n s i t y a t each node of the a m l o d e l . The v a l u e i s the C d i f f e r e n c e b e t i e e n the p o t e n t i a l s a t the nodes on e i t h e r s i d e of C node f o r the x component and the d i f f e r e n c e between the C p o t e n t i a l s a t the nodes above and below f o r the y component. C The p o t e n t i a l s are read f r o i f i l e POTEH.TXT and the f l u x e s C w r i t t e n t o the file DXDY.TXT. The n u i b e r of rows and c o l u i n s can C e i t h e r be en t e r e d f r o i the c o n s o l e or read f r o i the f i l e C IHPUT.DAT. C RE AL POTEH|30,65),DX(30,65),DY|30,65),XCELL,YCBLL ID TEGER iROVS,HCOLS,HFLUXLIRES CH ARACTER AHS C POTEH()...an a r r a y of the n o d a l p o t e n t i a l s C DX()...an a r r a y of the x c o i p o n e n t s of the nodal f l u x C DY(l...an a r r a y of the y c o i p o n e n t s of the nodal f l u x C XCELL...the s p a c i n g between the a r r a y c o l u i n s C YCELL the s p a c i n g between the a r r a y rows C 8ROWS...the n u i b e r of rows i n the a r r a y C HCOLS...the n u i b e r of c o l u i n s i n the a r r a y C R...the c u r r e n t row C C . t h e c u r r e n t row C XC BLL=.l YC ELL«.l C IR I T E C ' I / A / ) ' ) ' PROGRAM TO CALCULATE FIELD FLUXES' C get the n u i b e r of rows and c o l u i n s 201 HR I T B ( ' , ' I A \ ) ' ) ' EHTER DATA BY CONSOLE?-' RE ADC,'(A)') AHS IF ((AHS.HE.T).AHD.(AHS.HE.T))GOTO 201 C IF (AHS.EQ.'Y')THEH IRITE(»,'(A\)') ' HUMBER OF ROUS?-' READ!','UN,161') HROHS iRITE(»,'(A\)') ' HUMBER OF COLUMHS?-' READ(','(BN,I6!'I HCOLS EL SE OPBH|4,FILB='IHPUT.DAT') READ(4,100) HROHS,HCOLS CLOSE(4) EH DIF 100 FO RMAT(1X,3I5) 1B1 01-15-81 01:56:52 FLUX.FOR Thu 04-07-88 05:12:20 Pg 10 of 24 45-91 C OP ES(3,FILE='POTBS.TXT') IE 10(3,101) (|R,C,POTEH|R,C),C=l,HCOLS),R=l,HROIS) CL OSE(3) 101 FO RHAT(5|2I3,F10.6)) C C lot t o c a l c u l a t e the nodal f l u x d e n s i t i e s C C do the f o u r c o r n e r s f i r s t 1= 1 C= 1 DX IR,C)=P0TEH(1,2)-P0TBH(1,1) DY (R,C)=POTBR(2,l)-POTEH(l,l) C R= 1 O HCOLS DX IR,C)=P0TEH11,HCOLS)-POTEHI1,HCOLS-1) DY (R,C)=POTRH(2,HCOLS)-POTEHI1,HCOLS) C R= HROIS C= 1 DX (R,C)=POTRR(HROIS,2)-POTBR|HROIS,l) DY IR,0-POTEH(HROHS,1)-POTEH(HROHS-1,1) C R= HROHS C= HCOLS DX (R,C)-POTEH(HROIS,HCOLSI-POTEH(HROIS,HCOLS-1I DY (R,C)=POTEH(HROIS,HC0LS)-POTEH(HROIS-1,HCOLS) C C now do the o u t s i d e rows and c o l u i n s R= 1 DO 202 C-2,HCOLS-1 DX (R,C) = (POTBH(R,CU)-POTBH(R,C-l))/2. DY (R,C|=POTBH(R+l,C)-POTEH(R,C| 202 CO RTIHUE C R= HROIS DO 203 C=2,HCOLS-l DX (R,C)=(POTRH(R,C+l)-POTBH(R,C-l))/2. DY (R,C)=POTEH(R,C)-POTEH(R-l,C) 203 CO RTIHUE C C= 1 DO 204 R-2.HR0IS-1 DX (R,C)=P0TEH(R,C+1)-P0TEH(R,C) DY (R,C) = (POTBH(RU,C)-POTBH(R-l,C))/2. 204 CO RTIHUC 01-15-88 04:56:52 PLUX.FOR Thu 04-07-88 05:12:20 Pg 11 of 24 92-113 C O ICOLS DO 205 R=2,HROIS-l DX IR (C)=P0TBH|R,C)-P0TRH|R,C-1I DY (R,C}=lPOTEHlR+l,C)-POTEH|R-l,C))/2. 205 CO ITIROE C C not do the r e s t of the nodes DO 206 C-2,HCOLS-l DO 206 R=2,RROIS-l DX (R,C) = lPOTEHIR,CU)-POTBHlR,C - l))/2. DY |R,C)=IPOTEB|R+l,C)-POTBM(R-l 1C)l/2. 206 CO ITIBUE C C w r i t e the data t o DXDY.TXT OP BR(8,FILE-'DXDY.TXT',STATUS*'RBH') «R ITE(8,102) ((R,C,DX(R (C),DY|R,C),C=1,RC0LS),R=1,HR0*S| CL OSEI8) 102 FO RRAT|4(2I3,2F10.6I) ST OP 'FLUXES IR FILE DXDY.TXT' C BR D 01-15-88 04:00:20 HEWHPPL.FOR Thu 04-07-88 05:12:20 Pg 12 of 24 1-34 PR OGRAH HPFLOXLISES C T h i s p r o g r a i generates the f l u x l i n e s f o r an a n l o d e l g i v e n the C f i l e of n o d a l f l u x e s IDXDY.TXT). The i n p u t data can be e n t e r e d C f r o i the c o n s o l e or read f r o i a f i l e (IHPUT.DAT). The output f o r C the HP p l o t t e r i s w r i t t e n t o a f i l e (FLXLH.HP) which can then be C c o p i e d t o the p l o t t e r a f t e r r u n n i n g HPTI. C C The nodes f r o i which the f l u x l i n e s are t o e i a n a t e l u s t be the C f i r s t nodes s p e c i f i e d as nodes w i t h f i x e d p o t e n t i a l s i n IIPUT.DAT. C The n u i b e r of f l u x l i n e s and whether the l i n e s are t o go f r o i a C p o s i t i v e p o t e n t i a l to n e g a t i v e p o t e n t i a l or v i s a v e r s a • u s t be C en t e r e d f r o i the c o n s o l e . C C To c r a i the p r o g r a i i n t o 64k p l a c e s l i m i t a t i o n s on the the n u i b e r C of f l u x l i n e s . More l i n e s can be added by i n c r e a s i n g the f i r s t C dimension of XPLUXO and YPLOXO and r e d u c i n g the second p r o v i d e d C the n u i b e r of x and y p o i n t s i n each l i n e doesn't exceed the C second dimension. C RE AL DX(30,65),DY(30,65),RODPOT(50), 1 XCBLL,YCELL,XHEH(YHE«,XOLD,YOLD,X,Y,XSIZR,YSIZE, 2 YFLUX(15,200),MAXPOT,XFLUX(15,20O),XLIHE(15) IH TEGBR HROHS,HCOLS,HFXPOT,RHODE(50),CHODR(50), 1 HFLUXLIHES,IMAX,HFLUXPOIHTS(50),R rC,FIRSTCOL,LASTCOL, FIRSTROH CH ARACTER AHS LO GICAL HLIHE,HEHLIHB,FIRISH,PLTOMIH C DX(),DY()...the c o i p o n e n t s of the nodal f l u x C IODPOTO...the p o t e n t i a l a t the nodes w i t h f i x e d p o t e n t i a l s C XCELL.YCELL . . . X and y d i s t a n c e between nodes C XHBR,YHEH...the c o o r d i n a t e s of the end of the f l u x l i n e C XOLD,YOLD...the c o o r d i n a t e s of the p r e v i o u s end of the f l u x l i n e C X , Y . . . c u r r e n t c o i p o n e n t s of nodal f l u x C XSIZE,YSIZE...x and y d i s t a n c e between nodes 01-15-18 04:00:20 HEHHPFL.FOR Thu 04-07-88 05:12:20 Pg 13 of 24 35-75 C XFLOX(),YFLUX(}...arrays of the p o i n t s i n the f l u x l i n e s C RROBS,HCOLS...nuiber of rows and c o l u i n s i n the l o d e l C HFXPOT...nuiber of nodes w i t h f i x e d p o t e n t i a l s C RHODE)),CHODEO...location of the nodes w i t h f i x e d p o t e n t i a l s C HFLUXLIHES...nuiber of f l u x l i n e s C IMAX...the l a x i i u i n u i b e r of p o i n t s i n a f l u x l i n e 2*1IR0IS+ HCOLS) C HFLUIPOIHTS()...an a r r a y of the n u i b e r of p o i n t s i n each f l u x l i n e C R . C . t h e c u r r e n t row and c o l u i n C HLIHB...true i f the end of a f l u x l i n e i s between two c o l u i n s C IERLIXE t r u e i f the end of a f l u x l i n e i s s w i t c h i n g f r o i b e i n g C between two rows to between two c o l u i n s or v i s a v e r s a C FIMISH...true i f the c u r r e n t f l u x l i n e has f i n i s h e d C PLTOMIH...true i f the f l u x l i n e s go f r o i a p o s i t i v e p o t e n t i a l t o a C n e g a t i v e one XC E L L - . l YC E L L - . l C get the n u i b e r of rows and c o l u i n s , e t c . 200 IR I T E I V ( A \ ) ' ) ' EHTER DATA BY CONSOLE?-' RE ADC,MA)') AHS BR ITB<»,») IF ((AHS.HE.'Y'l.AND.IAHS.HE.'R'HGOTO 200 IF (AHS.EQ.'Y')TBEH BRITB(«,'(A\)'I ' RUBBER OF ROBS?-' R E A D ( V ( B B , I 6 ) ' ) RROBS B R I T R ( V ( A \ ) ' ) ' BOMBER OF COLUMRS?-' READ(','(BB,I6)') RCOLS B R I T E I V (A\)') ' HUHBER OF HODES WITH FIXED POTERTIALS?-' R E A D ( V ( B H , I 6 ) ' ) RFXPOT WRITE(*,'(/A/)') ' EHTER RODE COORD ARD FIXED POTENTIALS' DO 210 1=1,HFXPOT H R I T E I V I A U ' ) ' ROB ARD C0LUKH?-' READ 1 *,*) RNODEIN),CHODEIR) B R I T E I V ( A \ ) ' ) ' RODE POTERTIAL?-' R E A D ( V ) RODPOTIR) 210 CORTIHUE EL SE OPEH(4,FILE:'IBPUT.DAT') READU,101) HROBS,HCOLS,RFXPOT RBAD(4,102) (RHODE(H),CHODE(R),RODPOTIR),R=l,RFXPOT) ER DIF . 01-15-88 04:00:20 HEVHPFLFOR Thu 04-07-88 05:12:20 Pg 14 of 24 76-122 101 FO RHAT|1X,3I5) 102 FO RMAT|5(2I4,F6.3)/) C C g e t the n u i b e r of f l u x l i n e s and d i r e c t i o n of f l u x l i n e s IR ITE(»,'(A\)') ' NUMBER OF FLUX LIHES?-' EE AD C M HFLUXLIHES IR ITE | » ,M IR I T E I V U M ' ) ' FIRST COLUMH THAT FLUX LIHES START AT?-' RE ADC,M FIRSTCOL IR I T E ( V ) IR I T E | ' , ' ( A \ ) ' I ' LAST COLUMI THAT FLUX LIHES START AT?-' RE AD ISM LASTCOL IR ITE (* ,M IR I T I I V U M ' ) ' ROI THAT FLUX LIHES START OH?-' RE AD C M FIRSTROI IR 1TE(«,M 220 IR I T E I ' , ' ( A \ ) ' ) ' FLUX LIHES FROM PLUS TO MINUS?-' RE AD (VU) ' ) AHS IF I I A H S J R . T U I D . I A N S . H E . T I I G O T O 220 IF (AHS.EQ.'¥1)THEM PLTOHIH=.TRUE. EL SE PLT0MIH-.FALSE. EH DIF C C read i n the nodal f l u x e s OP EH(8,FILE :'DXDT.TXT') RE ADI8.103) ((R,C,DX(R,C),DY(R,C),C S1,HCOLS),R 31,HROIS) 103 FO RMAT(4(2I3,2F10.6)) C CA LL LIHESTART(DX,DY,XCELL,HFLUXLIHES,XLIHE, 1 FIRSTCOL,LASTCOL,FIRSTROI) C C Hoi to c a l c u l a t e the f l u x l i n e s C IR ITB(«,M ' START' DO 230 H-l,HFLUXLIHES C do s o i e i n i t i a l i z a t i o n XSIZE=XCELL YSIZE=YCELL XOLD=XLINE(H) YOLD=(FIRSTROI-l)*YCELL C XTBMP-AIHT(XLIHE(H)/XCELL) ICOLl=HIHT(XTEMP)+l ICOL2=ICOLl+l Xl-DX(FIRSTROI,ICOLl) 01-15-88 04:00:20 HEHHPFL.FOR Tbu 04-07-88 05:12:20 Pg 15 of 24 123-166 Yl=DY(FIRSTROW,ICOLl) X2=DX(FIRSTROH,ICOL2) Y2=DY(FIRSTROB,ICOL2) X C O L H A I B T (XLIHE U)/XCBLL))»XCELL XC0L2=XC0L1+XCELL X=(XLIHE(H)-XCOLl)* U2-X1)/(XCOL2-XCOL1)+X1 Y=(XLIHE(«)-XCOL1)MY2-Y1)/|XCOL2-XCOL1)+Y1 C FIHISH=.FALSE. BLIHE=.FALSE. IMAX=2«MCOLS+2*SROiS 1=1 XFLUX|H,l)=XOLD YFLUX(H,l)=YOLD C b e g i n the loop to c a l c u l a t e the p o i n t s f o r each l i n e 240 COHTIHUE C g e t the next p o i n t at the end of the f l u x l i n e CALL HEHCOORDSlX,Y,XOLD,YOLD,XHBH,YHEH,HBHLIHE,XSIZE, YSIZE, 1 PLTOMIN) C see i f t h i s l i n e i s f i n i s h e d CALL EHDLIHEI FINISH,HLIHB,XNBB,YNEB,HCOLS,HROWS,I,INAX, 1 XCELL,YCELL,N,HFLUXPOIHTS) C update the v a r i a b l e s f o r another t i i e t h r u the loop CALL UPDATE(I,H,HLHE,XHBB,YHBH,XOLD,YOLD,XSIZE,YSIZE, 1 XCELL,YCELL,NEILINE,XFLUX,YFLUX,DX,DY,X,Y,FINISH) IFIFIHISHIGOTO 250 GOTO 240 250 COHTIHUE WRITE(*,M ' FLUXLIHE HUHBER ',N 230 CO HTIHUE C tr ite ( * , M ' not g e n e r a t i n g HPGL p r o g r a i ' C C c a l l the p r o g r a i t o generate the output f i l e CA LL BASICINCOLS,MOWS,NFXPOT,CHODE,RNODE,NFLUXLINES, 1 XFLUX,YFLUX,HFLUXPOIHTS) ST OP EH D C C i t t t t M t t M t t i t i t t t t i t t t t i t i t i t t t i t t t i t t t i i t i t t t U t t i t t t i t t i t t t t t i t t t i SU BROUTIHE LINESTARTIDX.DY,XCELL,HFLUXLIHES,XLIHE, 1 FIRSTCOL,LASTCOL,ROW) RE AL FLUXDELTA,FLUXBBGIH,FLUXEHD,FLUXIHC,FLUXCUR,FLUXHEX, XLIHE|«), 1 X,DX[30 ,M,DY(30,M,XCELL 01-15-88 04:00:20 HEiHPFL.FOR Thu 04-07-88 05:12:20 Pg 16 of 24 167-209 IN TEGER COLCUR,COLNEX.FIRSTCOL,LASTCOL,HFLUXLIHES,RON C CO LCUR=FIRSTCOL CO LHEX=COLCUR+l FL UXBEGIH=SQRT(DX(ROW,FIRSTCOL)* *2+DY(ROM,FIRSTCOL)* * 2) FL UXEHD=SQRT(DXtROW,LASTCOL)**2+DY(ROW,LASTCOL)**2) FL OXMBX-SQRTI DX(ROW,COLHEX)" 2+DY(ROW,COLNEX)**2) FL UXCUR=0. FL UXIHC=0. DO 3 J=FIRSTCOL+l,LASTCOL FLUXIHC=FLUXIHC+SQRT (DX (RON, J) * * 2+DY (RON, J) " 2 ) 3 CO RTIHUE FL UXHC=FLUXINC/UPLUXLIHES-1) FL UXDELTA-FLUXIHC XL IHE(l)-XCELLMFIRSTCOL-l) H= 2 IT =0 C 1 CO RTIHUB IF (H.GT.HFLUXLIHES)GOTO 2 IF (ABS(PLUXHEX-FLUXCUR).LT.ABS(FLDXDELTA))TEEM COLCUR=COLHRX C0L8EX=C0LHEX+1 PLUXDBLTA SFLUXDELTA-(FLUXHEX-FLDXCUR) PLUXCUR=0. FLUXHRX=SQRT(DX I RON, COLHEX I" 2+DY (ROW, COLHEX) " 2 1 EL SE X=((FLUXDELTA+FLUXCURI/FLUXHBX)'XCELL XLIHE(R):(COLCUR-l)'XCELL+X H=H+1 FLUXCUR=FLUXCUR+FLUXDELTA PLUXDELTA=PLUXIHC EH DIF IT =IT+1 GO TO 1 2 CO RTIHUE ER D Q t t t t i i l t t t t l i i t t l t t t t t t U l t l K t t l l l t t l t t t t i t t t l t l t i U t t t t l i t t i t l t t i i t t SU BROUTIHE HEWCOORDS(X,Y,XOLD,YOLD,XHEH,YHEH,HEWLIHE, XCELL,YCELL, 1 PLTOHIH) C T h i s s u b r o u t i n e f i n d s the c o o r d i n a t e s f o r the next p o i n t i n the C f l u x l i n e . The passed p a r a i e t e r s are updated i n the sub UPDATE. C The ne* c o o r d i n a t e s are i n XHEW and YHEW. I f the l i n e 01-15-88 04:00:20 KEWHPFL.FOR Thn 04-07-88 05:12:20 Pg 17 of 24 209-251 changes f r o i C t e r i i n a t i n g betveen two c o l u i n s to t e r i i n a t i n g between two rows C or v i s a v e r s a , HEfLIME i s t r u e . C RE I L S,T,XOLD,TOLD.XMEW,YHEV,XPRIHB,TPRIMB,XSI6M,TSIGH. MAGX, 1 MAGY,FACTOR IM TEGER IYREw.IYOLD LO GICAL MBVLIHE,PLTOMIN C X,Y...the i n t e r p o l a t e d f l u x v a l u e s a t XOLD and YOLD C XOLD,YOLD...the c o o r d i n a t e s of the l a s t p o i n t on the l i n e C XREW,YMEtf...the c o o r d i n a t e s of the next p o i n t on the l i n e C XPRIME,YPRIME...the d i f f e r e n c e between XOLD and XNEV, e t c . C XSIGH,YSIGS...the s i g n of X and Y C MAGX,MAGY...the l a g n i t u d e of X and Y C F A C T O R . . . i n t e r i e d i a t e product C IYHEw\IYOLD... i n t e g e r p o r t i o n of YKEH and YOLD C IEVLIHB...true i f l i n e t h a t t e r m i n a t e s between two c o l u i n s now C t e n i n a t e s between two rows or v i s a v e r s a C PLTOMIN...true i f f l u x l i n e s go f r o i p o s i t i v e pot to n e g a t i v e pot C i n i t i a l i z a t i o n HE If LINE?. FALSE. XP RIME=0. YP RIME=0. XN B*"=0. YN EH=0. C IF (PLTOMIN)TBEN X=-X Y=-Y EN DIF XS IGH=SIGH(1.,X) YS IGN=SIGN(1.,Y) C C be g i n f i n d i n g next p a i r of c o o r d i n a t e s f o r the l i n e IF (X.EQ.O.)THEN C l o v e v e r t i c a l l y o n l y YPRIME=YCELL EL SEIF(Y.EQ.0.)THEN C i o v e h o r i z o n t a l l y o n l y XPRINB=XCELL EL SE , 01-15-88 04:00:20 HEIHPFL.FOR Thu 04-07-88 05:12:20 C c a l c u l a t e s o i e i n t e r i e d i a t e v a l u e s HAGX=ABS(X) RAGY=ABS(Y) FACTOR5AMOD(TOLD,YCELL) C IFIYSIGI.GT.O.ITBER PACTOR=YCBLL-FACTOR ERDIF C check t o see i f l i n e w i l l t e r i i n a t e on a "HEHLIHB" IF(HAGX.OE.HAGY)THEH C i t l a y not t e r i i n a t e on a new l i n e YPRIME=XCELL*MAGY/MXGX YHB»=YOLDtYSIGH»YFRIMB IYHEI=IHT(YHEM/YCELL) IYOLD=IHTIYOLD/YCELL) IF IIYHEM.HE.IYOLD)THEN C i t does t e r i i n a t e on a new l i n e HEHLIHB 5.TRUB. XPRIME=FACTOR*XCELL/YPRIME YPRIME=FACTOR ELSE XPRIKR=XCBLL ERDIF ELSE C i t does t e r i i n a t e on a new l i n e RBRLIRB 5.TRUE. YPRIME=FACTOR XPRIMB5FACTOR*HAGX/HAGY ERDIF EH DIF C C now c a l c u l a t e the new c o o r d i n a t e s XH EH'XOLD+XSIGH'XPRIHE YR ER-YOLD+YSIGH'YPRIHE RE TURH EH D C C C i t t t M t U M O t f t t t t t M t t M t f t t M t t t t M i t t t t t t t t t i U i M M t t i t t t t * t t l l t t t t SU BROUTIHE EHDLIHB(FIHISH,HLIHE,XREH,YHEH,HCOLS,HROHS,I, IHAX, 1 XCELL,YCELL,H,HFLUXPOIHTS) C T h i s s u b r o u t i n e checks to see i f the c u r r e n t f l u x l i n e i s C f i n i s h e d . I t does t h i s by c h e c k i n g the l a s t c a l c u l a t e d c o o r d i n a t e s C of the . l i n e . I f the x c o o r d i n a t e i s l a r g e r than the Pg 18 of 24 252-295 01-15-88 04:00:20 HB1HPFL.FOR Thu 04-07-88 05:12:20 Pg 19 of 24 295-336 • a x i i u i v a l u e C of x (i.e.,HCOLSOCELLI or s i a l l e r than 0, the l i n e i s f i n i s h e d . C S i i i l a r l y f o r the y c o o r d i n a t e . I f the l i n e i s f i n i s h e d , s o i e C v a r i a b l e s are updated. I t a l s o check whether the t a x i i u i n u i b e r C of p o i n t s p e r i i t t e d f o r a f l u x l i n e has been exceeded. C C RE AL ZHEI,YHEB,XCELL,TCELL,XKAX,YHAX, XCH,YCH IR TEGER HCOLS,RROHS,I,IKAX,H,HFLOXPOIHTS(») LO GICAL FINISH,HLIRB C l o s t v a r i a b l e s are as d e f i n e d above C XCH,YCH...local v a r i a b l e s f o r XRBH and YHER C I...the n u i b e r of p o i n t s i n the c u r r e n t f l u x l i n e C M...the c u r r e n t f l u x l i n e C XR AX=(HCOLS-l)'XCELL YM AX*(HROiS-1)'YCELL C i f the l a s t c o o r d i n a t e c a l c u l a t e d f o r the f l u x l i n e i s between C two c o l u i n s , then the x and y v a r i a b l e s i u s t be switched IF (HLIHE)THEH XCB=YHBR YCH SXHEH EL SE XCH=XBEH YCH-YNBH EH DIF C C see i f the l a x and t i n v a l u e s of x and y have been exceeded IF IXCB.LE.0..OR.XCH.GE.XMAX.OR.YCH.LE.0..OR.YCH.GE.YNAX) THER C i f s o , l i m i t the c o o r d i n a t e t h a t was exceeded PIHISH=.TROE. IF(XCH.LE.0)TBER XCH=0. ELSEIF)XCH.GE.XKAX)TBEH XCH=XKAX ELSEIFIYCB.LB.OTHER YCH=0. ELSE YCH=YMAX ERDIF C s w i t c h the x and y c o o r d i n a t e s i f necessary IFIHLIHBITHEN 01-15-88 04:00:20 Thu 04-07-88 05:12:20 HBHHPFL.FOR Pg 20 of 24 337-375 YHEV-XCB XHEH=YCH ELSE XHEM=XCH YHBN=YCH EHDIF ER DIP C see i f the f l u x l i n e has too i a n j p o i n t s IF t.HOT.FIHISHITHBN IF|I.GB.IMAX)PIHISH=.TRUE. ER DIF C update the n u i b e r of p o i n t s i n the f l u x l i n e i f l i n e i s f i n i s h e d IF (FIRISH)THER HFLUXPOIHTS(H)=I+l ER DIF C RE TURN EN D C C C i < t t * t t t t i t t t t t i t t t < t t t t * t t t t t l t t t t i t t t t t t t i t t t t t i t t l t t t t t t t < i I t t t t t i t SU BROUTINE UPDATEII,H,HLIHB,XHBH,YH8H,X0LD,Y0LD,XSIZE, YSIZE, 1 XCELL,YCELL,HEWLIIE,XFLDX,YPLUX,DX,DY,X,Y,FINISH) C T h i s s u b r o u t i n e updates the v a r i a b l e s b e f o r e e n t e r i n g the l o o p C a g a i n or e x i t i n g f r o i the l o o p . I t i n t e r p o l a t e s a v a l u e f o r the C f l u x a t the nev c o o r d i n a t e s f r o i the f l u x v a l u e s a t the n e a r e s t C nodes. C RE AL XHBH,YNBH,XOLD,YOLD,XSIZB,YSIZB,XCELL,YCELL rXFLUX(15, «), 1 YFLUXI15,»),DXI30,«),DY|30,»),TEMP,X,Y IR TEGER I,H,R1,R2,C1,C2 LO GICAL HLIHE.NEKLINE,FINISH C the v a r i a b l e s are as d e s c r i b e d above C C i n c r e a s e the n u i b e r of p o i n t s i n the l i n e b; one 1= 1+1 C i f c u r r e n t c o o r d i n a t e i s between two c o l u i n s , swap x and y va l u e s IF (HLIHE)TBBH TEMP=YREH 01-15-88 04:00:20 HEWHPFL.FOR Tbu 04-07-88 05:12:20 Pg 21 of 24 376-422 YHEi=XNBfl XREI-TEKP ER DIF C update the a r r a y of c o o r d i n a t e s f o r the f l u x l i n e s XF LUX(H,I)=XNBB YF LUX|N,I)=YHBN C see i f l i n e i s f i n i s h e d and q u i t IF (FIRISH) GOTO 260 C see i f f l u x l i n e not t e r i i n a t e s on a n o u a l l i n e IF |HEBLINE)HLIHB=.NOT.BLINE C n o i c a l c u l a t e f l u x v a l u e s a t the new c o o r d i n a t e s IF (HLINE)THEM R1=NIHT(YNEI7 YCELL)H C1=INT|XNEB/XCBLL)U C2=C1+1 X=DX(Rl,Cl)+(DX(Rl,C2|-DXlRl,Cl))*ANODlXNRB,XCELL)/XCELL Y=DY|R1,C1)+(DY(R1,C2)-DY(R1,C1))'AM0D(XHEN,XCELL)/XCELL EL SE R1=INTIYNBB/YCELL)+1 R2=R1+1 CMIHT(XNEB/XCELL)+1 X=DX(R1,C1)+(DX(R2,C1)-DX(R1,C1))*AH0D(YHE»,YCELL)/YCELL Y=DY(R1,C1)+(DY|R2,C1)-DY|R1,C1))*AH0D(YNEN,YCELL)/YCBLL EN DIF C C s w i t c h x and y v a r i a b l e s i f ne c e s s a r y IF (BLINE)TBBN TBKP=X X=Y Y=TENP XOLD=YNER YOLD=XNBB XSIZE=YCELL YSIZE=XCELL EL SE XOLD=XHBW YOLD sYNEN XSIZE=XCELL YSIZE :YCELL EN DIF 260 CO NTINUE RE TURN EN D 01-15-88 04:00:20 HEVHPFL.FOR Thu 04-07-88 05:12:20 Pg 22 of 24 423-466 C t U t t t i i O t t M t t U t t M M t i t i t i t t t t t M t t t i t i t i t t i t t t t t U t i U t t t t i t t t t t t SU BROUTIHE BASIC (NCOLS,HROHS,HFXPOT,CHODE,RHODE,IFLUXLINES, 1 XFLUX,YFLUX,HFLUXPOIHTS) C T h i s s u b r o u t i n e produces a file of HPGL g r a p h i c c o u a n d s f o r C the HP p l o t t e r s and w r i t e s t h e i t o a fi l e ( f l z l n . h p ) . I t asks C whether the model nodal p o i n t s are to be i n c l u d e d on the p l o t . C RE AL XFLUX(15,M,YPLUX(15,») IH TEGER HCOLS,HROHS,HFXPOT,CHODE11),RHODE(*),HFLUXLIHES, 1 NFLUXPOIHTS('),R,C LO GICAL DOTS C DOTS...true i f nodal p o i n t s are t o be i n c l u d e d on the p l o t CH ARACTER AHS C C see i f n o d a l p o i n t s are to be p l o t t e d 270 HR I T B I S ' I A U ' ) ' POTEHTIAL POIHTS OH PLOT?-' RE ADC,MA)') AHS IF KANS.HE.'Y 1) .AHD. (AHS.HB.'H'MGOTO 270 IF (AHS.BQ.'Y'ITHEH DOTS*.TRUE. EL SE DOTS=.FALSE. EH DIF C OP EH 19,FILE 5'FLXLH.HP',STATUS-'HEH') C C s e t the d e f a u l t parameters and the p l o t t e r s c a l e IR ITB|9,104) 1 DF; SC 0,10300,0,7560;' 104 FO RHATIA23) C C i n i t i a l i z e the o r i g i n c o o r d i n a t e s and send the pen t h e r e IC OLOR=l IX 0=1280 IY 0=2480 IR ITB(9,«) ' SP1;' HR IT8I9.105I ' PU ',1X0,',',IYO,'; PD;' 105 FO RHAT(A4,I5,A1,I5,A5) C C draw the o u t l i n e of the a n IC OLOE=l IX U=(HCOLS-1)»100+1XO IY H=(HROIS-1)»100+IYO IR ITE(9,106) ' P A ' . I X O . W I Y B V ; ' IR ITBI9.106) ' PA'.IXH . ' . ' . I Y H , ' ; ' 194 01-15-88 04:00:20 KESHPFL.FOR Pg 23 Thu 04-07-88 05:12:20 of 24 467-513 BR ITBI9.106) ' P A V X K . W I Y O , ' ; ' BR IT i ( 9 , 1 0 6 ) ' PAMXO.'.'.IYO, BR ITB|9,M 'PU;' 106 FO RMAT(A3,I5,A1,I5,A1) C C draw the nodal p o i n t s ; put 4 c o u a n d s on each l i n e IF (DOTS)TBER / IC OLOR=l IC OUBT=l DO 280 R=l,RROBS DO 280 C=l,HCOLS IR=IR-1)»100WO IC=(C-1)'100+IXO IF(ICOUHT.BQ.4)TBEH BRITE(9,105) ' PU ',IC,'.',IR,'; PD;' ICOUHTM ELSE BRITE(9,105) • PU \ I C , \ ' , I R , ' ; PD;' ICOURT=ICOURT+l ERDIF BRITBI9,*) ' PU;' 280 CO RTIRUE ER DIF C C draw i n the e l e c t r o d e s IC OLOR=l IX 2=|CBODE(1)-1I*100+IXO+10 IY 2=(RHODBll)-l)»100+IYO+10 BR ITEI9,*) 'SP3;* BR ITEI9.105) ' PU ',1X2,',',IY2,*; PD;' DO 290 1=2,RFXPOT 1X1=1X2 IY1=IY2 IX2=(CRODE(I)-1)»100+IXO+10 IY2=|RRODE(I)-l)*100tIYOU0 IF((RMODE(Z).EQ.RHODB(I-l)) 1 ((CRODE(I)-CBODE(I-l)).LE.l))TBEH BR ITEI9.106) ' P A , , I X 2 , ' , , , I Y 2 , ' ; ' ELSE BRITE(9,105) ' PU ',1X2,',',IY2,'; PD;' ERDIF 290 CO RTIRUE C C draw i n the e l e c t r o d e s IC OLOR=l IX 2=|CHODBll)-l)»100+IXO+20 IY 2=|RRpDE(l|-l)'100+IYO+20 01-15-81 04:00:20 HEWHPFL-FOR Thu 04-07-18 05:12:20 Pg 24 of 24 514-547 HR ITR|9,M *SP3;* HR ITEI9.105) 1 PO ',1X2,',',IY2,'; PD;' DO 300 1=2,HFXPOT 1X1=1X2 IY1=IY2 IX2=(CHODE|I)-1I»100+IXO+20 IY2=(RHODE|I)-1)«100+IYO+20 IF((RHODE(I).EQ.RHODBII-l)).AHD. 1 I(CHODE(I)-CNODE(I-l)).LE.l)1THEN HR ITEI9.106) ' PA'.1X2,',',IY2,';' ELSE HRITE(9,105) ' PU \ I X 2 , ' , \ I Y 2 , ' ; PD; 1 ERDIF 300 CO RTIHUE C C now d r a t a l l the f l u x l i n e s IC OLOR-2 DO 310 1=1,HFLUXLIHES IX2=XFLUX(I,1)»1000+IXO IY2=YFLUX|I,1)*1000+IYO IRITB|9,105) • PU M l J . ' . ' . m , 1 ; PD;' DO 320 J=l,HFLUXPOIHTS(I)-l R =JU 1X1=1X2 IY1=IY2 IX2=XFLUX(I,K)'1000+IXO IY2=YFLUX(I,R)»1000+IYO H RITEI9.106I ' P A ' . m / . ' . m , ' ; ' 320 CO RTIHUE HR ITE(9,») 'PU 0,0;' 310 CO RTIHUE CL OSEI9) RE TURH EH D 196 APPENDIX E: VALIDATION OF THE USE OF AN ULTRASOUND TRANSDUCER OVER THE BRACHIAL ARTERY TO DETECT BLOOD FLOW PAST THE CUFF In this study, the onset of the flow of blood past the occlusive cuff was detected with an ultrasound sensor placed by the d i s t a l edge of the cuff over the brachial artery. The ultrasound equipment was the Hoffmann-La Roche Inc. Arteriosonde 1010 Blood Pressure Instrument. To test the v a l i d i t y of t h i s method, the point of the onset of blood flow was determined concurrently with and compared with each of the four alternatives l i s t e d below. Not only were the methods compared for the coincident occurence of the f i r s t detected pulse as the cuff pressure was decreased, but also when applicable for changes i n basal values that might indicate a slow infusion of blood into the forearm past the cuff. In each case the outputs were displayed on the Hewlett-Packard Inc. model 7754A chart recorder. The f i r s t a l ternative was to detect blood flow by measuring the impedance of the forearm with an impedance plethysmograph. With th i s method, the f i r s t pulse i n d i c a t i n g the resumption of blood flow consistently occurred at the same time as that from the Arteriosonde. The value of the basal impedance seemed to change slowly with the occlusive cuff pressure held just above the occlusion pressure, but no sustained decrease.in impedance was observed. 197 The second alternative was to measure blood flow by monitoring the pressure i n a second cuff placed around the forearm and i n f l a t e d to 20-30 mmHg. Again, the f i r s t discernable pulse appeared at the same moment as that from the Arteriosonde. Similar to the basal impedance, the basal cuff pressure d r i f t e d with time but never showed a sustained increase with the occlusive cuff pressure marginally above the occlusion pressure. The t h i r d a lternative was to measure the blood flow i n a finger with a photoplethysmograph. If the occlusive cuff pressure was taken above the occlusion pressure and slowly decreased, the f i r s t pulse from the Arteriosonde occurred several pulses before the f i r s t from the photoplethysmograph. However, i f the pressure was quickly decreased to a pressure just above occlusion and then slowly decreased, the occurence of pulses from the two devices was coincident. It may be that after occlusion the a r t e r i a l blood slowly traverses the c a p i l l a r y bed into the venous system and the capacity of the drained a r t e r i a l system becomes s u f f i c i e n t to delay the appearance of any appreciable blood flow i n the fingers. The same e f f e c t was noticed with the fourth alternative, another doppler, the Minidoplex D500, monitoring blood flow i n the r a d i a l artery at the wrist. Detection of blood flow there was also delayed i f the cuff pressure was held above occlusion for very long. 198 A l l of the alternatives gave some supporting evidence that the method used was indeed v a l i d . 199 APPENDIX F: SOURCES OF NOISE AND INTERFERENCE ENCOUNTERED IN THE STUDY F l . RESPIRATIONARTIFACT Respiration r e s u l t s i n two kinds of interference, one additive and one m u l t i p l i c a t i v e . The additive component i s caused by motion as a r e s u l t of r e s p i r a t i o n . With the subject seated i n a chair with her elbow resting on a desk, r e s p i r a t i o n w i l l move her shoulder r e l a t i v e to her elbow. This motion involves the musculature of the upper arm and w i l l change the impedance of the tissue beneath a cuff placed about the upper arm. The m u l t i p l i c a t i v e component appeared as amplitude modulation of the impedance pulses. Respiration affects the a r t e r i a l blood pressure by more than one mechanism and can change the blood pressure by as much as 20 mmHg with each respiratory cycle during deep r e s p i r a t i o n [54] . As the a r t e r i a l pressure varies, the incursion of a r t e r i a l blood into the tissue beneath the tourniquet cuff varies, causing the modulation of the impedance pulses. This modulation i s , however, a legitimate r e f l e c t i o n of the changes i n the cuff pressure required to collapse the a r t e r i e s and should not be treated as noise. r The frequency content of the additive noise i s low since the average r e s p i r a t i o n rate i s about 0.2 Hz [55]. Its 200 magnitude can be large, however. F i g . F l shows an example of re s p i r a t i o n noise. The frequency content of the modulated impedance pulses i s very close to that of the unmodulated impedance pulses since amplitude modulation re s u l t s i n frequency side bands displaced from the c a r r i e r frequency (in this case, the frequency content of the impedance pulses) by the frequency of the modulation (the r e s p i r a t i o n Waveform) which i s to a f i r s t approximation sinusoidal. F i g . F2 shows an example of this noise. F2. MUSCLESPASM Many subjects found the muscle of the limb to which the tourniquet cuff was applied became i r r i t a b l e when the cuff was i n f l a t e d and experienced occasional muscle spasms. These produce an impulse-like noise of varying amplitude as shown i n Fi g . F3. F3. VASCULAROSCILLATIONS Occasionally, the p u l s a t i l e impedance waveform from a subject with the cuff i n f l a t e d to a pressure near the occlusion pressure was modulated by a low frequency waveform of varying amplitude and period which was not synchronous with r e s p i r a t i o n (Fig. F4). Similar o s c i l l a t i o n s have been observed i n blood flow i n the forearm after an occlusion with a i i i i i i 444 44 ji 44 4444 il iillllliijii H ill! ill! if i :l:i :: ; iimji I ]% It 1 nil li jliijilii 1 — jl •t iiiMii; i ill! illi II -! tHi till' :;:t : 5 1 1 1 1 IP I ittliiif!!! 11144 414 44-flS ii ill! iii! jl 1i Ml! iJHi il j iji! li 1/ p1 iiiiiiiiii 1 1 i si j 1 . ' " ft F i g . F l A d d i t i v e n o i s e from r e s p i r a t i o n . Bandwidth...0.05 Hz t o 100 Hz normal b r e a t h i n g deep b r e a t h i n g i n if 8 u t f l n i4i- -4 4! -i n lappa; M l l l t l l n p f j iii: Iii! il:! Iii: : ii iiii ill! ! ! ! i jpn; i r ><:\ iiii H P HI iii I [.; "ll iiii i it i P i i ifn i It ill! .: 1 '. i l .LJ: iiu 111-!!:! :U: Iji! V:V: Iii! !!••' i :;;{ :ni j i1 •>•• <••••1 •!!: ;•! j ! iiii 1!!! Iiii M i ! ! ! Iii li n i l i!ii TTCETM TTTf Hi • ; L I I 1 HIT illi! i i i i ! : ! ! ! I t f 1 • ' 1 i i l j i ! ! |l!i ! Jl Iiii if II Jill |1 ml i ]t jiti nil li III! ill! III! ill! Iii! ill F i g . F2 M u l t i p l i c a t i v e modulation from r e s p i r a t i o n . Bandwidth...1.5 Hz t o 30 Hz 202 I M ! 1 I I 1 F i g . F3 Artifact from muscle spasm. 203 tourniquet cuff was released and was attributed to o s c i l l a t o r y expansion of the a r t e r i o l e s . Vasomotion, a similar transient o s c i l l a t o r y blood flow i n the peripheral c i r c u l a t i o n , i s believed to be caused by the synchronous contraction of the smooth muscle i n the a r t e r i a l wall and to be metabolically mediated [56,57]. An alternate explanation i s that the increase i n a r t e r i a l d i s t e n s i b i l i t y with the reduced translumenal pressure causes the hydraulic resistance of the artery to be lowered to the point where the blood flow versus pressure r e l a t i o n s h i p become unstable and the flow o s c i l l a t e s [58]. The amplitude of the modulation was dependent on both the subject and the electrode configurations and i n the extreme, was equal to the amplitude of the impedance pulses when the cuff pressure was near the occlusion pressure. The change i n the frequency content of the r e s u l t i n g modulated waveform i s small since the modulating waveform i s approximately sinusoidal and i t s freqency i s low (about 0.1 Hz) for the same reasons discussed i n F l . In a manner sim i l a r to the modulation due to r e s p i r a t i o n , the o s c i l l a t i o n s corresponded to changes i n cuff pressure necessary to collapse the artery undergoing the o s c i l l a t i o n s . Because the o s c i l l a t i o n s are not expected to occur i n a l l the ar t e r i e s at the same time nor to be synchronous i f they do, the modulation must be treated as noise. 205 F4. PREMATURE VENTRICULAR CONTRACTIONS Premature ventricular beats are not uncommon i n healthy individuals and can r e s u l t from smoking, coffee, lack of sleep, and emotional i r r i t a b i l i t y [59]. The contraction of the heart i s normally a well coordinated series of events. If the timing of the normal chain of events i s disturbed for one or two beats, as with premature ventricular contractions, the cardiac output can be reduced, and the r e s u l t i n g a r t e r i a l blood pressure reduced u n t i l the normal contraction returns. A lower blood pressure would reduce the amplitude of the impedance pulses. F i g . F5 shows an example which may be due to a premature v e n t r i c u l a r contraction. Other arrhythmias can s i m i l a r l y compromise the cardiac output for one or two beats and would be expected to produce sim i l a r a r t i f a c t . F5. ALTERNATE CURRENT PATHWAYS The d i s t r i b u t i o n of current from the plethysmograph current source i s determined by the r e l a t i v e impedance of the alternate paths. While the p r e f e r e n t i a l path i s through the limb tissue d i r e c t l y between the source electrodes, other paths can ex i s t . With the plethysmograph on the upper arm, for instance, touching hands w i l l provide an alternate path for current from the d i s t a l electrode, through the hands, along 206 Fi g . F5 Artifact from PVC's. the other arm, through the thorax, to the proximal electrode. S i m i l a r l y , i f the hand touches ground, an alternate path w i l l e x i s t since the subject w i l l have a f i n i t e impedance from his body to ground (Fig. F6). If the impedance of the alternate path varies with time, the voltage from the plethysmograph representing the p u l s a t i l e impedance w i l l show this v a r i a t i o n . F6. VIBRATION Vibration coupled to subject can also cause noise. The eff e c t i s sim i l a r to motion a r t i f a c t i n that i t can cause motion which involves the musculature beneath the cuff that i s being sampled by the impedance plethysmograph. F7 . CUFF MOTION FROM. ARTERIAL EXPANSION A dominant source of noise for tourniquet cuffs applied to the leg was cuff motion due to the i n f l u x of a r t e r i a l blood at the proximal end of the cuff. The a r t e r i a l expansion pushed the cuff d i s t a l l y i n synchronism with the heart rate. The motion changed the position of the plethysmograph electrodes r e l a t i v e to the underlying tissue, causing an impedance change. The a r t i f a c t can not be distinguished from the impedance pulses; i t occurs at the same time and has the same waveform. 208 NOISE AND ARTIFACT - .I i II i i i i i 11 i'" i i I I i i " i i 11 | 11 i .II I-. T 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 TIME (seconds) F i g . F6 Art i fact from an alternate current path. 1 . Impedance pulses 2. Background noise with cuff pressure sufficient to suppress the impedance pulses 3. Art i fact from touching then removing the hand from ground F8. BAND ELECTRODES A r t i f a c t was experienced when band electrodes secured to a cloth template were used beneath a tourniquet cuff on the leg. It appeared much l i k e an impulse, possibly due to the high pass f i l t e r i n g i n the analog c i r c u i t s . The a r t i f a c t occurred when the overlying cuff was pressurized and possibly was due to the slow and i r r a t i c movement of the template and electrodes along the skin i n response to the pressure. F9. ELECTRONICIN S TRUMENTATION The dominant noise from the electronic instrumentation originated from the m u l t i p l i e r s used i n the plethysmograph. Replacing the m u l t i p l i e r s reduced the peak noise to -32 dB of the amplitude of the impedance pulses that are seen without a hyperemic response. 210 APPENDIX G: EQUIPMENT USED IN THE STUDY The instrumentation used i n the study, shown i n F i g . 5.1, i s l i s t e d below along with i t s major features. i) Hewlett-Packard Inc. model 7754A chart recorder and mainframe: a) four channel chart recorder; b) accepts four plug-in amplifiers; c) l i n e - i s o l a t i o n transformer. i i ) B i o e l e c t r i c amplifier plug-ins for the chart recorder, model 8811A: a) gain variable from 0.1 mv/division to 200 mv/div; b) lower frequency half-power point variable from 0.05 Hz to 150 Hz i n steps; c) upper frequency half-power point variable from 0.1 Hz to 10 kHz i n steps; d) voltage output of amplified and f i l t e r e d signal available. i i i ) C a rrier amplifier plug-ins for the chart recorder, model 8805B: a) accept model 1280 pressure transducers; b) continuously variable gain from 10 mmHg/div to 500 mmHg/div; c) three available frequency response settings; d) voltage output of amplified and f i l t e r e d signal available. iv) Data Translation Inc. DT2801 single board analog and d i g i t a l I/O system: a) plugs into an expansion s l o t i n an IBM AT-compatible personal computer; b) 12 b i t analog to d i g i t a l converter; c) 16 channel multiplexer. v) Tecmar Inc. PC-Mate dual stepper motor c o n t r o l l e r board: a) plugs into an expansion s l o t i n an IBM AT-compatible personal computer; b) two stepping motor c o n t r o l l e r s . vi) Hoffmann-La Roche Inc. Arteriosonde 1010 Blood Pressure Instrument: a) detects a r t e r i a l wall motion with an ultrasonic transducer. v i i ) W.A.Baum Co. Inc. mercury manometer. v i i i ) Tektronix Inc. model 7603 oscilloscope: a) two 7A26 dual trace v e r t i c a l amplifiers; b) 7A15A v e r t i c a l amplifier; c) 7B53A dual time base. ix) Standard Power Supplies Inc. SPS-60-5 f i v e v o l t power supply. x) Wavetek Inc. model 182A function generator. xi) Aspen Labs Inc. model A.T.S. 1000 tourniquet system: a) ± 5 mmHg accuracy; b) ± 6 mmHg hysteresis about set point; c) pressure adjustable from 50 mmHg to 475 mmHg i n steps of 1 mmHg. x i i i ) Applied Motions Products Inc. A061-ED0051 stepping motor. xiv) Tissue pressure transducer and related instrumentation. xv) Copam Inc. model PC 501-TURBO IBM AT compatible personal computer: i) 10 MHz clock; i i ) 1 Mbyte of memory; i i i ) 20 Mbyte hard disk; iv) 1.2 Mbyte floppy diskette drive. 213 APPENDIX H: FIRST AND SECOND DUAL-BLADDER CUFF EXPERIMENTS HI. EXPERIMENTPROCEDURE The procedure for the f i r s t two experiments was similar except for the means of attaching the electrodes to the arm. The subject was seated with her forearm resting comfortably on a desk. At the midpoint of the upper arm, the brachial artery was located by palpation and the position marked. In the f i r s t experiment, the electrodes, with electrode leads attached, were positioned r e l a t i v e to the marked postion and adhered to the arm. The electrode leads were secured to the arm with micropore tape. After placing the tissue pressure transducer along the arm above the electrodes, the cuff was c a r e f u l l y aligned such that the middle seam was over the marked artery position and then wrapped around the arm and tightened such that one finger could be s l i d underneath the proximal end but three could not. In the second experiment, the electrodes were attached to the cuff before the cuff was applied thereby assuring proper alignment between the cuff and electrodes. A thin clear p l a s t i c sheet was placed over paper on which the proper electrode configuration was marked. Electrodes were attached to the sheet above the marked electrode positions and electrode leads were attached to the electrodes. After securing the electrodes and leads with micropore tape, the array was attached to the cuff with tape having adhesive on 2 1 4 both sides. The p l a s t i c sheet was then c a r e f u l l y removed and the cuff placed over the arm such that the array was properly positioned r e l a t i v e to the marked artery. After pressing against the cuff to adhere the electrodes to the arm, the cuff was wrapped around the arm and tightened as above. Coupling gel was applied to the face of the ultrasound sensor and the sensor was placed over the brachial artery d i s t a l to the cuff, secured with micropore tape, and connected to the Arteriosonde. The electrode leads were then attached to the leads from the remote box housing the current source and instrumentation amplifiers and the equipment turned on. The proper operation of the plethysmograph was v e r i f i e d by looking at the input to the demodulators with an oscilloscope. The cuff hoses were then connected between the cuffs and the Tourniquet Systems, and the d i s t a l bladder i n f l a t e d to 2 0 0 mmHg. The bladder pressure was then reduced u n t i l blood flow resumed as noted by the Arteriosonde. After noting t h i s pressure, P n , and adjusting the B i o e l e c t r i c Amplifier gains for near maximum de f l e c t i o n , the cuff pressure was released. The proximal and d i s t a l bladder pressures were then set to P n + 3 0 mmHg and P n + 9 0 mmHg respectively and a 1 0 second epoch of the impedance pulses recorded on the chart recorder and the computer. Both bladder pressures were reduced by 1 0 mmHg and another 1 0 second epoch of data recorded. This procedure was repeated u n t i l the d i s t a l bladder pressure was 215 10 to 20 mmHg below the occlusion pressure. At this point the bladder pressures were released. In the second experiment, to c o l l e c t data to est a b l i s h a reference, the process of recording the data was repeated except now with the d i s t a l bladder pressure remaining at Pn+90 mmHg while the proximal bladder pressure was reduced i n steps of 10 mmHg. After disconnecting the electrode leads, the pressure p r o f i l e s were recorded for proximal and d i s t a l bladder pressures of Pn and Pn+60 mmHg respectively, and P n-60 and P n respectively, with the tissue pressure transducer as close as possible to the electrodes without touching them. After removing the cuff and electrodes, the arm was checked for any injury such as bruising or excoriation and the posi t i o n of the electrodes and cuff checked for proper alignment from the imprint l e f t by the cuff. H2. DATAANALYSIS The analysis of the data from the f i r s t and second experiments proceeded i n the same way as the analysis of the data from the t h i r d single-bladder cuff experiment. The data were f i l t e r e d using an in t e g e r - c o e f f i c i e n t low-pass d i g i t a l f i l t e r with i t s f i r s t zero at 30 Hz. After inverting the samples and removing the of f s e t , the pulse amplitudes from 216 both the leading and f a l l i n g edges were found for each 10 s epoch. An average amplitude and variance for each epoch was calculated and written to a separate f i l e along with the maximum and minimum pulse amplitudes. H3. RESULTS As with the data from the t h i r d experiment using the single-bladder cu f f s , the graphs of peak impedance pulse amplitude versus r e l a t i v e cuff pressure for the f i r s t and second experiments were evaluated for o f f s e t , rate of onset, and point of onset. Again the inter-subject v a r i a b i l i t y i n the point of onset was large and for a simi l a r reason. Although the pressures i n the bladders were adjusted for a 60 mmHg difference, the difference between the peaks i n the pressure p r o f i l e beneath the cuff varied from 20 to 90 mmHg. The eff e c t of varying the difference i n the peaks would be to change the point of onset, which i s what was observed. Other s i g n i f i c a n t sources of error included the accuracy of the tourniquet systems i n regulating the cuff pressures, sampling the blood flow and impedance signals only at discrete pressures 10 mmHg apart, and the time taken to complete the experiment for each subject. 217 With the error i n each tourniquet system as large as 4 mmHg, the error i n the difference between bladder pressures could be as high as twice this value or 8 mmHg. Because the blood flow signal was used to detect the occlusion pressure and was only sampled every 10 mmHg, the occlusion pressure to which the cuff pressures are referenced could be i n error by almost 10 mmHg. I f , for example, the occlusion pressure was 149 mmHg, and the blood flow signal was only sampled at 140 and 150 mmHg, 140 mmHg would be used as the occlusion pressure since no blood flow would be detected at 150 mmHg but i t would be at 140 mmHg. Also, for the second experiment i n which the impedance pulse amplitude maximum was used as a reference, the actual maximum could be l i e between two sampled points, and the reference would be taken as the maximum pulse amplitude of the two sampled points. As much as 10 min could pass from the beginning of the measurement of the f i r s t epoch to the measurement of the l a s t , and i t i s possible that some parameter af f e c t i n g the rela t i o n s h i p between tissue impedance and cuff pressure could have changed (blood pressure, for example). Because of the v a r i a b i l i t y i n the point of onset, i t was d i f f i c u l t to evaluate the e f f i c a c y of the two methods of deriving a reference. With the method of using separate reference electrodes, however, the reference signal did not remain constant with cuff pressure as shown i n Fig . HI and i t 218 i s u n l i k e l y to be a consistent or accurate means of normalizing the data. The method used i n the second experiment of looking for a maximum as only the pressure i n the proximal bladder was reduced was also used i n the t h i r d experiment where i t was found to be e f f e c t i v e . 219 CXjrr PRES. RKIATIVK TO OCCLUSION (mmHg) • MD + GMC O LD A FT X AS V RS F i g . HI Output from the reference electrodes f o r s i x of the 18 subjects showing the v a r i a t i o n with c u f f pressure. 220 APPENDIX I: SINGLE-BLADDER CUFF EXPERIMENTS II. INTRODUCTION I n i t i a l experimentation was done with single-bladder c u f f s . With these early experiments, there were two main objectives: 1) to i d e n t i f y the variables that govern the r e l a t i o n s h i p between the tissue impedance, cuff pressure, and the onset of flow; and 2) to f i n d the electrode configuration that i s best suited for predicting the onset of blood flow. With the development of the tissue pressure transducer, a t h i r d objective became to investigate the e f f e c t of the pressure p r o f i l e on the r e l a t i o n s h i p between the tissue impedance and cuff pressure. To explore the dependence of the impedance on parameters related to the occlusive cuff, two types of cuffs were used: a tourniquet cuff and a blood pressure cuff. Three experiments were performed i n which the cuff pressure was decreased and the r e s u l t i n g change i n the amplitude of the impedance pulses noted. In the f i r s t , using 5 subjects, 8 configurations were examined: 4 for each of the two cuff types (Fig. I I ) . From these r e s u l t s , four of the o r i g i n a l 8 configurations were chosen for the second study 221 using 9 subjects. One further single cuff experiment was performed with 10 subjects. The inconsistency i n the r e l a t i o n s h i p between impedance pulse amplitude and cuff pressure had been i d e n t i f i e d and i t was believed to be due to the v a r i a b i l i t y i n the proximal edge of the tissue pressure p r o f i l e between subjects and between successive applications of the cuff on the same subject. In the l a s t experiment, the tissue pressure p r o f i l e s were measured as well as the changes i n the tissue impedance with cuff pressure. Only the i n - l i n e longitudinal electrode configuration was used. For the f i r s t two experiments the impedance pulse amplitude was measured from the chart recordings taken during the experiments. For the l a s t experiment the data was sampled and recorded d i g i t a l l y and software algorithms were devised and used to extract the pulse amplitudes. 12. CUFFTYPES The 34 inch (86.4 cm) single-bladder tourniquet cuff used i n the experiments was from Aspen Labs, catalog number 60-4004-008. This 11 cm wide cuff was curved to f i t the taper of a limb and had an 8.5 cm wide bladder. The bladder was contained within a f l e x i b l e cuff with a r i g i d backing and i t s edges were secured to the cuff. In order to provide a separate connection for pressure sensing by an automatic tourniquet, the bladder had 2 ports with Luer type locking pneumatic 222 So Sr Se Se So - \ \ — B (e A 8 Se Se D S o ^ * ^ r-10 cm scale A 1s the Aspen cuff outline B 1s the blood pressure cuff outline So...Source electrode Se...Sense electrode Sr...Reference Electrode The dashed l i n e Indicates the position of the brachial artery Fig. i i The electrode configurations examined 1n the single-bladder experiments. a) Configuration 1; transverse b) Configuration 2; transverse 1n-l1ne c) Configuration 3a; transverse o f f s e t with the Aspen cuff d) Configuration 3b; transverse o f f s e t with the blood pressure cuff e) Configuration 4; longitudinal In-line 223 connectors. Hook and loop fasteners (Velcro) were used to secure the cuff around the limb. The blood pressure cuff, a Baumanometer Calibrated V-lok Cuff available from W.A. Baum Co. Ltd, was a large adult size, 66.4 cm long and 16.6 cm wide. The 32.5 cm by 15.3 cm bladder was free to move within the clot h containment and could i n fact be e a s i l y removed. As with the tourniquet cuff, i t had 2 ports with Luer type locking pneumatic connectors and was secured around the limb with hook and loop fasteners (Velcro). Even though these large cuffs are recommended only for subjects with large arms, they were used for the entire population sample, with arm size varying from 24.1 cm to 35.5 cm, to avoid introducing other cuff dependent variables. For the smaller arms, the cuff had to wrapped more than once around the arm and care had to be taken to assure the edges of each lap li n e d up with the one beneath i t . 13. ELECTRODE CONFIGURATION S Conductive adhesive electrodes (3M-Littman 2322 and Lec Tec Tracets) were used i n the arm study not only because emphasis could be placed on d i s t i n c t tissue volumes with th e i r small size, s p e c i f i c a l l y , the brachial artery, but also because they could be configured to provide a reference by taking advantage of a negative impedance sampling f i e l d . The electrode configurations were selected to compare longitudinal 224 placement with transverse placement, to compare sampling large tissue volumes with sampling small tissue volumes, and to compare d i f f e r e n t methods of finding a reference signal with which to normalize the impedance data. The s p e c i f i c configurations were determined from preliminary empirical investigation and modelling. Because the computer model uses a f r o n t a l or s a g i t t a l section of the arm, the transverse configurations could not be modelled; however, simply estimating the lead vectors provided insight into the performance of the transverse electrode configurations. Configurations 3 and 4 offered the p o s s i b i l i t y of providing a reference s i g n a l . Because of a negative s e n s i t i v i t y i n the impedance sampling f i e l d , the impedance signal from configuration 3 included a posit i v e peak at the beginning of each pulse complex. The magnitudes of these peaks were examined to see i f they could be used to normalize the negative peaks that followed. This means of establishing a reference has the advantage of using the same electrodes for both the reference signal and the desired p u l s a t i l e impedance signal and thus eliminates any v a r i a b i l i t y due to the differences i n electrode c h a r a c t e r i s t i c s or positioning. In configuration 4, an additional electrode was placed proximally to the sense electrodes with the same spacing. Because a r t e r i a l blood must pass beneath this electrode before reaching the sense electrodes, the impedance signal between 225 this electrode and the more proximal sense electrode would give a representation of blood flow past them that could be compared to the impedance signal from the sense electrodes. The use of t h i s signal as a reference was investigated. Configurations 1 and 3 sample large volumes while configurations 2 and 4 sample the volume around the brachial artery. By sampling around a major artery, the proportional change i n impedance as the artery f i l l s with blood i s larger than i n the configurations which sample a larger volume. Configurations 1 and 2 sample transversely while configurations 3 and 4 sample l o n g i t u d i n a l l y . The sense electrodes i n configuration 3 are placed transversely with a longitudinal o f f s e t and the configuration actually samples at an angle to the long axis of the arm. Multi-electrode arrangements for the sense electrodes were not used because each electrode should only be connected to a high impedance to avoid introducing electrode a r t i f a c t . Buffering each electrode would increase the instrumentation complexity. 14. DATACpLLECTlONANDANALYSIS The subject was seated with her l e f t arm resting on a desk. When electrode configurations 2 and 4 were used, the electrodes were placed over the palpated brachial artery of the l e f t arm. The mid-point of the cuff was then aligned 226 r e l a t i v e to the electrodes as i t was put i n place. The accuracy of the placement was checked at the end of the experiment by examining the impressions l e f t on the subject's skin by the seams of the i n f l a t e d cuff. The data from any t r i a l with cuff placement i n error by more than 1 cm were rejected. Templates to a s s i s t i n placing the electrodes used with configurations 1 and 3 were made from a tensor bandage. A section of the tensor was sewn into an armlet to be placed over the upper arm and aligned r e l a t i v e to the a n t i c u b i t a l fossa, the depression on the inside of the elbow. At the desired location of the electrodes, holes were cut i n the armlet and the edges of the holes stitched. Because tensors tend to stretch lengthways without changing th e i r width appreciably, the longitudinal placement of the electrodes on the arm remained approximately the same for d i f f e r e n t arm circumferences. The electrode leads were attached to the electrodes by paper c l i p s with wires soldered to them and secured with micropore tape. Eventually (in the dual-bladder cuff experiments), the leads were either soldered d i r e c t l y to the electrodes or attached to them with short sections of paper c l i p s bent i n a manner to take up much less room. The general procedure for a l l the experiments was the same. After attaching the electrodes, the electrode leads were 227 connected to the electrodes such that they would come out from under the d i s t a l edge of the cuff. The cuff was then applied by c a r e f u l l y placing the midpoint of the cuff (marked on the cuff) r e l a t i v e to the electrodes. As the cuff was wrapped around the arm, care was taken to a l i g n the edges of each wrap. The cuffs were wrapped with enough tension such that one finger could be slipped beneath cuff but three could not. With the t h i r d experiment, the tissue pressure transducer was placed just above the i n - l i n e l o n g i t u d i n a l l y placed electrodes such that the e l e c t r i c a l and pneumatic connectors of the transducer were outside the d i s t a l edge of the cuff. The ultrasonic transducer for the Hoffman-La Roche Arteriosonde was placed over the brachial artery just d i s t a l to the cuff and secured with micropore tape. The electrode leads were connected to the appropriate leads from the remote box which contained the current source and instrumentation amplifiers. Hoses were used to connect the cuff connectors to the Aspen Labs Automatic Tourniquet System. After the equipment was turned on, the input to the demodulator was displayed on an oscilloscope to v e r i f y proper operation, and the gains of the b i o e l e c t r i c amplifiers were adjusted for near maximal d e f l e c t i o n without l i m i t i n g on the chart recorder paper with the cuff pressure set to approximately occlusion pressure. 228 After the cuff was deflated for a moment, i t was i n f l a t e d to about 200 mmHg with the Aspen Labs Automatic Tourniquet System. The cuff pressure was then reduced i n steps of either • 10 mmHg or 5 mmHg and 10 second epochs recorded on the chart recorder (and also with the computer i n the l a s t experiment). The pressure was decreased to about 15 mmHg below occlusion. The pressure p r o f i l e was measured i n the l a s t experiment by moving the t i p of the transducer as far towards the proximal edge of the cuff as possible. After determining the tissue pressure at the f i v e d i s t i n c t points on the transducer, the cuff was deflated and the transducer moved so that after the cuff was r e i n f l a t e d , the tissue pressure at f i v e additional points could be determined. After the cuff was removed, the arm was checked for any injury such as excoriation or bruising, and the impression of the cuff l e f t on the arm was examined for proper placement of the cuff r e l a t i v e to the electrodes. For the f i r s t two experiments, the heights of the impedance pulses were taken from the chart recordings. Using dividers, a l l the pulses within each epoch that were not contaminated with noise were measured and an average taken. Software was written for the l a s t experiment to automatically determine the average pulse height i n each epoch, the variance of the pulse heights, maximum pulse height, and the minimum pulse height. 229 The programs for data a q u i s i t i o n were written i n C-language and made use of exi s t i n g subroutines for multiplexing, A/D conversion, software reset, and error handling. To extract the pulse heights i n software, the data were f i r s t f i l t e r e d using an in t e g e r - c o e f f i c i e n t low-pass d i g i t a l f i l t e r with i t s f i r s t zero at 30 Hz [60] . Because the signals were converted from analog form into o f f s e t binary, the offset was removed by subtracting 2048. They were then inverted so that the impedance pulses would be p o s i t i v e . After finding the maximum value of any sample i n the epoch, the peaks were found by scanning the record using a threshold of the maximum value divided by three. When a sample value larger than the threshold was found, the maximum sample value encountered from that point on was recorded as the scanning continued u n t i l the samples were again lower than the threshold. This maximum value corresponded to a peak of an impedance pulse and was recorded along with i t s address. The heights of both the leading edge and the f a l l i n g edge were determined. The address of the base of the leading edge was determined by finding the second i n f l e c t i o n point p r i o r to the peak and the address of the base of the f a l l i n g edge was determined by finding the point after the peak where the slope becomes p o s i t i v e . This i s i l l u s t r a t e d i n Fig . 12. The heights are determined by subtracting the values at the bases from the 230 g . 12 Algorithm f o r determining pulse amplitude from the r i s i n g and f a l l i n g edges. 1 ) Impedance pulse 2) F i r s t d e r i v a t i v e 3) Second d e r i v a t i v e The base of the r i s i n g edge i s taken at the second zero of the second d e r i v a t i v e p r i o r to the peak of the impedance pulse (point x ) . The base of the f a l l i n g edge i s taken where the f i r s t d e r i v a t i v e becomes p o s i t i v e a f t e r the impedance pulse peak (point y ) . The amplitudes are c a l c u l a t e d by s u b t r a c t i n g base values from the peak value! 231 peak value. After finding the heights of a l l the impedance pulses i n an epoch, the program recorded i n a f i l e the average height, the maximum height, the minimum height, and the unbiased estimate of the variance [61]. The addresses of the peaks i n each epoch were checked for accuracy by viewing the epoch using the graphic c a p a b i l i t y of the program, Lotus 123 (Release 2, Lotus Development Corporation, Cambridge, Ma). If any were inaccurate, or i f any of the data was corrupted with a r t i f a c t , the erroneous peak heights were removed and the calculations redone manually. The occlusion pressure was taken to be the cuff pressure at which blood flow was f i r s t detected with the Hoffman-La Roche Arteriosonde. The e f f i c a c y of the reference was evaluated by determining the independence of the reference signal amplitude from the cuff pressure, and by examining the inter-subject v a r i a b i l i t y after normalization with the reference. 15. RESULTS 15.1 Firstexper^™?^^ For each t r i a l , the pulse heights were plotted against the cuff pressure and the graphs evaluated for o f f s e t , rate of onset, point of onset, and noise where: 232 1) the o f f s e t i s the magnitude of the impedance pulses at high cuff pressures r e l a t i v e to the magnitude at occlusion; 2) the rate of onset i s the rate of change of the height of the impedance pulses near the resumption of blood flow; 3) the point of onset i s the cuff pressure near the resumption of flow at which the amplitude of the impedance pulses begins to increase rapidly; and 4) the noise i s the random background interference subjectively evaluated at the occlusion pressure. In general, for the data recorded when the tourniquet cuff was used, the o f f s e t was larger than that of the blood pressure cuff, the rate of onset was poorer, but the point of onset was farther from the occlusion pressure. Data from one of the subjects i s shown i n F i g . 13. From the f i r s t experiment, the four electrode configurations chosen were: 1) tourniquet cuff, configuration 2, transverse i n - l i n e placement with l o c a l i z e d sampling; 2) tourniquet cuff, configuration 3, transverse-offset placement with large area sampling; 3) blood pressure cuff, configuration 3, transverse-o f f s e t placement with large area sampling; and 4) blood pressure cuff, configuration 4, longitudinal i n -l i n e placement with l o c a l tissue sampling. 4 conf igurat ions with Subject A Blood Prcnura Cuff 90 110 130 ISO 170 cuff pnum (mmHg) 4 conf igurat ions with Subjec t A A*p*n Cuff cuff prauura (iranHg) F i g . 13 Data.from one of the subjects showing the results from a l l eight electrode configurations. The l e t t e r and number Identifying the configuration are placed on the data point at the occlusion pressure. 234 15.2 Second ex In the second experiment, the data was normalized with the impedance pulse amplitude at the occlusion pressure for comparison. Evaluation of the rate of onset with th i s normalization was d i f f i c u l t because of the considerable v a r i a t i o n i n the point of onset between subjects. No subjective evaluation of the noise was done as i t was i n the f i r s t experiment. The data for each configuration showed considerable inter-subject v a r i a t i o n i n both the o f f s e t and the point of onset. The performance with each of the configurations i s shown i n F i g . 14. The dominant cause of the inconsistency i n the o f f s e t and point of onset was the v a r i a b i l i t y i n the proximal edge of the pressure p r o f i l e beneath the cuff along the arm. This i s discussed i n more d e t a i l i n the following section on the t h i r d experiment. Other possible causes, which may be present but have less e f f e c t , are cuff motion a r t i f a c t and an e f f e c t i v e impedance s e n s i t i v i t y f i e l d much larger than expected. Cuff motion a r t i f a c t was a dominant source of error i n the studies on the leg and originated from the cuff moving d i s t a l l y each time the art e r i e s at the proximal edge of the cuff expanded during systole. The motion causes a pulse s i m i l a r i n shape-to the impedance pulse and appearing at the same time. Its e f f e c t would be to increase the o f f s e t and 235 Blood Pressure Cuff, Transverse Electrodes Blood Pressure Cuff, In-Llne Transverse Tourniquet Cuff, Longitudinal Electrodes Tourniquet Cuff, Transverse Electrodes VERTICAL SCALE: Amplitude of P u l s a t i l e Impedance Normalized with Value at Occlusion HORIZONTAL SCALE: Cuff Pressure Relative to Occlusion Pressure 1n mmHg VARIATION OF PULSATILE IMPEDANCE WITH CUFF PRESSURE: Average of Nine Subjects with Maximum and Minimum Responses indicated F i g . 14 The r e s u l t s of the second single-bladder cuff experiment. The data 1s f o r nine subjects and shows excessive Inter-subject v a r i a t i o n f o r a l l the configurations. 236 decrease the rate of onset with the normalization that was used. The point of onset would remain the same but would be more d i f f i c u l t to determine because of the larger o f f s e t . A larger e f f e c t i v e impedance sampling f i e l d would extend the s e n s i t i v i t y to conductance changes into tissue that i s farther from the sense electrodes than expected. The eff e c t would be a larger o f f s e t , a premature point of onset, and a lower rate of onset (with normalization with the pulse amplitude at occlusion). Evaluation of the reference from configuration 4 was l e f t to the t h i r d experiment. With the transverse-offset electrode configurations, 3a and 3b, the height of the i n i t i a l p o s i t i v e pulse, which sometimes was not even present, was very i r r e g u l a r between subjects and was not a r e l i a b l e reference. 15.3 Thirdexperiment The e f f e c t of an inconsistent proximal edge of the pressure p r o f i l e was investigated i n the t h i r d experiment. F i g . 15 shows the pressure p r o f i l e s for 10 subjects. The v a r i a b i l i t y i n spite of consistent application and cuff pressure i s apparent. F i g . 16 shows the r e s u l t i n g pulse height versus r e l a t i v e cuff pressure curves with the numbers in d i c a t i n g the r e l a t i v e amplitude of the pressure p r o f i l e at a distance of 2.2 cm from the proximal edge of the cuff (the number 1 i s the largest amplitude, while the number 10, shown 2 2 0 PRESSURE P R O F I L E S Cuff Pressures 1 9 0 to 2 0 0 mmHg KD Distance Along Arm (cm) GM O LO A FT AS RS Distance Along Arm (cm) BD O KH A LS MB f i g . 15 The tissue pressure p r o f i l e s beneath the single-bladder cuff f o r 10 subjects. There 1s a large variation 1n the leading edge of the p r o f i l e s . 238 IMPEDANCE PULSE AMPLITUDE—RIGHT PEAK Single Bloddor Tourniquet Cuff Relative Cuff Preasure (mmHg) F i g . 16 The variation 1n Impedance pulse amplitude with cuff pressure f o r the t h i r d single-bladder c u f f experiment. The curves are numbered 1n the order of the r e l a t i v e magnitude of the leading edge of t h e i r respective pressure p r o f i l e s . The number "0" 1s number 10 1n the order. as 0, represents the smallest). The trend i n the delay i n the point of onset as the cuff pressure i s reduced towards the occlusion pressure i s apparent except for curve 1. Since only four data points for the pressure p r o f i l e leading edge were available, the anomaly of curve 1 could be due to an a r t i f i c i a l l y high tissue pressure reading because of, for example, a crease i n the cuff material overlying the sensor contact. 15.4 Reference To normalize the impedance pulse amplitude from the sense electrode pair of the i n - l i n e electrode configuration, the amplitude was divided by the amplitude of the corresponding pulse from the reference electrode pai r . The r e s u l t s were hard to interpret because the normalization would not correct the v a r i a t i o n present i n the point of onset. The reference pulse heights were not completely constant and did show dependence on the cuff pressure (Fig. 17). 240 Single-Bladder Tourniquet Cuff CUFF PIUES. RELATIVE TO OCCLUSION (mmHg) • MD + GHC O L D A FT X AS V R S F i g . 17 The output from the reference electrodes for six of the subjects showing the variation In amplitude with cuff pressure. APPENDIX J: LEG TOURNIQUET EXPERIMENTS 241 J l . INTROpU Some experimentation was done using the Aspen Labs 34 inch single-bladder tourniquet cuff on the thigh. Band electrodes were used instead of the conductive adhesive EKG electrodes. Because the taper of the thigh i s much greater than that of the upper arm, there was a large a r t i f a c t due to cuff motion that was not seen with the arm studies. As well, the band electrodes intermittently caused a large a r t i f a c t . These problems w i l l have to be resolved before the tissue impedance versus cuff pressure r e l a t i o n s h i p with the leg can be properly investigated. J2. THECUFF The 34 inch single-bladder tourniquet cuff was used for the leg studies. The large adult sized blood pressure cuff was t r i e d but would move on the leg when i n f l a t e d . Because the bladder edges were not fixed to the blood pressure cuff, the bladder would begin to r o l l down the leg when i n f l a t e d . The tourniquet cuff was much more p o s i t i o n a l l y stable due to i t s design, e s p e c i a l l y from the curve along i t s length that helped i t to f i t a tapered limb, although even th i s cuff would occasionally come loose i f placed on a limb with a pronounced 242 taper. In order to decrease the l i k e l i h o o d of the cuff s l i p p i n g and to reduce the cuff motion a r t i f a c t discussed below, the cuff was applied as proximally on the limb as possible where the taper of the thigh tended to be the smallest. J3. ELECTRpDES It was necessary to use band electrodes instead of the conductive adhesive electrodes used i n the arm studies. Conductive adhesive electrodes have the advantage of emphasizing the impedance change i n tissue near the surface of the arm, such as around the brachial artery. With the thigh, however, the a r t e r i e s are much deeper. At the proximal end of the thigh where the cuff was to be placed, they l i e close to the center of the thigh. In F i g . J l , i n which one electrode configuration i s modelled for both the arm and leg, i t can be seen that the s e n s i t i v i t y pf the impedance sampling f i e l d around the artery i n the leg i s much lower than that i n the arm. A permanent band electrode array was made by gluing aluminum tape (presumably aluminum deposited on a mylar backing) to a clot h backing. The tape was sensing tape (7/32 inch Sensing F o i l , part no. 44-1155, Radio Shack Divi s i o n of the Tandy Corpoaration) used as a leader for 1/4 inch audio tape. By making the clot h backing the same shape as the tourniquet cuff, the cuff could e a s i l y and reproducibly be leg with artery shown arm with artery shown Fig. J l Comparison of the same electrode configuration on the arm and leg. The scale on the l e f t side diagrams f o r the leg i s twice that of scale on the r i g h t hand drawings f o r the arm. to u> 244 aligned with the electrode array on every application. Applying the disposible band electrode tape (Contact Products Ltd. M6001 Electrode Tape) accurately was a very d i f f i c u l t procedure. Furthermore, the electrode tape was very uncomfortable when the cuff was i n f l a t e d or the leg flexed and extended. The electrode array was applied without any conductive electrode paste. The band electrodes often caused spurious impedance signals a f t e r the cuff was i n f l a t e d even though none was present before i n f l a t i o n . The a r t i f a c t often could be eliminated by reapplying the electrode array but i t was not always possible to predict i t s occurence p r i o r to i n f l a t i o n of the cuff. J4. CUFFMOTIONARTIFACT When the ar t e r i e s along the proximal edge of the i n f l a t e d cuff became engorged with blood and expanded during systole, the cuff would be pushed d i s t a l l y causing a change i n the impedance signal from the sense electrodes. The signal was i n phase with the desired impedance pulse signal and had a sim i l a r shape, making i t d i f f i c u l t to dis t i n g u i s h the two. The cuff motion was quite v i s i b l e by s i t t i n g i n a chair and viewing the edge of the cuff against some stationary l i n e d paper placed on the f l o o r beneath the cuff. Also, the a r t i f a c t could be noticeably decreased by securing the cuff to the torso with a tensor bandage placed around the lower back. J5. RESULTS With the cuff placed as proximally as possible on the thigh, the impedance pulse amplitude would increase as the cuff pressure was reduced towards the occlusion pressure. Before further investigation of the r e l a t i o n s h i p i s warranted, however, the problem of electrode and cuff motion a r t i f a c t w i l l have to be addressed. 246 APPENDIX K: CONTROL SOFTWARE Kl . CONTROLSOFTWARE The PC-Mate Stepper Motor Controller w i l l accept programs from the host computer to control the stepping motors without any intervention needed from the host computer. This mode was used i n the ramping operations and c a l l s to the subroutines, MOTORX(argument) and MOTORY(argument), that were used to quickly i n f l a t e and deflate the c u f f s . The d i r e c t or nonprogrammed mode of the PC-Mate Stepper Motor Controller was used with the control algorithm to control the bladder pressures. Five subroutines were used by the main program. MOTORX(argument) and MOTORY(argument) were used to quickly change the pressures with the program mode of the PC-Mate Stepper Motor Controller where the arguments are the desired pressures. DELAY(argument) was used to delay for a fixed period where the argument i s the delay i n seconds. SENDX(argument) and SENDY(argument) were used to send ASCII data to the co n t r o l l e r s where the arguments are the ASCII characters. A general flow chart of the software i s shown i n F i g . K l . Each main section i s explained i n more d e t a i l below with reference to the program at the end of this appendix. K2. INITIALIZATION As shown in.- F i g . Kl (a) , the program requests that the signal into each A/D channel be taken to zero to establish zero pressures, impedance, and blood flow i n f l a t e the d i s t a l and proximal bladders to 240 mmHg and 180 mmHg i adjust the bladder pressures i n t e r a c t i v e l y s t a r t to ramp down the proximal-bladder pressure to est a b l i s h a reference F i g . Kl l a ) I n i t i a l i z a t i o n 2 4 8 F i g . Kl (b) Sampling 249 s t a r t looking for the base F i g . Kl (c) Looking for a peak I calculate the pulse height F i g . Kl (d) Looking for the base take an average of the l a s t three peak medians J save the average for the reference i f the average i s the largest yet F i g . Kl (e) Reference f i n d error by subtracting the pulse height from the reference value calculate change i n bladder pressures necessary to reduce the error to zero change the bladder pressures F i g . Kl (f) Control 253 r e i n f l a t e proximal bladder 1 ' wait one second s t a r t control F i g . Kl (g) Reinflate the proximal cuff and s t a r t control stop control wait 40 seconds ramp down the bladder pressures write the sampled impedance and pressure data to f i l e s ^ stop ^ Fi g . Kl (h) suspend control and write the sampled data to f i l e s 255 baseline values for the program (lines 42-63) . The b i o e l e c t r i c amplifier outputs are switched off and the pressure transducers are exposed to atmospheric pressure. Once the baseline values have been measured, the program prompts the user to return the equipment to i t s operational mode. The proximal and d i s t a l bladders are then i n f l a t e d to 180 and 240 mmHg respectively (lines 66-71). In the program, the proximal bladder i s i d e n t i f i e d with an "x" and the d i s t a l with a "y". Each bladder pressure i s then adjusted i n t e r a c t i v e l y i n turn (lines 74-104). The program requests the new bladder pressure be entered on the console and adjusts the bladder pressure accordingly. When the pressure i s adjusted c o r r e c t l y to produce the desired tissue pressure p r o f i l e peak, the adjustment can be terminated by entering two i d e n t i c a l pressures consecutively. After both bladder pressures have been adjusted, the program asks whether the pressures are to be adjusted again. If yes, the process i s repeated. If no, the program begins ramping down the proximal cuff pressure. The pressure i s ramped down by sending a sequence of ASCII code to the x-motor c o n t r o l l e r (lines 107-128). The code causes the motor to move a small number of steps (XSTEPS) and then delay for a shprt time (XDELAY). This sequence i s repeated a number of times (XITERS), causing the pressure to slowly decrease. Once the code i s given to the c o n t r o l l e r , no 256 further intervention i s required by the computer other than to check the c o n t r o l l e r status r e g i s t e r to see i f the ramping i s f i n i s h e d (line 153). K3. SAMPLING Fig . Kl (b) Prior to taking another sample, the number of samples taken since the l a s t peak was found and since the threshold was l a s t adjusted i s checked (lines 148-150). If the number exceeds 60 i n both cases (2/3 s), the threshold i s decreased by 1/2. The sample ( f i l t e r _ i n [ j ] ) i s then input to an integer c o e f f i c i e n t low-pass d i g i t a l f i l t e r with i t s f i r s t zero at Hz (lines 158-173). The difference equation for the f i l t e r y(n) = 2*y(n-l) -y(n-2) +x(n) -2*x(n-12) +x(n-24) where y(*) i s the f i l t e r output and x(*) the input. The z-domain transfer function i s : H(z) = ( l - z - 1 2 ) 2/(1-z- 1 ) 2 K4. LOOKINGFORAPEAK F i g . Kl (c) In a manner s i m i l a r to the program described i n 6.4, the program detects peaks by continuously looking for a sample greater than a threshold (line 181 and l i n e 191). Once one i s \ 257 found, the maximum sample encountered i s recorded (lines 193-196) u n t i l the samples again f a l l below the threshold. The maximum i s a peak. K5. LOOKINGABASE Fi g . Kl (d) Once a peak i s found, the program looks for the base of the peak, which i s the f i r s t sample following the peak at which the f i r s t derivative becomes po s i t i v e (lines 215-216). The difference equation for the f i r s t derivative i s [62]: f i r s t _ d e r i v a t i v e = x(n) -2*x(n-l) +2*x(n-2) +x(n-3) The pulse height i s determined by subtracting the base value from the peak value (line 222). The median of the l a s t three peak amplitudes i s then determined (lines 223-232) and the threshold that i s used to f i n d the next peak becomes 1/4 of the median (line 233). K6. REFERENCE F i g . Kl (e) If the program i s s t i l l ramping down the proximal-bladder pressure to f i n d a reference, then an average of the l a s t three median values of the peak amplitude i s taken (lines 243-248) . The maximum average encountered (max_average) while looking for a reference i s taken to be the reference (line 249) . 258 K7. CONTROL F i g . Kl (f) , If the program i s i n the control mode i n which the bladder pressures are being regulated by the computer, i t checks the y-motor c o n t r o l l e r status r e g i s t e r to see i f the motors have fi n i s h e d moving after the l a s t pressure correction (li n e 261). If not, an error value i s calculated (line 265): error = height[2] - ref_value where height[2] i s the most recently determined pulse height and ref_value i s 1/2 the reference. From the error, the necessary number of motor steps are calculated (lines 279 and 289) to change the cuff pressure and decrease the error. The program then sends the required p o s i t i o n change to the motor co n t r o l l e r s (lines 266-295). K8. REINFLATETHEPROXIMALCUFFANDSTARTCONTROL Fi g . Kl(g) When the proximal-bladder pressure has fini s h e d ramping down, the system i s i n i t i a l i z e d for control (lines 304-330). The proximal-bladder pressure i s returned to i t s s t a r t i n g value (line 309) and a one second delay i s i n i t i a t e d to allow the system to s e t t l e (line 315). The motor speeds are increased (lines 327-328). 259 K9. SUSPEND CONTROL AND WRITE THE SAMPLED DATA TO FILES Fi g . Kl (h) A delay of 40 s i s i n i t i a t e d (lines 339-340). Afterwards, both the bladder pressures are ramped down (lines 342-378) and the sampled data are written to f i l e s (lines 380-395). 03-15-11 22:00:26 \C86\COmOL\COmOL6.C Pg 1 Toe 03-15-88 22:17:30 n a i n o f 10 1-41 1 f i n c l u d e ' s t d i o . h ' 2 l i n c l n d e "dthead.h' 3 i n t xpress*180,ypress=240,xdelay=89,ydelay=449; 4 i n t x s t e p s * 2 , y s t e p s = 2 , x i t e r s = - 2 0 8 , y i t e r s = - 1 6 7 ; 5 i i t x x e r o * 0 , j x e r o * 0 , p l t i e r o = 0 , a r t r e r o = 0 , d c o n s t = 4 ; 6 7 8 i t i a l ) 9 1 10 FILE ' o u t p u t . f i l e l ; 11 FILE «output_file2; 12 FILE »output_file3; 13 FILE ' i n p u t f i l e ; 14 15 i n t j , f i l t e r J n [ 2 5 ] , f i l t e r j > u t [ 4 1 , h e i g h t [ 3 ) ; 16 i n t t h r e s h o l d ; 17 i n t f i r s t j e r i v a t i v e ; 18 i n t p e a k _ y a l u e , a b o v e _ t h r e s h o l d ; 19 i n t n o s t . n e d i a n , l e a s t ; 20 i n t l o o k p . l o o k b ; 21 i n t p e a M 3 ] , a v e r a g e , n a x _ a v e r a g e , r e f e r e n c e ; 22 i n t peak a d d r e s s , r e f v a l u e , c o n t r o l , c o u n t , x [ 2 0 0 ] , y [ 2001 ; 23 i n t xstepn»[200l,ystepss[200); 24 i n t t e n p , b t t f f e r x [ 2 0 0 ] , b u f f e r y [ 2 0 0 ] , e r r o r a [ 2 0 0 ] ; 25 i n t i , g > h , c o n [ 2 0 0 ] , p r e s s x [ 2 0 0 ] , p r e s s y [ 2 0 0 ] ; 26 i n t n , p l e t h [ 8 0 0 0 ] , h e i g h t o u t [ 2 0 0 ] , h e i g h t a d d r e s s ! 2001 ; 27 i n t s e t a d c ( ) , s e t c l o c k ( ) , d t r e s e t ( ) , c h e c k p ( ) ; 28 i n t a d i ( ) , s e t p o r t ( ) , d w r p o r t ( ) , s a n p l e s , o x p r e s s , o y p r e s s ; 29 i n t o n e _ s e c o n d , s i z e _ o f _ e r r o r , k ; 30 char buffer[8],ans»er[l),to a i r [ l ] , a d j u s t j r e s [ l ] ; 31 32 l o n g i n t nunber x s t e p s , n u n b e r j s t e p s , e r r o r ; 33 34 i=0; 35 f h i l e ( i < 2 0 0 ) ( c o n ( i ) = 0 ; x ( i ] = 0 ; y(i)»0; x i t e p i s [ i ] = 0; y s t e p s s ( i ) = 0 ; b u f f e r x [ i ] = 0 ; b u f f e r y [ i ) = 0 ; ' e r r o r a [ i ) = 0 ; i++; p r e s s x ( i ] = 0 ; p r e s s y ( i l = 0 ; I 36 i=0; 37 38 d t r e s t t l ) ; /» i n i t i a l i i e A/D board 39 40 H / I M t i t t l l t t t l H t t i t t l t t t t i t l t t l t t t l l t t t i t t i l t t t l i t t t t i t 03-15-18 22:00:26 \C86\COmOL\COmOL6.C Pg 2 Tae 03-15-88 22:17:30 w i n of 10 41-80 *"/ 42 /» The f o l l o w i n g code s e t s the b a s e l i n e v a l u e s «/ • h i l e l H l t o i i r [01«rMlt to j i r [u , 1«» , ! , ) , l I p r i n t f ( ' \ n Z e r o 1/D Channels? * ) ; s c a n f ( * l s " , t o a i r ) ; I x z e r o = adi ( 8 , 0 , 1 ) ; y z e r o - adi ( 9 , 0 , 1 ) ; p l t z e r o = a d i ( 1 0 , 0 , l ) ; a r t z e r o = adi ( 1 1 , 0 , 1 ) ; rtile(i<200) I p m s x l i ) = x z e r o ; p r e s s y [ i ] - y z e r o ; i++; | i * 0 ; p r i n t f ( " \ n x z e r o I d y z e r o I d p l t z e r o I d a r t z e r o I d \ n \ x z e r o , y z e r o , p l t z e r o , a r t z e r o ) ; t o a i r ( 0 ] = V ; •hIl«[!((to a i r ( 0 ] = = ' y ' ) l i ( t o a i r [ 0 ] « T ) ) ) I p r i n t f ( " \ n S t o p c o c k s Closed? * ) ; s c a n f ( " l s \ t o a i r ) ; I ' " / t t t t t t t i i i i t i t t i t l t t t t t t t t n m t t t t i i t t t t t t i t I t i t i i M i l t u / o u t p o r t b (0x300, 0x18); I* Reset the X CY512 '/ outp o r t b (0x301, OxAE); /» Reset the Y CY512 '/ out p o r t b (0x303, 0 ) ; /* C l e a r X i n t e r u p t */ out p o r t b (0x304, 0 ) ; /* C l e a r Y i n t e r u p t »/ l o t o r y ( y p r e s s ) ; /* i n f l a t e the d i s t a l c u f f '/ • o t o r x ( x p r e s s ) ; / ' i n f l a t e the p r o x i i a l c u f f * / / i t t t i t t t i t t t t i i i l t t t t i t i t t i i t t t t i t t t t i t t t t t t t t t t t t t t t t t /• a d j u s t the p r e s s u r e s f o r the c o r r e c t p r o f i l e peaks »/ a d j u s t j r e s l O K y ' ; w h i l e ! (adju3t_pre6[0)=='y') I i ( a d j u s t j r e s [0}=='Y')) I o j p r e s s * 0; 03-15-18 22:00:26 \C86\COmOL\COmOL6.C Pg 3 Tue 03-15-88 22:17:30 l a i n of 10 11-122 81 o z p i e s s = 0; 82 i h i l e l o y p r e s s != y p r e s s ) 83 I 84 l o t o r y ( y p r e s s ) ; 85 p r i n t f H n Y p r e s s u r e I d " , y p r e s s ) ; 86 p r i n t f ( ' Bnter new p r e s s u r e " ) ; 87 oypress * y p r e s s ; 81 scanf H i " , b u f f e r ) ; 89 y p r e s s = a t o i ( b u f f e r ) ; 90 I 91 t h i l e l o x p r e s s !• z p r e s s ) 92 I 93 l o t o r x ( x p r e s s ) ; 94 p r i n t f ( " \ n X p r e s s u r e I d " , x p r e s s ) ; 95 p r i n t f I * E n t e r new p r e s s u r e ' ) ; 96 oxpress • x p r e s s ; 97 s c a n f f t s ' . b u f f e r ) ; 91 x p r e s s * a t o i ( b u f f e r ) ; 99 I 100 p r i n t f C \ n O o you want t o a d j u s t p r e s s u r e s again? ' ) ; 101 s c a n f l ' t s ' . a d j u s t j p r e s ) ; 102 I 103 104 / i t t t i i i i t M i t t i i t t t i t t t t t t U t i t t t H i t i i U t t t t U t i t H i t t 105 106 107 / t t i t t t t i t t t « i M t t t u t t t t i t t t t * u t t t * t » i i t » t u u n t m t t 101 /* ra s p down the p r o x i i a l b l a d d e r p r e s s u r e »/ 109 110 o u t p o r t b (0x300, OxAB); /* Reset the X C7512 */ 111 o u t p o r t b (0x301, 0x3E); /' I n h i b i t ! ' / 112 o u t p o r t b (0x300, OxBE); /' S e l e c t X */ 113 o u t p o r t b (0x303, 0 ) ; /' C l e a r X i n t e r u p t >/ 114 o u t p o r t b (0x304, 0 ) ; /* C l e a r 7 i n t e r u p t '/ 11$ sendx('B\nR l\nS l\nF 255\aA\n+\n"); 116 i f ( l i t e r s > 0) 117 s e n d x I W ) ; 111 e l s e 119 I 120 s e n d x f - \ n " l ; 121 x i t e r s = ( - l i t e r s ) ; 122 I 03-15-88 22:00:26 \C86\COITROl\COITROL6.C Pg I ?ue 03-15-88 22:17:30 lain of 10 123-161 123 sendx CE\n'); sendx CI '); itoa (xsteps, buffer); 121 sendx (buffer); sendx (a\nG\n! '); itoa (xdelaj, buffer); 125 sendx (buffer); sendx ("\nL "); itoa (xiters, buffer); 126 sendx (buffer); sendx (\3\n0\nQ*); sendx ('D\n'); 127 \2l / I t l t t t i t l i t t t i l i l l t l t l H t H t t l t t l t i i i t i t i t t t l t l t t t l U t t »/ 129 130 131 lookp»l; lookb=0; peakjralue«0; abo?e_threshold=0; 132 first_derivative»-l; average=0; iax_average=0; referenced; 133 peak[0]«0; peak(l]=0; peak[2)«0; 134 height[0]=20; height[1]=20; height[2]=20; 135 peak_address-0; ref_value=0; control=0; count=0; 136 threshold-10; one second=l; i * 0; n=0; j = 0; 137 138 / ' initialixe filter '/ 139 rtile(i<25) I filter_in[i)=0; i++; | i=0; 140 i i i l»(i<4) I filter out[i]=0; in ; | i=0; 141 142 143 / O M M t t t t O t t i H t U i i U t M i t t t t t t i i i t M t i i t t t t i i i t t t t t / 144 /* begin saipling */ 145 146 thile(i<8000) /» begin scan */ 147 ( 148 /'too long between peaks?'/ 149 count+t; 150 if(count>60) I counts; tbreshold=threshold/2; I 151 152 /'raip finished?'/ 153 if(inportb(0x300) & 0x20) referenced; 154 155 /'take a saiple'/ 156 filter in[j]--(adi(10t0,l)-pltzero); 157 158 /'filter the data'/ 159 filter out(0]=filter_out(l]; 160 filter_outlU«filter.outI21; 161 filter out[2]-filter out(3]; 162 if(j < 12) g=25-|12-j); 163 else g»j-12; 164 if(j < 24) h=25-(24-j); 03-15-88 22:00:26 \C86\COITROL\COmOL6.C Pg 5 Tue 03-15-88 22:17:30 l a i n of 10 165-209 165 e l s e h-j-24; 166 f i l t e r o a t ( 3 ) = ( 2 * f i l t e r o a t | 2 ) ) - f i l t e r _ o u t ( l l 167 • f i l t e r i n [ j l -168 I2»filter_in[g])+filter_in[h); 169 p l e t h l i ] " f i l t e r o u t [ 3 ] ; 170 i f ( j — 2 4 1 171 172 elie 173 174 175 176 / M t i i m i l t t t t i t t t t t t i t M i t t t i t t t i l t t t t t t t t i l t i t t t i i i l t / 177 /* not look f o r a peak */ 178 179 i f ( l o o k p ) 180 1 181 i f ( a b o v e threshold«0) 182 1 183 i f { f i l t e r out[3] > threshold) 184 I 185 peak_value«filter_out[3); 186 above t h r e s h o l d s ; 187 u t 1 i loo 189 i else 190 1 191 i f ( f i l t e r out[3] > threshold) 192 1 193 i f ( f i l t e r out[3) > peak value) 194 I 195 peak v a l u e = f i l t e r out(3]; 196 1 197 1 198 else 199 1 200 lookpsO; 201 lookb=l; 202 above t h r e s h o l d ^ ; 203 1 204 1 205 1 206 207 208 / t t l M t t i i t t t M i i t t t H t t t H i i t t i t i i t U t t i t l i M t t l i t t t H i 209 /' loo k i n g f o r a bate? 03-15-16 22:00:26 \CS6\COmOL\COmOL6.C Tue 03-15-11 22:17:30 l a i n Pg 6 of 10 209-250 */ 210 211 i f ( l o o k b ) 212 { 213 i f ( f i r s t d e r i v a t i v e < 0) 211 I 215 f i r s t d e r i v a t i v e = - f i l t e r out(01-2» f i l t e r o u t [ l ] 216 4 2 * f i l t e r o u t [ 2 ] + f i l t e r o u t [ 3 ] ; 217 I 211 e l s e 219 I 220 height[0]«height[l]; 221 h e i g h t l l ] h e i g h t [ 2 ] ; 222 h e i g h t ( 2 ] - p e a k _ v a l u e - f i l t e r . o u t [ 3 ] ; 223 nost= h e i g h t [ 2 ] ; 224 n e d i a n * h e i g h t [ l ] ; 225 l e a s t - h e i g h t [ 0 ] ; 226 i f | n o s t < n e d i a n ) 227 I 221 nost«nedian; 229 nedian«height(2]; 230 I 231 i f ( l e a s t > l o s t ) n e d i a n * n o s t ; 232 e l s e i f ( l e a s t >iedian) n e d i a n = l e a s t ; 233 threshold=nedian/4; 234 l o o k p = l ; lookb=0; count=0; 235 f i r s t _ d e r i v a t i v e = - l ; 236 h e i g h t _ o u t ( n ) = n e d i a n ; 237 h e i g h t a d d r e s s [ n ] * i ; 231 239 240 /•!»»»» t t » » t » t H l i » t t » i i » t i l H * » t * » » t » l » * » » i » * » * » * » * t M l 241 /' l o o k i n g f o r the r e f e r e n c e ? »/ 242 241 i f ( r e f e r e n c e ) 244 1 245 p e a k ( 0 ] = p e a k ( l ) ; 246 p e a k ( l ] = p e a k [ 2 ) ; 247 peak[2]«nedian; 241 a v e r a g e * ( p e a k [ 0 ] • p e a k [ 1 ] t p e a k | 2 ] ) / 3 ; 249 if(nax_average<average) nax_average= average; 250 r e f _ v a l u e * n a x _ a v e r a g e / 2 ; 03-15-88 22:00:28 \C86\COHTROL\CORROL6.C Pg 7 Toe 03-15-81 22:17:30 l a i n of 10 251-291 251 I 252 253 254 / i U i i M t t i t t t t i M i i t t i i t t t l i t t i t t t i M H t t H M t t t t i M I t i 255 /» i n the control node? '/ 256 257 i f ( c o n t r o l ) 258 I 259 /» notor still noving?'/ 260 x[n]=inportb(0x300)Mx40; 261 y[n|=inportb(0x301)l0x40; 262 i f ( ! y [ n l ) 263 ( 261 iedian=|height(0]+heightll]t height[2])/3; 265 error»height[2]-ref value; 266 i f (error > 0) 267 ( 268 sendx( at\n'); 269 sendy("+\n'); 270 I 271 else 272 I 273 sendxIMn*); 274 sendytMn'); 275 error s (-error); 276 I 277 k=2; 271 /'calculate hoi far to turn the notor'/ 279 nunber_x_steps=(error*343)/k/ nax.average; 210 /'turn the notor'/ 211 tenp=nunber_x_steps; 212 itoaltenp,buffer); 213 pressx{n]=adi(8,0,0|; 214 pressy[n]=adi(9,0,0); 215 s e n d x C l ' ) ; 216 sendx(buffer); 217 sendxC\nC\n"); 288 /'calculate hoi far to turn the notor'/ 319 nunber_y_steps=(error»281)/k/ naxjverage; 290 /'turn the notor'/ 291 tenp=nunber_y_steps; 03-15-88 22:00:26 \C86\COHTIOL\COITROL6.C Pg 8 Toe 03-15-88 22:17:30 l a i n o f 10 292-333 292 293 294 295 296 297 I 291 n H ; 299 I 300 I 301 302 303 / t i t t t t U l t t U i t M t M t i t M i i t M t t t t i t M i t i t t t t t t t t t t i M t »/ 301 /» f i n i s h e d the r e f e r e n c e ? on t o c o n t r o l »/ 305 306 i f ( ( r e f e r e n c e U ( c o n t r o l ) 307 I 308 p r i n t f ( ' i n f l a t i o n . . . \ n " ) ; 309 n o t o r x ( z p r e s s ) ; 310 l o t o r y ( y p r e s s ) ; 311 312 i i i l e ( ! ( i n p o r t b ( 0 x 3 0 0 ) I 0x20)) ; 313 314 p r i n t f ( " w a i t a s e c o n d . . . \ n " l ; 315 d e l a y ( o n e second); 316 317 c o n t r o l ^ ; 318 j « 0; lookp«l; l o o k b * 0 ; peak_value»0; 319 above t h r e s h o l d ^ ; c o u n t s , • 320 h e i g h t [ 0 ] * 2 0 ; h e i g h t [ l ] : 2 0 ; h e i g h t [ 2 ] * 2 0 ; 321 f i r s t _ d e r i v a t i v e = - l ; peak_address*0; t h r e s h o l d - 1 0 ; 322 t h i l t ( j < 2 5 ) ( f i l t e r i n [ j ] * 0 ; J « ; | j=0; 323 w h i l c ( j < 4 ) I f i l t e r o u t ( j H ; j H ; I j=0; 324 325 p r i n t f C o n t o c o n t r o l . . . \ n " l ; 326 /'speed up the n o t o r s * / 327 sendxI'R 200\nS 25\nF 43\nA\n+\nI l \ n 6 \ n ' ) ; 328 s e n d y C R 200\nS 25\nF 43\nA\n+\nN l\nG\n'); 329 I 330 331 / t i t t t t t t t M t i l i t M t t t i t i t m t M t i t t i t t t t l t t O i U t t i i t t t « / 332 i « j 333 I i t o a l t e n p , b u f f e r ) ; s e n d y C l ' I ; t e n d y ( b u f f e r ) ; s e n d y l ^ n C V n 1 ) ; 03-15-8! 23:00:26 \C86\COmOL\COITROL6.C Pg 9 Tue 03-15-81 22:17:30 l a i n o f 10 334-375 334 335 33 j / t t t t i l t t i t t l i t t t H l t t l t t t l t i U t t t l t t i t t t t i t t l t t t H U t H 337 / ' f i n i s h e d s u p l i n g ; d e l a y 40s; r a i p dorn the p r e s s u r e s */ 338 339 one_second=40; 340 d e l a y lone second); 341 342 xdelay»356; y d e l a y M 4 9 ; x i t e r s = - 1 5 6 ; y i t e r s = - 1 2 6 ; 343 iendx('H\nR l\nS l\nF 2 5 5 \ n i \ n * \ n " l ; 344 i f ( l i t e r s > 0) 345 s e n d x C A n " ) ; 346 e l s e 347 b e g i n 348 s e n d x C - \ n " ) ; 349 x i t e r s » l - x i t e r s ) ; 350 end 351 sendx C B \ n ' ) ; sendx C I ' ) ; i t o a ( x s t e p s , b u f f e r ) ; 352 sendx ( b u f f e r ) ; sendx C\nG\nX <); 353 i t o a ( x d e l a y , b u f f e r ) ; sendx ( b u f f e r ) ; sendx ( • \ n l ' ) ; 354 i t o a ( l i t e r s , b u f f e r ) ; sendx ( b u f f e r ) ; 355 sendx (\3\n0\nO,'); 356 357 o u t p o r t b (0x301, O x i E ) ; /* Reset the Y CY512 '/ 358 o u t p o r t b (0x300, 0x3E); /' I n h i b i t X 359 o i t p o r t b (0x301, OxBI); /' S e l e c t Y '/ 360 o u t p o r t b (0x303, 0 ) ; /' C l e a r X i n t e r u p t '/ 361 o u t p o r t b (0x304, 0 ) ; /' C l e a r Y i n t e r u p t '/ 362 sendy |"H\nR l\nS l\ n F 255\nA\n+\n"); 363 i f ( y i t e r s > 0) 364 sendy C+\n'); 365 e l s e 366 b e g i n 367 s e n d y C - \ n ' ) ; 368 y i t e r s = ( - y i t e r s ) ; 369 end 370 sendy C B \ s ' ) ; sendy (*I ' ) ; i t o a ( y s t e p s , b u f f e r ) ; 371 sendy ( b u f f e r ) ; sendy C\nG\nX ' ) ; 372 i t o a ( y d e l a y , b u f f e r ) ; sendy ( b u f f e r ) ; sendy ( "\nL • ) ; 373 i t o a ( y i t e r s , b u f f e r ) ; s e n d j ( b u f f e r ) ; 374 sendy C,3\n0\nQ'); 375 03-15-88 22:00:26 \C86\CORROl\COmOL6.C Tue 03-15-88 22:17:30 u i n Pg 10 of 10 376-397 376 outportb (0x305, OxBB); /» S e l e c t I & 7 ' / 377 s e n d j l'D\n'); sendx ("D\n'l; 378 379 / t i t t t l t l t l t i l i t t t t t l t t H l t l i t t t t l H t t i t l t t l H t i t t t t t t l t 380 /» write the inpedance and pressure saiples to files »/ 381 382 printf('writing files\n'|; 313 outputjilel=fopen("plethchk.prn",*w'); 38( iftloutput.filel) ( printf("could not open plethchk. prn\n"|; exitO; I 385 386 Otttput_file3=fopen('heightsl.prn\'w*); 387 if(!output file!) I printf('could not open heightsl. prn\n'); exitO; I 388 389 for(i=0;i<J000;iH) 390 fprintf(output_filelf'td\n\pleth[i]); 391 fclose(output file!); 392 393 forlH;i<n;i++) 394 fprintf(output_file3,'td *d\n",presix(i],pressy( 395 396 397 i l l ; fclose(output_file3); 

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:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0097900/manifest

Comment

Related Items