Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. DOES THE PROCESSING OF HYPNOTIC ANALGESIA REQUIRE ATTENTION-DEMANDING RESOURCES? A dual-task analysis of hypnotic-susceptibility-mediated differences in executive attentional processing between hypnotic and nonhypnotic analgesia A thesis presented in part fulfilment of the requirements for the degree of Master of Arts in Psychology at Massey University Louis Smeets 2000 Department of Psychology Massey University Palmerston North II DEDICATION This thesis is dedicated to the memory of Ernest R. Hilgard, Kenneth S. Bowers and Nicolas P. Spanos, who all sadly passed away in recent years. Their research efforts spanning several decades have contributed immensely to our understanding of clinical hypnosis in general and hypnotic analgesia in particular, and have inspired many others, myself including, to pursue further research in these areas. iii ACKNOWLEDGEMENTS I would like to thank my supervisor, Malcolm Johnson, for introducing me to research in pain mechanisms and management, for allowing me considerable freedom to pursue my own interests and directions, and for his valuable editorial comments. My sincere thanks also go to Dr. Steve Humphries for his helpful comments and encouragement during the early stages of this project. Daniel Hean for developing the software programme for the tracking task. Blair Vernal for assisting with some of the initial screening sessions. Dr. Bill Zika who, despite his busy schedule, made time to record the hypnotic inductions used during the initial screening sessions on tape. Or. John Spicer for his sound advice on data analysis, and Dr's. Eric Woody and Auke Tellegen for their clarification on queries about the Waterloo-Stanford Group C Scale of hypnotic ability and the Tellegen Absorption Scale. I am also grateful for the support received from Barry Andrews who was always helpful and patient during the many requests for available classroom or lecture- theatre space to hold the hypnotic-susceptibility screening sessions. And finally, but most importantly, all the participants who freely gave, what became for many, very substantial amounts of their valuable time. This research project was approved by the Human Ethics Committee at Massey University (HEC96/45) and was funded by a combined grant from the Psychology Department and the Graduate Research Fund. Dedication Acknowledgments Contents List of figures List of tables Preface Abstract Introduction CONTENTS Chapter One Pain mechanisms and management 1.1. Development of pain theories and treatments 1 .2. Advances in the understanding of nociceptive transmission 1.3. The Gate-Control theory of pain 1.4. Current understanding of pain 1.4. 1 . Phasic and tonic pain systems - first and second pain 1.4.2. The affective-emotional component of pain experience 1.4.3. Plasticity in the CNS and the development of acute and chronic pain. 1.4.3.1. Different pain states, changes in response patterns, and underlying neurophysiological mechanisms 1.4.3.2. Changes in neurotransmitter systems 1.4.3.3. Functional plasticity in the CNS 1.4.3.4. Plasticity at the gene-expression level 1.4.4. The diathesis-stress model of chronic pain 1.4.5. Pain in absence of sensory input - the neurosignature 1.4.6. The neurophysiology of pain - pain and the brain 1.5. Influence on pain management iv Page ii Ill iv xiii xvii xix 1 3 4 8 14 14 16 21 21 27 29 31 38 40 44 55 Chapter Two Psychological strategies for coping with pain 2.1. Behavioural and cognitive-behavioural approaches 2. 1 . 1 . Behavioural strategies 2.1.2. Cognitive strategies 2.1.2.1. Focusing 2.1.1.2. Distraction 2.1.3. Factors affecting the efficacy of cognitive-behavioural pain-coping strategies 2.1.4. 2.1.5. 2.1.6. Evidence for effectiveness Anxiety The cognitive costs of attention diversion and thought suppression 2.1. 7. Distraction versus sensation monitoring 2.2. Hypnotic analgesia 2.2.1. Characteristics 2.2.2. Evidence for effectiveness 2.2.3. Mechanisms Chapter Three Attention, multiple-task performance, and interference v Page 62 62 62 63 63 63 65 66 67 69 70 73 73 74 76 3. 1. Attention processing 85 3.1.1. Control of stimulus selection: bottom-up biases and 86 top-down control 3.1.2. Models of attention 3.2. Multiple-task performance 3.3. Task interference 3.4. Pain and attentional interference 3.4.1. Direct access to awareness 3.4.2. The interruptive quality of pain 3.4.2.1. Alternative research methodologies: The task paradigm 3.4.2.2. Factors that may modulate the interruptive quality of pain 88 92 94 96 96 97 99 99 VI Page 3.5. Processing of emotion-arousing information and efficacy 103 of emotion-arousing distractors 3.6. Neurophysiology of attentional processes 105 3.6.1. Posterior and anterior (fronto-limbic) attentional systems 105 3.6.2. Major functions of attention and their anatomical substrates 106 3.6.2.1. Orienting 106 3.6.2.2. Selection 107 3.6.2.3. Vigilance 3.6.3. Controlled and automatic processing: Event-related brain potentials studies of the different phases of attentional control with evidence for brain structures and neurotransmitter systems involved 3.6.3.1. Arousal and familiarisation to novel input: the habituation of the orienting reaction 3.6.3.2. Activation and readiness: the maintenance of targeted orienting 3.6.3.3. Comfort and effort: innovative attention 3.6.3.4. Summary Chapter Four Consciousness 107 108 109 112 113 116 4.1. The search for a model of consciousness 118 4.2. Global Workspace theory 121 4.2.1. Why need for competition for access to the Global Workspace? 125 4.2.2. States of absorption 125 4.2.3. Hypnosis 126 4.2.4. Dissociation 127 4.2.5. Compliance and involuntariness 4. 3. Neurobiological interpretation of Global Workspace theory 128 129 4.3.1. Support for access competition 129 4.3.2. Evidence of widespread dissimination of conscious information 130 Chapter Five Hypnosis 5. 1 . The nature and characteristics of hypnotic phenomena 5.2. Hypnotic susceptibility and its assessment 5.2.1. The development and stability of hypnotic susceptibility 5.2.2. Evidence for a genetic contribution 5.2.3. Relationship to main personality dimensions 5.2.4. Assessment and factor-analytic structure of hypnotic suscepti bi I ity 5.3. Absorption 5.3.1 . Correlates of absorption 5.3.2. Absorption and altered states of consciousness vii Page 132 136 136 137 138 139 143 143 144 5.3.3. Possible moderators of the correlation between absorption and 145 hypnotic susceptibility 5.3.4. The spectral analysis of hypnotic performance with respect 146 to absorption 5.3.5. Dealing with the inherent conflict between suggested behaviour 152 and actual experience 5.4. Dissociation 154 5.5. Unconscious influences in hypnosis 157 5.5.1 . Conscious and unconscious influences 157 5.5.2. Efficacy of direct versus indirect suggestions 159 5.5.3. Conclusions from experimental studies using the "real-simulator" 162 control design 5.6. Models of hypnotic responding 164 5.6.1. Neodissociation theory 165 5.6.1.1. Hilgard's neodissociation model of hypnosis 165 5.6.1.2. The dissociated-experience model 168 5.6.1.3. The dissociated-control model 172 5.6.2. The social psychological model 179 5.6.2.1. Effortful goal directed actions and misattribution of 179 experience viii Page 5.6.2.2. Further results from real-simulator control studies 182 and conclusions reached 5.6.2.3. Critique of the social-psychological view of hypnosis 190 5.6.3. Predictions regarding attentional involvement 5.6.4. Signs of rapprochement: sociocognitive and dissociation explanations may apply to different ends of the continuum of hypnotic responding 5.7. The neuropsychophysiology of hypnosis 191 193 198 5. 7 .1. Evidence of neuropsychological changes during hypnosis 198 5.7.2. Evidence of neurophysiological changes during hypnosis 202 5.7.2.1. Do people with high and low hypnotic susceptibility 203 differ in brain activity during the nonhypnotic state? 5. 7.2.2. Are there changes in brain activity that distinguish 205 hypnosis from the nonhypnotic state? 5.7.2.3. Do people with high and low hypnotic susceptibility 206 show fundamental differences in the patterns of brain activity during hypnosis? 5.7.3. A neuropsychophysiological model of hypnosis 210 5.7.3.1. Summary of main findings leading to its development 210 5.7.3.2. Crawford et al's neuropsychophysiological model of 212 hypnosis Chapter Six The current study 6.1. Miller and Bowers (1986, 1993) 6.2. Design of the current study 6.2.1. Methodological issues 6.2.1.1. Secondary attention-demanding task 6.2.1.2. Method of experimental pain stimulation 6.2.1.3. Experimental design 214 216 216 217 217 218 6.2.2. Comments on the rationale for the hypothesised interference 220 effects 6.3. Hypotheses 223 Method Chapter Seven Screening for Hypnotic Ability 7. 1. Introduction 7.2. Initial Assessment 7.2.1. Subjects 7 .2.2. Measures 7.2.2.1. Harvard Group Scale of Hypnotic Susceptibility: Form A 7.2.2.2. Tellegen Absorption Scale 7.2.2.3. Toronto A!exithymia Scale 7.2.2. Procedure 7.3. Follow-up Assessment 7.3.1. Subjects 7.3.2. Measure - Waterloo-Stanford Group C Scale 7.3.3. Materials 7.3.4. Procedure Chapter Eight Experimental Stage 8. 1 . Experimental design 8.2. Subjects 8.3. Apparatus 8.3.1. Pain stimulus delivery 8.3.2. Tracking task and controi of pain stimulus delivery 8. 4. Measures 8.4.1. Pain ratings (intensity and unpleasantness) 8.4.1.1. Pain intensity 8.4.1.2. Pain unpleasantness 8.4.2. Absorption in tracking task 8.4.3. Strategy use 8.4.4. Hypnotic depth IX Page 225 226 226 227 227 228 230 230 231 232 232 234 234 237 238 239 239 241 243 244 244 245 245 245 246 8.5. Procedure 8.5.1. Familiarisation 8.5.2. Assessment of individual pain levels - the pain-threshold and pain-tolerance trials 8.5.3. Main experiment Results 8.5.3.1. Nonhypnotic analgesia condition 8.5.3.2. Hypnotic analgesia condition Chapter Nine Tracking-Performance Data 9.1. Introduction 9.2. Tracking-performance data 9.2.1. Accuracy of input and missing data 9.2.2. Violations of assumptions 9.2.2.1. Normality 9.2.2.2. Homogeneity of variance 9.2.2.3. Transformation of variables 9.2.2.4. Random sample and independence of scores 9.2.3. Analysis 9.2.3.1. Data from full tracking wave for all three trials combined 9.2.3.2. Exploration of tracking-wave differences 9.2.3.2. Data from middle parts of tracking wave for all three trials combined 9.2.3.4. Data for middle parts of tracking wave for trial one only 9.3. Summary of main findings x Page 248 248 249 251 251 252 254 255 255 255 256 256 257 259 260 260 266 268 272 278 Chapter 10 Pain ratings and other data collected 10.1. 10.2 10.3. 10.4. 10.5. Pain ratings 10.1.1. Pain intensity 10.1.2. Pain unpleasantness Absorption in tracking task Strategy use Hypnotic depth Summary of main findings Chapter Eleven Hypnotic ability screening data 11.1 Harvard Group Scale of Hypnotic Susceptibility: Form A 11.2 Tellegen Absorbtion Scale 11 3 Waterloo-Stanford Group C Scale Discussion Chapter Twelve Discussion xi Page 280 280 284 288 291 293 299 301 303 305 12.1 Review of hypotheses and findings 308 12.1.1. Hypotheses 1 and 2 - Differences in tracking performance 308 (i.e., attentional interference) 12.1.2. Discussion 310 12.1.2.1. Accuracy of hypnotic susceptibility assessment 310 12 .1 . 2. 2. Actual hypnotic status during experimental session 312 12.1.2.3. Pain stimulation 313 12.1.2.4. Reliance on attentional effort and sensitivity of the 314 tracking task 12.1.3. Hypotheses 3 and 4 - Pain ratings 317 12.1.4. Discussion 318 12.2. General discussion 12.2.1. Attentional capture of visual motion 12.2.2. Methodological issues 12.2.3. Subsequent support for the dissociated-control model of hypnosis 319 319 320 323 12.2.4. Evidence arguing against the notion that the process of hypnosis does not require executive attentional effort 12.2.5. Final conclusion and recommendation References Appendices Hypnotic Susceptibility Screening - Initial Assessment 1. Subject Information Sheet 2. Consent Form Hypnotic Susceptibility Screening - Follow-up Assessment 3. Subject Information Sheet 4. Consent Form Experimental Session 5. Subject Information Sheet 6. Consent Form 7. Medical Checklist xii Page 327 330 333 428 430 431 432 433 434 435 xiii UST OF FIGURES Page 1. Superfiscial layers of the dorsal horn receiving postsynaptic fascilitation 11 from tactile fibres and inhibition from nociceptive fibres. Deeper layers receiving postsynaptic facilitation from both noxious A5 and nonnoxious Aa and AB fibres. Substantia gelatinosa neurons presynaptically inhibit projection neurons. Output of projection neurons crosses to the other side of the spinal cord and ascends up the spinal column. 2. Simplified schematic representation of the gate-control theory of pain transmission 3. 4. 6. 7. 8. Major ascending and descending pathways in the central nervous system relating to nociception. Schematic representation of information processing in the neuromatrix. Relationship between basic constructs of Global Workspace theory with widely used equivalent terms. The spectrum of hypnotic performance. The spectral analysis of hypnotic performance with respect to absorption. Item-characteristic curve for Communication Inhibition suggestion, HGSHS:A. 12 13 42 122 146 147 148 9. Item-characteristic curve for Music Hallucination suggestion, WSGS. 149 10. Diagram depicting competition for limited resources between the attention- 220 demanding requirements of pain processing and tracking-task performance. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. Diagram depicting effective use of hypnotic suggestions by subjects with high hypnotic susceptibility, as hypothesised. Diagram depicting the use of nonhypnotic strategies to cope with painful stimulation, while simultaneously performing an attention­ demanding tracking task. Experimental design showing all treatment combinations. Electrode placement for iontophoretic pain stimulation, showing the anode bowl on the volar surface of the subject's arm. Full tracking wave and block of three pain-stimulus trials for each condition. Visual analogue scale for subjects' rating of pain intensity. Difference in mean tracking-deviation scores between pain-off and pain-on situations for all treatment conditions by hypnotic ability. Data representing the full tracking wave for all three stimulus trials combined. Difference in mean tracking-deviation scores between treatments, for pain-on situations at both pain levels by hypnotic ability. Data representing the full tracking wave for all three stimulus trials combined. Between-treatment comparison of mean tracking-difference values, with 95% confidence intervals, at both pain levels by hypnotic ability. Data representing the full tracking wave for all three stimulus trials combined. Section of the tracking wave showing the division between extreme and middle parts of the waveform. xiv Page 221 222 237 241 243 244 263 264 266 267 xv Page 21. Difference in mean tracking-deviation scores between pain-off and pain-on situations for all treatment conditions by hypnotic ability. Data representing the middle parts of the tracking wave for all three stimulus trials combined. 269 22. Difference in the mean tracking-deviation scores between treatments, 270 for pain-on situations at both pain levels by hypnotic ability. Data 23. 24. 25. 26. 27. representing the middle parts of the tracking wave for all three stimulus trials combined. Between-treatment comparison of mean tracking-difference values, with 95% confidence intervals, at both pain levels by hypnotic ability. Data representing the middle parts of the tracking wave for all three stimulus trials combined. Difference in mean tracking-deviation scores between pain-off and pain-on situations for all treatment conditions by hypnotic ability. Data representing the middle parts of the tracking wave for trial one only. Difference in mean tracking-deviation scores, for pain-on situations at both pain levels by hypnotic ability. Data representing the middle parts of the tracking wave for trial one only. Between-treatment comparison of mean tracking-difference values, with 95% confidence intervals, at both pain levels by hypnotic ability. Data representing the middle parts of the tracking wave for trial one only. Mean pain-intensity ratings for both hypnotic-ability groups in all treatment conditions. 272 274 275 277 283 28. Mean pain-intensity ratings, with 95% confidence intervals, for both 284 hypnotic-ability groups in treatment conditions with tracking task on. xvi Page 29. Mean tracking-unpleasantness ratings for both hypnotic-ability groups 287 in all treatment conditions. 30. Mean pain-unpleasantness ratings, with 95% confidence intervals, 288 for both hypnotic-ability groups in treatment conditions with tracking task on. 31. Mean ratings for absorption in the tracking task, with 95% confidence 290 intervals, for both hypnotic-ability groups in all treatment conditions. 32. Mean pain-intensity ratings for all treatment conditions by hypnotic 296 depth. 33. Mean pain-unpleasantness ratings for all treatment conditions by 298 hypnotic depth. 34. Distribution of item-pass percentages for HGSHS:A. 302 35. Distribution of item-pass percentages for WGSC by hypnotic-ability 307 group. 1. 2. 3. 4. 5. 6. 7. LIST OF TABLES Contrasting capabilities of conscious and unconscious processes. Item difficulty, estimated item reliability, and biserial correlation with absorption for all items of HGSHS:A and WSGC. Correlation of trait and situational variables with hypnotic responsiveness as a function of hypnotisability leveL Summary based on Kirsch, Silva, Corney, and Reed (1995), experiment 4. Stimulus delays and ramp rates for pain threshold-tolerance trials. Means and standard deviations of tracking-deviation scores in pain-off and pain-on situations for both hypnotic-ability groups in all treatment conditions. Data representing the full tracking wave for all three stimulus trials combined. Mean tracking-performance scores for the difference between pain-off and pain-on situations. Data representing the full tracking wave for all three stimulus trials combined. Means and standard deviations of tracking-performance scores for both hypnotic ability groups in all treatment conditions. Data representing the middle parts of the tracking wave for all three stimulus trials combined. xvii Page 119 148 151 250 262 265 269 8. Mean tracking-performance scores for the difference between pain-off 271 and pain-on situations. Data representing the middle parts of the tracking 9. wave for all three stimulus trials combined. Means and standard deviations of tracking-performance scores for both hypnotic-ability groups in all treatment conditions. Data from the middle parts of the tracking wave for trial one only. 273 10. 11. 12. 13. 14. 15. 16. 17. 18. Mean tracking-performance scores for the difference between pain-off and pain-on situations. Data representing the middle parts of the tracking wave for trial one only. Means and standard deviations of subjects' pain-intensity ratings for all treatment conditions by hypnotic ability. Means and standard deviations of subjects' pain-unpleasantness ratings for all treatment conditions by hypnotic ability. Means and standard deviations for absorption in tracking-task ratings for all treatment conditions by hypnotic-ability group. Number of times subjects in both hypnotic-ability groups used deliberate pain-coping strategies in each treatment condition. Means and standard deviations of subjects' pain-intensity ratings for all treatment conditions by hypnotic depth. Means and standard deviations of subjects' pain-unpleasantness ratings for all treatment conditions by hypnotic depth. Number and percentage of subjects that passed each item on HGSHS:A. Item-pass percentages and rank order for current sample of the HGSHS:C, and normative data from Australian, Californian, and Harvard samples. xviii Page 276 282 286 289 292 294 297 301 303 19. Item-pass percentages and rank order of item difficulty for the current 305 sample of WGSC, a normative sample for WGSC, and a comparative sample using the SHSS: C xix Page 20. Item-pass data for WGSC by hypnotic-susceptibility group. 306 21. Means and standard deviations of pain reduction scores. 325 From Eastwood, Gaskovski, and Bowers (1998). PREFACE For most people, myself included, the first contact with hypnosis is through watching a demonstration of hypnosis either on television or as part of a live audience. Such performances can be quite spectacular and are certainly entertaining. However, to me they remained interesting stage phenomena, much like those performed by a skilled magician, and did not arouse any scientific interest into hypnosis. Being interested in pain management thought, one sooner or later comes across accounts of hypnotic analgesia, that is the use of hypnotic suggestions to achieve relief of pain and distress. Whatever ones opinion about hypnosis, it has been thoroughly proven, both in clinical and experimental settings, that some people do achieve significant and clinically important benefits when using hypnotic suggestions over and above those available to the average person using nonhypnotic coping methods. The interesting question, which has been debated by some researchers for decades, is what are the processes whereby hypnotic analgesia is achieved, and are these fundamentally different from those involved in the execution of nonhypnotic coping strategies. More recently, pain research has increasingly emphasised the role of attention in pain processing and in particular the ability of pain to have priority access to processing resources and dominate conscious processing at the expense of other activity. This interference with other ongoing activity is one of the major pain-related handicaps experienced by people with chronic pain. The concept of attention, and in particular the distinction between controlled and automatic processing, is crucial to an understanding of both pain processing and hypnosis and provides an important and fascinating approach for studying the two. This thesis was written as part of a study investigating differences in attentional interference effects between hypnotic and nonhypnotic analgesia. The main hypothesis tested is whether: "Hypnotic analgesia, unlike nonhypnotic pain-coping strategies, can be achieved without reliance on high-order (executive) attentional resources and therefore results in no or only minimal interference with other ongoing and attention-demanding activities. xx xxi A proper understanding of this topic and the wider context wherein it occurs requires some knowledge of the following key aspects: (1) pain and pain management; (2) pain coping strategies; (3) attention, and in particular access to, and interference with , attentional resources in multiple task environments; (4) consciousness; and (5) hypnosis. The introduction to this thesis follows this outline. Chapters one to six form the introduction. Some sections of the introduction provide additional and more in-depth information (particularly on the neurophysiology of pain, attention, and hypnosis) that are useful for a fuller appreciation of these topics and can assist the reader in understanding how these main aspects are linked together. However, strictly seen these are not necessary for a direct understanding of the main research question. For the convenience of the reader, these sections are marked with a red asterisk ( G ) following the section heading. They include sections 1.4, 3.6, 5.2, 5.5, 5.6.2.2., 5.7, and all of chapter 4. Chapter 1 provides an overview of the main aspects of our current understanding of pain processing and the control mechanisms involved. Particular reference is given to inhibitory control processes descending from cortical and subcortical brain structures. As Chapter 6 will show, there is evidence for hypnosis-related differences in the effectiveness of such inhibitory control mechanisms. The final section briefly covers how the advances in pain research have influenced pain management practices. Section 1.4 G provides more in-depth information on the affective-emotional dimension of pain, on the changes that take place when a pain condition becomes chronic, and on the neurophysiology of cortical and subcortical brain structures involved in pain processing and responding . It is in the management of chronic pain where hypnotic analgesia may have its greatest advantage. The specific question of the current study is part of an underlying research effort to enhance our understanding of pain mechanisms and derive at more effective methods for the control of particularly chronic pain. Developments in the area of neurophysiology are leading to a more specific understanding of the theoretical mechanisms of pain, which in turn contributes to the development of more specific and effective pain management practices. It is for this reason that these areas are given substantial coverage in the introduction of this thesis. xxii Chapter 2 "Psychological methods of pain control" consists of two main parts. The first part introduces behavioural and cognitive pain coping strategies, highlights factors that may influence their utility, and evaluates evidence for the effectiveness of such strategies. It then describes some of the influences of anxiety on pain responding, outlines the cognitive costs of using attention diversion and pain suppression strategies, and contrasts the effectiveness of attention diversion versus sensation monitoring strategies. The second part describes the main characteristics of hypnotic analgesia, looks at both clinical and experimental evidence for its effectiveness, and outlines proposed mechanisms whereby hypnotic analgesia may reduce pain. Chapter 3 "Attention, multi-task performance, and task interference." This chapter briefly describes the main characteristics of information processing: competition for limited capacity processing resources, and the selection of information for further processing. It continues with a description of how this latter process is influenced by both bottom-up stimulus-driven biases and by top-down control. It then highlights the development of models of attention with particular reference to Shiffren and Schneider's (1977) distinction between controlled and automatic processing, and outlines the main components of Norman and Shallice's (1986) hierarchical model of supervisory attentional control. This is followed by a description of how interference and the demands of concurrent task performance are treated by traditional limited­ capacity models of attention, and by models based on multiple resource theory. The next section describes the interruptive quality of pain, its specific (hard-wired) capacity to capture attention, and factors that may moderate the interruption of ongoing activity. This is followed by a brief section on biases in the processing of emotion-arousing information and preliminary findings regarding the efficacy of distraction tasks with an emotional theme. Section 3.6 G deals with the neurophysiology of attention. This section reviews the different dimensions of attentional processing and their anatomical correlates, including arousal and targeted readiness which are also important aspects in pain processing, and novelty which is important for attentional capture and effective distraction strategies. xxiii Particular attention is given to the mechanisms involved in the control of attention and findings supporting the existence of anterior and posterior attentional systems. This section provides a summary of the background knowledge that has led to the development of the neuropsychophysiological model of hypnosis described in section 5.7.3.2. Chapter 4 briefly introduces the topic of consciousness and relates conscious and unconscious processes with respectively controlled and automatic attention. It does so with particular reference to Bernard Baars' Global Workspace theory of consciousness and briefly describes how this conceptualisation can be used to explain such phenomena as hypnosis, absorption, dissociation and involuntariness. Many actions and processes are either well established (learned and familiarised) or may be programmed (hard-wired) as is the case with pain so that they, once activated, can be executed on an automatic and subconscious level. As will be covered in section 5.6, some researchers and theorists argued that hypnosis is one of these processes. Chaper 5 "Hypnosis" starts with a description of the nature and characteristics of hypnotic phenomena, and the factors that may contribute to the experience of hypnosis. The next three sections deal more in-depth with the three main factors: hypnotic susceptibility, absorption, and dissociation respectively. Section 5.6 compares and critically evaluates the main models of hypnosis: the dissociated experience and dissociated control models and the social-psychological model of hypnosis, and the predictions they make regarding the involvement of attention. The next part highlights some more recent findings that indicate that the opposing views of social psychological and special process (i .e. dissociation) explanations both appear to apply, but at different ends of the continuum of hypnotic responding. Section 5.2 G covers the assessment and measurement of hypnotic susceptibility. Experimental studies of hypnosis phenomena commonly use scores on standard hypnotic susceptibility scales as the criterion for allocating subjects to experimental conditions on the basis of their hypnotic ability. This section explores the argument as to how well such measures capture the important components of hypnotic responding. xx iv This is relevant because there is increasing support for the notion that (1) individual differences in hypnotic responding reflect differences in kind (i .e., underlying mechanisms) rather than in dimension (i.e., position along the continuum of a single trait), and (2) there exist subsamples of highly hypnotisables exhibiting distinct patterns of responding and brain activity that are not differentiated by the standard hypnotic susceptibility tests which treat highly hypnotisables as a homogeneous group. Section 5.5 G "Unconscious influences in hypnosis" indicates how human behaviour in general , and hypnotic responding in particular, can be influenced by information that is perceived and processed outside of normal conscious awareness. It highlights how the social psychological explanation of hypnosis emphasises the importance of Type I unconscious influences such as demand characteristics, expectancies, and social compliance; but, unlike the dissociation model of hypnosis, denies the influence of Type II unconscious influences involving genuine alterations in the way information is processed such as the down-regulation of nonessential functions and a shift towards increased primary process thinking. This section also reviews experimental research into the relative efficacy of direct and indirect hypnotic suggestions, and highlights how a type of control experiment called the real-simulator design can be used to assess the influence of demand characteristics. Section 5. 7 G It is the area of neurophysiological research that provides important new insights in the, otherwise largely stagnated, debate about the mechanisms underlying hypnotic responding in general and hypnotic analgesia in particular. This section reviews neurophysiological evidence for fundamental changes in brain activity that: (1) can distinguish the hypnotic from the nonhypnotic state, and (2) can distinguish between individuals with low and high hypnotic susceptibility in each of these states. It concludes with a summary of a neuropsychophysiological model of hypnosis that is based on the result of these studies. Chapter 6 "The current study" starts with a description of the two studies by Miller and Bowers (1986; 1993) that form the basis for the current experiment. xxv This is followed by a description of the main aim of the current study and the ways in which the methodology was changed in an effort to increase the sensitivity of the design and allow for greater specificity when analysing the effects of the experimental manipulation. The last section outlines the specific hypotheses of the current study. Chapters 7 and 8 make up the method section for respectively the hypnotic susceptibility screening stage and the experimental part of the study. Chapters 9, 10, and 11 comprise the results section. Chapter 9 covers the results of analysis of tracking performance data relating to the main research question. Chapter 10 lists the results of the assessment of pain intensity and pain unpleasantness ratings as well as data on subjects' level of absorption, strategy use, and hypnotic depth. Chapter 11 briefly summarises the results of hypnotic susceptibility measurements during the screening part of the study. The data in Chapters 10 and 11 does not directly relate to the main research question, but does provide additional information used in interpreting the results and supports the arguments made and conclusions reached in the discussion section. Chapter 12 starts with a discussion of results relating to the main research question (hypotheses 1 and 2) and evaluates possible reasons for the absence of hypothesised differences in interference effects. The effectiveness of the tracking task is examined with reference to the characteristics of attentional capture of visual motion, and recommendations are made for future research and improvements to the current design. This is followed by a discussion of the analyses of pain ratings (hypotheses 3 and 4 ). Finally, recent research of attentional processing during hypnosis is evaluated with particular emphasis on neuroimaging studies providing direct measures of localised cortical activation during hypnosis and performance of attention-demanding tasks. 1 ABSTRACT There is substantial evidence that hypnotic analgesia can be effective in reducing pain and distress in both experimental and clinical settings in at least a sizeable portion of the population. However, the mechanism whereby hypnosis achieves this are not well-understood and various explanations have been proposed. These offer fundamentally different predictions about the attentional involvement of hypnotic analgesia, which are highly relevant to pain research as the disruption of ongoing activity is one of the more debilitating aspects of pain. While cognitive-behavioural coping strategies may attenuate pain of short duration, their effortful deployment further interferes with ongoing activity, and there are strong indications that their effectiveness rather rapidly decreases as pain perseveres. If, as dissociated-control theory proposes, hypnotic analgesia does not require attentional effort for its execution , it would provide significant advantages for individuals who can effectively achieve it (i .e. , those who are highly susceptible to hypnotic suggestions). This hypothesis was further tested in an experimental study using a dual-task scenario and repeated-measures design. One hundred and ninety student volunteers were first screened for hypnotic susceptibility using the Hardvard Group Scale of Hypnotic Susceptibility: Form A, and seventy-eight also completed a more demanding follow­ up assessment using the Waterloo-Stanford Group C scale. This resulted in fifty individuals who qualified for participation in the experimental part of the study by scoring as either high or low hypnotisable on both these measures. Of these, 12 lows and 14 highs went on to take part in an experimental study that had high and low hypnotisables performed a cognitively demanding tracking task while using either hypnotic analgesia or cognitive-behavioural strategies to cope with iontophoretically administered pain. Interruption of tracking performance during each coping method was used as a measure of central attentional resources needed to execute that coping strategy. Results did not find evidence for the hypothesised absence of interference effects among high hypnotisables using hypnotic analgesia. Possible reasons are examined and exploration of data indicates that the tracking task was not difficult enough to require significant and continuous attention, and lacked sensitivity to distinguish interference effects between treatment conditions. 2 Findings do not allow a conclusion of support for either explanation of the mechanisms underlying effective hypnotic analgesia. Highly hypnotisable subjects using hypnotic analgesia did achieve significantly greater reductions in both the intensity and unpleasantness of the pain than low hypnotisables using hypnotic analgesia or high and low hypnotisables using cognitive-behavioural coping strategies. Characteristics of the attentional capture of visual motion are discussed and suggestions made for future research and improvements to the design of the current study. Considerable attention is given to findings of a large body of neurophysiological studies of brain activity and a proposed neuropsychophysiological model of hypnosis. When combined, results of these studies indicate that the mechanisms of attentional control involved in the process of hypnosis are fundamentally different from those involved in the use of standard cognitive­ behavioural strategies, but that both processes do require central attentional effort and resources. 3 INTRODUCTION CHAPTER ONE PAIN MECHANISMS AND MANAGEMENT 1.1. Development of Pain Theories and Treatments Our understanding of pain and the mechanisms that influence the experience of pain has increased enormously, specially during the second half of this century. Till then, pain was seen to result from a straight-through sensory projection system (Melzack, 1993; Melzack & Wall, 1965). Specificity theory proposed that pain perception resulted from activation of receptors that only responded to intense noxious stimulation and transmitted impulses in a direct line to an exclusive pain centre in the brain. The amount of pain experienced was seen to be directly proportional to the amount of noxious stimulation or tissue damage. Pain was solely a function of sensory input and varied according to the quality and intensity of the sensory stimulus. Accordingly, the majority of clinicians adhered largely to a medical model that treated pain as a disease. As a result, pain treatment approaches consisted mainly of attempts to remove the pain stimulus, either pharmaceutically (e.g., by administering analgesics, narcotics, or nerve blocks) or through reparative or destructive surgical procedures. Yet, such medical and surgical interventions proved to be effective in no more than 50% of chronic pain patients (Weisenberg, 1977). Psychological factors were dismissed as merely reactions to pain, and comorbidities were generally ignored (Long, 1994). Patients who reported pain for which there was no apparent organic cause were frequently told it was all in their head, and were dismissed as "frauds" or send to psychiatrists (Merskey & Chandarana, 1992). Observant clinicians, however, noticed that this conceptualisation of pain did not account for some phenomena they frequently observed in practice. 4 Surgical lesions made at almost every level of the central and peripheral nervous system had proven unsuccessful in abolishing the severe pain that frequently follows trauma or lesions to peripheral or central nerve fibres (e.g., neuralgia, phantom limb pain, causalgia; Melzack & Wall, 1965). These factors argued strongly against a strict direct-line, stimulus-response model of pain perception. A new theory was needed that could account for, among others, such phenomena as: individual differences in pain responding; pain that occurred or persisted long after the stimulus was removed or the injury had healed; pain referred to unrelated areas where no pathology existed; and the observation that low-threshold non-nociceptive stimuli, such as light touch, did at times trigger severe pain. Several versions of pattern theory emerged out of this quest. All included some form of central summation of stimuli, which was proposed to take place in the dorsal horns of the spinal cord. Levingstone (1943) suggested that continued stimulation set up reverberating circuits in the dorsal horns. Activity evoked in such structures could then be triggered by normally non-nociceptive stimuli and generate the abnormal firing patterns that are interpreted centrally as pain. Noordenbos (1959) proposed that large-diameter fibres inhibit small-diameter fibres and that the substantia gelatinosa in the superficial dorsal horn plays an important role in this central summation and input control. These theories were still rather vague and none included an explicit role for the brain other than as a passive receiver of messages. Nevertheless, they moved the theoretical conception of pain in the right direction and introduced the idea that nociceptive transmission can be modulated and is influence by both nociceptive (slow) and non-nociceptive (fast) fibre systems. 1.2. Advances in the Understanding of Nociceptive Transmission It is useful at this stage to distinguish between nociception and pain, as the two are not synonymous. Nociception refers to neural activity in the pain-mediating nervous system that is evoked by activation of nociceptors. Nociceptors are free nerve endings that are preferentially sensitive to stimuli that are damaging (noxious) to normal tissues or to stimuli that would become noxious if prolonged (International Association for the Study of Pain, 1986, p. S220). 5 Pain is a subjective experience based on cognitive mechanisms that are influenced by the level of arousal in this nociceptive system (Bromm, 1995). Pain, like other sensations, can be modulated by a wide range of behavioural experiences. Not all nociception is experienced as pain, and nociception itself is neither a necessary nor a sufficient condition for the conscious experience of pain. It is not until the brain interprets the noxious stimulation, and provides it with an affective evaluation, that nociception becomes pain. The two main types of nociceptors in humans and other mammals are: (1) mechanical and thermal nociceptors, which are activated by high intensity mechanical stimuli (e.g., pinprick or cut) or noxious heat or cold (> 45°C or < 5°C); and (2) polymodal nociceptors, which react to chemical events activated by tissue damage resulting from high-intensity mechanical , thermal , or chemical means (Bowsher, 1989). Nerve fibres can be classified from A to D according to the speed at which they conduct nerve impulses (Tyrer, 1992). Nociceptors transmit impulses along small-diameter axons. In the case of mechanical and thermal nociceptors, these are thinly myelinated, relatively fast-conducting AS fibres; whereas for the polymodal nociceptors they are unmyelinated, slow-conducting C fibres. Mechanoreceptors form another type of receptors. These are activated by low­ threshold non-nociceptive stimuli (e.g. , touch or light pressure) and transmit stimuli along large-diameter, myelinated, fast-conducting Aa and A,Bfibres. Thus, the small­ diameter nociceptive (AS and C) fibres convey high-threshold pain and thermal sensations whereas the large-diameter non-nociceptive (Aa and A,8) fibres convey low-threshold mechanical sensations. Together they are referred to as primary afferents. There is a fourth category of nociceptors called silent receptors who are not normally activated by noxious stimulation, but whose firing threshold can be dramatically reduced by inflammation or various chemical insults. The dorsal (posterior) horn of the spinal cord (see Figure 1) can be divided into several layers, called the laminae of Rexed, based on structural differences and their function in the transmission and modulation of afferent information. These laminae are interconnected and exhibit a high degree of interaction. For an in-depth description of nociceptive transmission and dorsal-horn organisation see, for example, Brown (1981 ), Willis (1985) or Willis and Coggeshal (1978). 6 Both nociceptive and non-nociceptive primary afferents enter the spinal cord via de ventrolateral division of the dorsal (posterior) horn. Within the spinal cord they ascend for one or two segments and then terminate on to neurons in several laminae in the dorsal horn. From here they projection to second-order projection neurons also referred to as transmission or tract (T) cells which relay the incoming sensory information along five major pathways or tracts to higher centres in the brain. The lateral spinothalamic tract is concerned with the transmission of well-localised discriminative pain and temperature sensations. The anterior spinothalamic tract mainly carries light touch stimuli. Recent evidence suggests that some of its fibres ascend ipsilaterally all the way to the midbrain, where they cross in the posterior commissure and project primarily on intralaminar neurons in the thalamus, with some fibres reaching the periaqueductal gray matter of the midbrain. This tract is suggested to convey aversive and motivational nondiscriminative pain sensations, although some question its existence as a separate entity (Afifi & Bergman, 1998). The spinoreticular and spinomesencephalic tracts are also important in pain transmission. The latter, consisting of axons of neurons in laminae I and V, is thought to contribute to the affective component of pain (Basbaum & Jessel!, 2000). The spinothalamic tract ascends throughout the spinal cord and brainstem to project on neurons in the ventral posterior lateral (VPL) nucleus of the thalamus. Axons of the VPL neurons further project to the somatosensory cortex. Various neurotransmitters and neuromodulators are involved in the transmission and modulation of nociceptive stimulation. The majority of primary sensory neurons in the dorsal horn release glutamate, an amino acid that functions as a rapidly acting excitatory neurotransmitter. In addition, many dorsal-horn neurons reacting to small­ diameter (A8 and C) fibres also release neuropeptide neurotransmitters, notably substance P (SP), somastatin (SOM), and vasoactive intestinal peptide (VIP), which are believed to mediate slow synaptic transmission (Afifi & Bergman, 1998). Substance P and glutamate are released from the intraspinal terminals of nociceptors following noxious stimulation and have an excitatory action on dorsal-horn neurons and facilitate pain transmission (Dickenson, 1996a). Excitatory neuropeptides such as SP enhance and prolong the actions of glutamate. 7 Nociceptive activity can be inhibited by descending pathways containing serotonergic fibres from the nucleus raphe magnus and nucleus gigantocellularis of the medulla oblongata, noradrenergic fibres from the nucleus locus coerulus in the rostral pons and caudal midbrain, and enkephalinergic fibres from the periaqueductal gray matter in the midbrain (Afifi and Bergman, 1998). They transport, among others, endogenous opiate peptides that are released following activation of various brain structures. The main families of endogenous opiates are endorphins, enkepkalins, and dynorphins. They are potent inhibitors of pain receptors. Met-enkephalin and SOM inhibit the release of SP from primary afferents thereby inhibiting activity in dorsal horn neurons. A number of other neuropeptides are also involved in the modulation of pain transmission including neurokinins, galanin, and calcitonin gene­ related peptide (CGRP). While know to modulate pain mechanisms, the exact functions of many of these neurotransmitters/modulators are still largely uncertain (Dray, 1996). For a more thorough description of the function and distribution of neurotransmittters and neuropeptides see, for example, Strand (1999). Glutamate is likely to be the main neurotransmitter of the projection neurons (Dickenson, 1996a). Some of the projection neurons are nociceptive specific (i.e. , they are only excited by A8 or C fibres) . Other projection neurons in the deeper laminae (particularly lamina V), called wide-dynamic-range (WDR) or convergent neurons, receive input from both high-threshold nociceptors that facilitate pain perception and low-threshold mechanoreceptors that normally inhibit pain transmission. Noxious stimulation of A8 and C fibres results in the high-frequency discharge of WDR neurons associated with the perception of pain, whereas tactile stimulation of Aa and A/] fibres normally results in only low-frequency activation (Abram, 1990). The WDR neurons have receptors for both SP and glutamate and the interplay of these receptors determines the excitatory component of spinal pain transmission. One of the subtypes of glutamate receptors is the N-methyl-D­ aspartate (NMDA) receptor. Low-frequency stimulation of C fibres produces an acute constant response of the WDR neurons. The NMDA receptor does not appear to participate in this normal baseline response. However, repeated (i.e. , high-frequency) noxious stimulation sensitises the WDR neurons and lowers their threshold so that they then also respond to low-threshold stimulation. 8 This results in a dramatic increase in both magnitude and duration of the neuronal response, a phenomenon referred to as 'wind-up' (Diamond & Coniam, 1997; Dickenson, 1996a). The wind-up of WDR neurons is NMDA receptor dependent. Ion channels of NMDA receptors are normally closed by voltage-dependent magnesium (Mg2+) blocks, but the release of peptides and glutamate, combined with a postsynaptic depolarisation, opens the positively charged ion channels and allows for an influx of calcium ions (Ca2+). This sets off several intracellular actions that lead to a persistent increase in the excitability of the WDR neuron (see Carstens, 1996; Dickenson, 1996b). Under these conditions, even gentle non-noxious stimulation of normal skin (e.g., the rubbing of clothes against the skin), which produces only a low­ level afferent barrage, can result in the high-level spinal output that is experienced as pain. This pain condition is known as allodynia, and is one example of the plasticity in the behaviour of central nervous system (CNS) neurons. There are indications that such NMDA events play a crucial role in the maintenance of prolonged pain, which is frequently accompanied by hyperalgesia (i.e., increased sensitivity to noxious stimulation) and allodynia (Dickenson, 1996a). There is clear evidence for the effectiveness of NMDA antagonists in inflammatory pain, neuropathic pain, and experimental models of allodynia (e.g., Yaksh, 1989). NMDA antagonists (anti­ hyperalgesics) differ from conventional analgesics in that they prevent or block the hypersensitivity, but leave baseline responses unaffected. 1.3. The Gate-Control Theory of Pain In the early 1960's, neurophysiological studies started to provide evidence for modulation of dorsal-horn neurons (Jessel & Kelly, 1991). Melzack and Wall (1965) integrated the earlier work by Noordenbos and others with the new evidence for neural modulation at the spinal level and developed it into a theory that could account for many of the previously confusing clinical observations. They called this the gate-control theory. Although a lot of detail has been added to the initial formulation of this theory, its original propositions remain largely intact today. However, rather than seeing this as an empirically proven theory, it is more helpful to view it as a conceptual model that provides a useful representation of the vastly complex and only partially understood neural interactions that take place at the spinal level. 9 When entering the dorsal root, the fast conducting A5 fibres excite synapses of dorsal-horn neurons in laminae II and IV, while some also terminate on laminae I, III, and V The slow-conducting C fibres terminate on and excite neurons in particularly laminae II, but some also project to laminae I and III. The substantia gelatinosa (lamina II) is made up almost exclusively of inhibitory and excitatory interneurons. The more inner laminae (particularly lamina V) are increasingly excitable by non­ noxious, low-threshold, stimulation from Aa and AJ3fibres. The majority of neurons in laminae V are WDR neurons which receive low-threshold input from large-diameter A/3 fibres as well as, both direct and indirect, input from the small-diameter nociceptive afferents (A5 and C fibres). Some of the AS fibres provide direct input to projection neurons in the marginal layer (lamina I) which also receive indirect input from C fibres via stalk-cell interneurons in lamina II. There exists a cascading interaction (inhibition and excitation) between neurons in these progressively deeper layers within the dorsal horn which, with the exception of lamina 11, all send fibres to the brain (Diamond & Coniam, 1997). Lamina V contains primarily second-order projection neurons that send information up the spinal cord to the brainstem and thalamus for further processing. What will be transmitted upwards depends on the summation of informatioin received from the more dorsal laminae, direct input from periheral stimulation, and central inhibition descending from higher up in the nervous system. The combined interaction of all these factors determines the integrated firing pattern of the projection neurons. When this exceeds a critical preset level, it triggers a sequence of responses in what Melzack and Wall (1965) referred to as the Action System. This involves many areas of the brain including specialised systems that are involved in the sensory-discriminative, affective­ motivational, and cognitive-evaluative dimensions of pain. It also directs motor pathways that facilitate motor responses including somatic and autonomic activity. The entire system operates through complex feedforward and feedback mechanisms (see Figure 3, p. 13). Furthermore, nociceptive signals can also be modulated at successive synaptic relays along the central spinal pathway up the dorsal columns. The entire nervous system exhibits a high level of integrity, particularly with regard to pain sensation. 10 The gate-control theory of pain transmission proposes a control system, located in the substantia gelatinosa (lamina II) of the superficial dorsal horns, that modulates the flow of nerve impulses from peripheral fibres to higher centres in the CNS. This gate-control mechanism is sensitive to the level of activity in both nociceptive and non-nociceptive fibres, and is modulated by the interaction of excitatory and inhibitory interneurons plus inhibitory control pathways that descent from higher centers. The following is a brief description of the updated version of the gate-control mechanism. Both small- and large-diameter fibres project into the substantia gelatinosa (SG) which is suggested to be the modular center for pain. The small­ diameter, nociceptive AS and C fibres directly (postsynaptically) inhibit SG neurons, whereas the large-diameter Aa and A/3 fibres directly facilitate (excite) these SG neurons. The large-diameter, non-nociceptive fibres and small-diameter, nociceptive A5 fibres also provide postsynaptic excitatory input to the T cells (see Figures 1 and 2). Neural connections in the substantia gelatinosa can modulate the transmission of primary sensory afferents unto the second-order projection cells by regulating (closing) the gate in two basic ways. Normally there is always some ongoing stimulation carried by tonic, slowly adapting fibres which keeps the gate open. The discharge of these large-diameters fibres initially will fire the T cells through the direct route and then partially close the gate through their activation of SG neurons which provide presynaptic inhibition of T cells. Substance P, the probable neurotransmitter of the C fibres (Bowsher, 1989, Tyrer, 1992), is found abundantly in dorsal-root ganglion neurons of C fibres and in the areas of the dorsal horn where they terminate. It facilitates transmission in second­ order projection neurons (T cells), opens the gate, and increases the perception of pain. Nociceptive A5 fibres directly activate projection neurons, and C and A5 fibres also inhibit activity of the SG cells. In the case of the latter, this occurs both directly and by some segmental collaterals that activate enkephalin-releasing interneurons (IE, see Figure 3). J) c. c de~cenclin~ \n\,i'b ·,\or) c.ont!'o\ ~ri no - t\.i c:i\CJIY\i(. t r <:tc.1 spino - ~et; c.u\~r ,.('QC.t 11 Figure 1. Superficial layers of the dorsal horn receiving postsynaptic facilitation from tactile f ibres and inhibition from nociceptive fibres . Deeper layers receiving postsynaptic facilitation from both noxious Ao and non-noxious fibres . Substantia gelatinosa neurons presynaptically inhibit projection neurons . Descending control and some A,B collaterals act ivate inh ibitory interneurons. Output of projection neurons crosses to the other side of the cord and ascends up the spinal column. However, collaterals of the large-diameter, non-nociceptive A/3 fibres , whose main axons ascend the dorsal columns, can activate GABA-energic interneurons that presynaptically inhibit the central terminals of nociceptive C fibres, thereby reducing their capacity to secrete the SP that would normally inhibit the SG cells . Thus, one way of closing the gate is through non-nociceptive (low-threshold high­ frequency) input from large-diameter tactile fib res. Therefore, rubbing the skin or applying gentle heat to the area around a painful spot can alleviate the pain . This process is used therapeutically by techniques such as massage, transcutaneous electrical nerve stimulation (TENS) , or vibratory stimulation (Wall & Sweet, 1967; Woolf, 1989). Aa+ Ap c DC Ap DESCENDING INHIBITORY CONTROL Ao ~~~~~~~----"-:..__~~~~~~~~~~-- T = projection neuron (T cell) SG = substantia gelatinosa neuron IE = enkephal in releasing interneuron IG = gaba-ergic inhibitory interneuron - = facilitation - = inhibition Figure 2. Simplif ied schematic representation of the gate-control theory of pain transmission. 12 A second and much more powerful way involves closing the gate from the inside through descending inhibitory control. Prior to reaching the SG, the main axons of some of the very fast conducting A/3 fibres branch off and travel up the dorsal columns and connect with certain structures in the medulla, pons, and periaqueductal grey matter of the brainstem (see Figure 3) . Neurons in these brainstem structures have ascending projections to the thalamus, which modulates pain responding (Lenz, 1992). When activated either directly or by low-threshold high-frequency peripheral stimulation , they release endogenous opioids that activate descending inhibitory pathways (see Dickenson, 1994). Descending inhibition can reduce or prevent nociceptive signals from being transmitted upwards to conscious levels of the brain (see Basbaum & Fields, 1984; Fields & Basbaum, 1978, 1989; Willis , 1985, 1995). This feed-forward loop enables these central control centres to activate descending inhibitory control (feedback loop) before activity in the projection neurons has triggered the action system (Bullingham, 1985, Melzack & Wall , 1965). grey matter locus ceruleus fast Ap fibres descending inhibition MIDBRAIN PONS reticular formation MEDULLA spinothalamic tract spinoreticular tract SPINAL CORD Figure 3. Major ascending and descending pathways in the central nervous system related to nociception. 13 The superficial dorsal horn contains a high density of enkephalin- and dynorphin­ containing inhibitory interneurons. Enkephalins have similar properties to morphine, they modulate the effect of SP and consequently tend to close the gate and reduce pain (Tyrer, 1992). Descending neurotransmitters and some collaterals from Aofibres excite these inhibitory interneurons, whose enkephalin-releasing terminals inhibit activity of the C fibres (see Figure 1 ). This occurs both presynaptically, by decreasing their ability to release SP, and postsynaptically, by hyperpolarising the membrane of the dorsal-horn neurons and increasing membrane permeability thereby reducing the size of the excitatory postsynaptic potential (see Netter, 1983, p. 160). Descending axons of serotonergic, noradrenergic, and enkephalinergic fibres also exert direct inhibitory control through contacts with the dendrites of projection neurons. 14 There is another way that descending control can reduce the activity of the WDR projection neurons, namely through noxious stimulation elsewhere in the body at sites that are distant from the excitatory receptive fields of these WDR neurons. This is referred to as diffuse noxious inhibitory control (DNIC). It operates through a control loop that is independent from the descending inhibitory control that originates from structures in the midbrain and medulla that are part of the endogenous pain­ inhibition system mentioned above (LeBars, Bouhassira, & Villaneuva, 1995). The DNIC loop involves the subnucleus reticulus dolaris, which contains neurons that are exclusively responsive to nociceptive stimuli and have "whole-body" receptive fields. These send descending projections that are relayed through the dorso-lateral funiculus and terminate in the dorsal horn at all levels of the spinal cord. 1.4. Current Understanding of Pain 0 1.4.1. Lateral and medial pain systems - phasic and tonic pain The two types of nociceptive afferents (A8 and C) are not only activated by distinct types of stimuli , but transmit impulses along separate pathways and facilitate distinct types of pain responses . These are referred to as phasic or first pain and tonic or second (slow) pain . Phasic or first pain is activated by stimulation of mechanical or thermal nociceptors. These are located in the skin , subcutaneous tissues, and around muscles and joints and respond to pinprick or phasic heat stimuli (> 45°C or < s°C). They transmit impulses along A8 fibres that have a conduct velocity of approx. 5-25 mis (Adriaensen , Gybels, Handwerker, & van Hees, 1983; Holmes, 1990). Phasic pain , which occurs immediately following an injury, has a mean latency of 240 msec (Bromm, 1995). It is brief, has a sharp stinging or pricking quality, and can rapidly rise and fall in intensity. It causes phasic withdrawal reflexes designed to quickly withdraw from the emergency and avoid further damage. 15 Melzack (1990) suggest that the lateral pain-signalling system is most active during such phasic pain. The tracts of the lateral system travel upward along both sides of the brainstem's central core and project onto the sensory cortex. Thus, activation of the lateral system can give rise to rapid, sudden, sharp pain in clearly identified sites. Phasic or first pain provides information about the sensory qualities of the pain including the location, duration, and intensity of the noxious stimulus (Price, 1976). Following injury, the body's own opioids activate the descending control system, which quickly dampens activity in the lateral system. This rapid inhibition is necessary from an evolutionary and biological standpoint. In situations where other demands are particularly urgent and dominant (e.g., life threatening), powerful stress­ induced analgesia is activated (Bolles & Fanselow, 1980). This enables an animal to defend itself or escape from danger, rather than become overwhelmed by fear or the pain of the injury and, therefore, an easy prey. In humans it manifests itself, for example, in the fact that people often report little or no pain immediately following a serious injury, or can rescue somebody else while being apparently oblivious of their own injuries. Only in such exceptional circumstances of extreme environmental stress can severe pain be relegated and momentarily lose its attention demanding character. As the imminent danger diminishes, pain will again demand attention. Tonic pain is activated by tissue damage relayed by polymodal nociceptors that are distributed in tissues throughout the body with the exception of the CNS. Stimuli are transmitted along small-diameter, unmyelinated, C fibres that have a slower conduction velocity of about 0.5-2 m/s (Adriaensen et al. , 1983; Holmes, 1990). Tonic or second pain, therefore, has a mean latency of approx. 1.200 msec (Bromm, 1995). It has a burning or throbbing quality, and causes tonic muscle contractions that strongly motivate us to recuperate and refrain from activities that could further aggravate the injury. They promote natural healing and limit the spread of infection. The medial pain signalling system is thought to be largely responsible for tonic pain that can linger long after the initial injury has occurred, and is seen to control the affective-motivational component of pain (Melzack, 1990). Its tracts ascend through the central core of the brainstem and send impulses to structures in the limbic system which influences emotions. 16 Because of its slow conduction velocity, the medial pain signalling system is not well­ suited for immediate responses. Rather, activation of the medial system results in the emotional responses that influence actions taken following the initial threat. Tonic or second pain can in many cases be completely or significantly suppressed by psychological factors or peripheral skin stimulation (Price, 1976). In summary: The lateral pain system is proposed to include the primary (S1) and secondary (Sn) somatosensory cortex, and be involved in the processing of sensory­ discriminative aspects of pain. The medial pain system is proposed to include limbic structures such as the anterior cingulate cortex and the mid and anterior insula, and be involved in the processing of the affective-motivational aspect of pain. 1.4.2. The affective-emotional component of pain experience So far, I have mainly dealt with nociception; however, this by itself is neither necessary nor sufficient for the experience of pain. The neural activity is only the initiator that opens the door to the conscious experience of pain. Emotional distress is a fundamental part of the pain experience. It can be a cause of pain, a consequence of pain, and a state concurrent with pain (Craig, 1993). The International Association for the Study of Pain (!ASP) clearly acknowledged the importance of affective distress in pain by defining pain as: "an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage" (IASP, 1986, p. 8217). The sensory perception of pain (nociception) would not be experienced as pain without the evaluation of emotional distress. It is not so much the sensory intensity of pain, but rather its negative emotional quality and disruption of normal ongoing activity that lead to suffering. Although it is useful to distinguish between nociception and pain (Fields, 1987), this should not lead to dualistic thinking as the two are highly interrelated and are not amenable to a tidy separation (Fernandez & Turk, 1992). The sensory experience may give rise to the emotional distress of pain, but emotional distress can amplify the sensory experience. Emotions are seen as adaptive and motivating organisers of experience and behaviour (Chapman, 1996). 17 Emotions compel action and expression, and have an important role in memory and learning processes. What is stored in long-term memory depends heavily on its emotional value. Memories of emotional events have a powerful influence on current and future behaviour and expectations. Associated cues can bring them back to immediate and heightened attention. Emotional events are also highly effective reinforcers (positive or negative) that influence operant learning (Fordyce, 1990). The emotional reaction to pain is related to the perceived threat of the situation. Perceived threat to our biological integrity results in heightened arousal and a generalised stress response that prepares us for immediate and effective actions needed to safeguard our wellbeing (e.g., the fight or flight response, withdrawing from injurious behaviour, and/or efforts to seek help). Fear conditioning supports survival by fostering avoidance of potentially dangerous situations. However, the fear accompanying pain can become associated with non-noxious stimuli through classical conditioning and lead to avoidance of activities that are not painful in themselves, but have become associated with pain. Furthermore, environmental stimuli , including the behaviour of others, can become reinforcers or punishers that either directly or indirectly reward pain behaviour (e.g., attention from others, possibility of financial compensation) or punish it (e .g. , loss of mobility and possibly loss of job and/or status) . These influences can become powerful shapers of coping behaviour that either enhance recuperation or lead to maladaptive coping styles and produce chronicity of the pain condition. Emotional experiences are not simply universal (i .e. , inherent and immutable), but are to an important extent determined by societal and cultural influences (Mesquita and Frijda (1992). There is reasonable evidence to suggest that the experience of pain is at least partially a social construction (Craig, 1986). The complex interactions of emotions, previous pain experiences, cultural and societal influences, and subjective expectations and interpretations influence whether and how a particular person experiences pain (see e.g. , Craig, 1995; Crawford Clark & Bennett Clark, 1980; Levine & De Simone, 1991; Otto & Dougher, 1985; Zatzick & Dinsdale, 1990). 18 There is both clinical and experimental evidence for the interaction of emotions and pain. Romano and Turner (1985), for example, observed a bidirectional relationship between pain and depression. Depression can provoke pain by increasing pain sensitivity and reducing pain tolerance, and pain can serve as a stressor that evokes subsequent depression. Zelman, Howland, Nichols, and Cleeland (1991) found that tolerance to cold-presser pain was reduced by viewing depressive statements and enhanced by viewing reduced elation statements. Thus the affective-emotional response to injury or tissue trauma is a complex interaction of the perceived threat of the situation; the level of physical arousal and resulting stress response; and the combination of learned behaviours, memories of previous experiences, expectations, and social influences. The discrimination between sensory and affective pain processing first arises at the dorsal horn of the spinal cord. From here sensory information is transmitted along the spinothalamic tracts, while information destined for affective processing ascends the spinal cord along the spinoreticular tract (Chapman, 1995). The afferent pathways of the spinoreticular tract arrive at different levels of the brain stem from where they project along four major pathways to various areas in the neocortex. These reticulocortical pathways use various neurotransmitters, all of which play a role in the complex experience of emotion during pain. The noradrenergic pathway, consisting of a dorsal and a ventral noradrenergic bundle, is most closely linked to negative emotional states (Gray, 1987). The dorsal noradrenergic bundle (DNB) is the ascending projection from the locus coeruleus (LC). The DNB has widespread projections throughout the limbic system and neocortex. Through these; the LC, which accounts for about 70% of all noradrenaline in the brain, may exert an almost global influence on brain activity (see Chapman, 1995). The LC reacts to stimuli that threaten or signal damage to the biological integrity of the individual. These do not have to be nociceptive specific, but nociception inevitably increases activity in LC neurons. Enhanced activity in the LC and DNB results in negative emotional arousal and hypervigilance that can even progress into panic. 19 The DNB is the mechanism for vigilance and orientation to affectively relevant and novel stimuli . It also regulates attentional processes and facilitates motor responses (Foote & Morrison, 1987; Gray, 1987). This enables the individual to exercise global vigilance for harmful or threatening stimuli and to respond quickly and effectively when needed. Excitation of the LC is also evident in anaesthetised animals and does not appear to require cognitively mediated attention (Chapman, 1996). The emotional experience of pain is thus linked with the awareness of immediate biological threat and shares central mechanisms with vigilance. The ventral noradrenergic bundle (VNB) innervates the hypothalamus. The medullary reticular formation projects through the ventral noradrenergic bundle (VNB) to the hypothalamus (Bonica, 1990b). These projections supply up to 90% of all the catecholaminergic innervation to the hypothalamus and provide the major neurophysiological link between tissue trauma and the hypothalamic response (Chapman, 1996). The hypothalamus plays an important role in the affective response to pain. The medial and lateral hypothalamus receive input from nociceptive projection neurons at all levels of the spinal cord (Burstein, Cliffer, & Geisler, 1988). The paraventricular nucleus (PVN) of the hypothalamus is an important coordinating centre that links nociceptive input with autonomic arousal and hormone release in the hypothalamus, pituitary and adrenal cortex also referred to as the HPA axis (see Chapman, 1995). The PVN evokes autonomic arousal through neural as well as hormonal pathways, and neural activation and hormone release in HPA structures is regulated by intricate feedback mechanisms. The HPA axis takes executive responsibility for coordinating behavioural readiness with physiological capability, awareness, and cognitive function (Chapman, 1996). The PVN receives afferent information from other nuclei in the hypothalamus, several nuclei in the reticular system including the LC, and from the hippocampus and amygdala. The PVN synthesises corticotrophin-releasing hormone (CRH), and in response to afferent input that signals threat or injury (e.g., noxious stimulation) secrets CRH into the portal circulation. This, in turn, stimulates the anterior pituitary, which leads to the release of several neuropeptides including adrenocorticothrophic hormone (ACTH) into the systemic circulation (see Chapman, 1996, p. 71 ; Strand, 1999, pp. 170-174). 20 ACTH stimulates the adrenal medulla to release adrenaline and noradrenaline and corticosteroids such as hydrocortisone and corticosterone. Corticosteroids inhibit the inflammatory response and provide inhibitory feedback to regulatory processes that can return immune system disturbances to normal, preventing excessive responses to stress that could produce self-injury, that is, autoimmune disease {Strand, 1999). The widespread activation of excitatory and inhibitory processes produce a complex adaptive stress response that involves both neural and endocrinologic changes. Activation of structures in the HPA axis can alter the normal balance in endocrine function to help ready the organism for the extraordinary behaviours that will maximise its chances to cope with the threat at hand (Selye, 1978). Intricate feedback mechanisms help to regulate hormone activity and re-establish normal homeostasis again. If the spinoreticular activation is particularly strong, these processes can lead to powerful stress-induced analgesia that minimises peripheral sensitisation and attenuates nociceptive signalling, thereby helping the organism to cope with the threat without the distraction of severe pain. When the activation persists after the immediate threat is gone and is not, or cannot be, controlled with medication, the stress response will reduce, but some low to moderate level of stress response action is likely to remain. When this persists for days, weeks, or longer, the prolonged dysrhythmia of the HPA axis may become counterproductive and, in particular, disrupt circadian rhythms. This is evident in the frequent reports of sleep disturbances, fatigue, general lethargy, poor concentration, loss of appetite, and diminished sexual interest by chronic pain patients, although some may undoubtedly also result from side effects of medication taken. These consequences of HPA axis dysrhythmia contribute to the development and maintenance of depression that is a frequent comorbidity of chronic pain (Chapman, 1996). In summary, tissue trauma results in: (1) activation of both spinothalamic and spinoreticular pathways, (2) the concurrent generation of sensory and affective processes that subserve adaptive functions, (3) activation of predominantly noradrenergic structures in the limbic system to produce the affective dimension of pain, and ( 4) in a hypothalamically mediated stress response that plays a role in pain chronicity (Chapman, 1996). 21 Thus, pain is not synonymous with nociception, and cognitive and affective factors play a crucial role in the experience of pain. Accordingly, interventions for the management of pain should target both emotional and sensory processes (Chapman , 1995). Exactly how the nociceptive activity becomes conscious and induces feelings of pain (the mind-body problem) remains still largely unresolved (Crick & Koch, 1992). For further reading on the particular utility of hypnosis in integrating mind and body aspects of pain and healing, see Rossi and Cheek (1988). 1.4.3. Plasticity in the CNS and the development of acute and chronic pain 1.4.3.1. Different pam states, changes in response patterns, and underlying neuroph ysiological mechanisms There are many different pain conditions or syndromes, each with their own specific physiological and psychological characteristics (e.g ., low back pain, neuralgia, migraine, visceral pain, deafferentation pain, cancer pain, fibromyalg ia etc) . An understanding of their underlying mechanisms, combined with psychosocial information, is invaluable in determining the most effective pain-management program for each pain patient or patient group. There are different criteria along which the various pain conditions can be grouped into broader categories, for example: (i) acute and chronic pain; (ii) nociceptive and nocigenic pain; or (iii) Phase 1, Phase 2, and Phase 3 type pain, which combines the previous two classifications. Acute pain sensations are usually related to a specific and locally definable injury. They evoke a generalised pain response that involves various characteristic physiological and behavioural changes such as elevation of blood pressure, peripheral vasoconstriction, inhibition of gastrointestinal activity, glycogen release by the liver to increase blood sugar, and extreme anxiety or hostility. These are intended to alert the organism to impending damage and prepare it for rapid response. 22 Acute pain normally subsides gradually. It responds well to analgesics and usually does not develop tolerance to them (Bullock, 1996; Swonger & Matejski, 1988). Chronic pain is prolonged, defined by persistence, and often progresses in severity as pathology increases. It is generally less clearly localised and described (Swonger & Matejski, 1988). It is now well-understood that chronic pain is not simply acute pain that lingers on, but involves distinct characteristics, neural changes, and pain behaviours that require different treatment approaches. Nocigenic or nociceptive pain results from direct activation of nociceptors due to tissue damage caused by trauma, inflammation, or disease, while neurogenic or neuropathic pain is due to lesions of nerves in the peripheral or central nervous system and can occurs in the absence of nociceptor stimulation. Neurogenic pain includes pain conditions such as deafferentation and central pain syndromes (e.g., reflex sympathetic dystrophy, phantom limb pain, postherpic neuralgia, thalamic pain) and often has a burning or electric sensation (Basbaum & Jessel!, 2000). Cervera and Laird (1996) have proposed a conceptual framework that can be used to understand the changes in neurophysiological mechanisms that occur along the transition from acute to chronic pain states. They have identified three major stages or phases of pain, which they refer to as Phase 1, Phase 2, and Phase 3 pain. This classification is based on the changes in the nature and time course of the originating stimuli and the subsequent fundamental changes in response properties of various components of the nociceptive system. The three pain phases range from nociceptive pain in a stable, intact, and normally functioning nociceptive system to more abnormal, neuropathic, pain states. Although each phase has its own distinct underlying neurophysiological mechanisms, these are not mutually exclusive and several mechanisms may coexist in the same individual. Phase 1 pain: Phase 1 pain is of short duration and relates to the processing of brief noxious stimuli (i.e., acute pain of nociceptive origin). It is characterised by a close relationship between the level of activity in peripheral nociceptors and the subjective experience of pain. Experimental pain research has predominantly studied Phase 1 pain. 23 The features of Phase 1 pain can best be explained by models based on specificity, that is a direct transmission along neural circuits that specifically process simple noxious events with possible modulation occurring at synaptic relays along the way. Localised, brief, noxious stimulation does not result in the extrasynaptic spread of neuropeptides (volume transmission) in dorsal-horn neurons that is evident following inflammation or prolonged noxious stimulation (Sandkuhler, 1996). Phase 2 pain: Phase 2 pain is persistent and relates to intense and prolonged noxious stimulation resulting from tissue damage or inflammation (i .e., chronic pain of nociceptive origin). Repeated, high-frequency stimulation not only greatly increases afferent inflow, but also changes the sensitivity of nociceptive neurons both peripherally and centrally. Phase 2 pain is characterised by this shift in stimulus­ response function that is triggered and maintained by repeated, high-frequency input and results in increased excitability. Prolonged noxious stimulation and inflammation also activate the release of various neuropeptides that modulate postsynaptic transmission. Sandkuhler and colleagues have observed a synchronisation of discharges in multireceptive neurons following inflammation that was not present following acute, noxious stimulation. (Sandkuhler, 1996). Synchronisation of converging neurons is a powerful mechanism whereby information transfer may be strengthened. While, SP superfusion significantly enhanced background activity, by itself it did not result in synchronisation of discharges in converging neurons. Peripherally, there are two main changes in the response pattern of nociceptors that contribute to Phase 2 pain. First, a lowering in threshold to the extent that stimulation in the area of injury that would normally only be mildly painful results in an excessive response and becomes much more painful (primary hyperalgesia; LaMotte, Shain, Simone, & Tsai , 1991 ). Mildly painful stimulation to normal skin surrounding the injured area may also become more painful (secondary hyperalgesia; Cervera, Meyer, & Campbell , 1994; LaMotte et al. , 1991 ; Torebj6rk, Lundberg, & LaMotte, 1992). This sensitasation of nociceptors after injury or inflammation results from the release of a variety of chemicals including bradykinin, histamine, prostoglandins, leukotrines, acetylcholine, serotonin and SP by the damaged cells and tissue in the vicinity of the injury (see Basbaum & Jessell , 2000). 24 Second, this increased responsivity is further enhanced by the appearance of spontaneous activity. Together they result in a continuous barrage of nociceptive afferent input that is likely to contribute to the development of spontaneous pain (Cervero & Laird, 1996). There is evidence that certain noxious stimuli can evoke not only an acute excitation of nociceptors, but also cause prolonged sensitisation of both "normal" and "silent" nociceptors (for a review see Cervera, 1994). Central sensitisation also plays a role in the development of Phase 2 pain. Repeated, high-frequency, peripheral stimulation also sensitises nociceptive neurons in the spinal cord so that they become more easily excited. Activity in low-threshold mechanoreceptors (e.g., by light touch) is now able to evoked not only tactile sensations, but also the high-frequency output that is associated with pain. This condition is known as allodynia. While allodynia only occurs in responds to a stimulus (i.e., patients with allodynia feel no constant pain), patients with hyperalgesia often perceive pain spontaneously (Basbaum & Jessel!, 2000). This development of spontaneous pain, allodynia, and hyperalgesia is dependent on ongoing afferent input from the site of injury and can be abolished by the injection of a local anaesthetic into the site of peripheral nociceptive activity (Gracely, Lynch, & Bennett, 1992). The knowledge that certain pain states activate subsequent hyperalgesia triggered the idea of using pre-emptive analgesia (i.e., analgesia administered before a painful procedure) to reduce, for example, postoperative pain (Wall, 1988). This is an interesting concept, particularly if it would also apply to neuropathic pain. So far, outcome studies have revealed no significant advantage for the pre-emptive use of non-steroidal anti-inflammatory drugs or local anaesthetics, but some tentative positive results have been reported for the pre-emptive use of opioids (for a review see McQuay, 1994). More research is needed using improved designs and larger sample sizes that provide greater statistical power. Phase 3 pain: Whereas nocigenic pain (i.e., Phase 1 and Phase 2 pain) results from peripheral injury, Phase 3 pain, also called neurogenic or neuropathic pain, results from loss of sensory input due to lesions of peripheral nerves or damage to the CNS itself. A characteristic feature of Phase 3 pain states is the lack of correlation between injury and pain. 25 There are a number of Phase 3 pain states and their particular features and underlying mechanisms depend not only on the location and nature of the nerve damage, but also on genetic, developmental , emotional , and psychological influences, as well as the influence of instrumental learning , conditioning, and the development of pain memories. Therefore, even seemingly similar injuries might result in greatly different expressions of pain and disability. There are three main groups of mechanisms that account for the abnormal sensory symptoms that typically accompany most Phase 3 pain states: (1) reactive changes in response to nociceptive afferent input, (2) pathological changes in damaged neurons, and (3) altered response patterns resulting from a functional reorganisation in the CNS following the loss of normal afferent input. Reactive changes in response to nociceptive input are a normal consequence of injury or inflammation. This process also operates in neuropathic pain and is likely to account for the allodynia and secondary hyperalgesia frequently seen in neuropathic pain patients. The pathological changes in damaged neurons and the functional reorganisation in the CNS, however, are largely unique to Phase 3 pain (see e.g. , Devor, 1988). Pathological activity in nociceptors and abnormal ongoing activity in damaged, large­ diameter, non-nociceptive afferents can results in the particularly prolonged and intense, secondary, hyperalgesia-like and allodynia-like changes observed in neuropathic pain patients following loss of afferent input (Willis, 1994). We have already seen that the level of ascending spinal nociceptive transmission is determined by the balance between excitatory and inhibitory systems. Differences in this interaction help to explain why Phase 3 pain is often much harder to attenuate than Phase 2 pain, and why, for example, NMDA receptor antagonists are much more effective in the control of inflammation than in the control of neuropathic pain. Tissue damage and inflammation cause activation of nociceptive afferents that excite dorsal-horn neurons. In neuropathic pain this peripheral nociceptive input either is reduced or absent (e.g., following complete transection of the spinal cord) . Remaining peripheral input and abnormal firing of damaged neurons may aggravate the condition. Additional excitation can arise from descending excitatory tracts that can be activated by stimulation of the motor cortex, stimulation in the reticular activation system, or emotional responses (Willis, 1991 ). 26 Preliminary findings indicate that changes in descending excitation may be involved in the expression of Phase 2 and Phase 3 pain (Cervero & Penderleith, 1985; Cervera & Wolstencroft, 1984). Even more important than the differences in excitatory input are the very significant changes in inhibitory input. Neuropathic damage result in the reduction of a number of key inhibitory controls, including: (i) loss of afferent input resulting in loss of large­ diameter non-nociceptive inhibition; (ii) a reduction in the number of, particularly pre-synaptic, inhibitory opioid receptors; (iii) a reduced ability of morphine to produce analgesia at the spinal level; and (iv) a reduction in the transmitter levels of GABA and glycine, leading to malfunctioning of inhibitory interneurons (Dickenson, 1996a). In contrast, inflammation results in increased activity in inhibitory mechanisms once a certain level of excitation is produced. This might represent physiological attempts to counter the increased peripheral drive. Neurons with low excitability become more excitable, but, within one to three hours of inflammation, neurons with a high degree of peptide or NMDA-receptor-mediated excitation (wind-up) become less active (Stanfa, Sullivan, & Dickenson, 1992). Inflammation increases the ability of morphine to inhibit nociception as it rapidly induces a novel peripheral site of opioid action and enhances the spinal effect of morphine (Dickenson, 1994). Thus, whereas neuropathic damage results in reduced inhibition that helps to maintain a prolonged and constant level of high excitation, inflammation results in a brief period of high excitation followed by an increase in inhibition and a gradual decline in excitation. The effectiveness of morphine analgesia is enhanced after inflammation, yet reduced in neuropathic states. The magnitude of supraspinal descending inhibition over spinal nociceptive neurons is directly related to the amount of afferent input they receive (Brinkhus & Zimmermann, 1983; Cervera & Plenderleith, 1985). Thus, loss of afferent input not only causes changes at the spinal level, but also leads to reduced descending inhibition from supraspinal mechanisms. This disinhibition may be an important factor in the high spontaneous discharges and exaggerated afterdischarges in dorsal-horn neurons recorded in animals with peripheral or central nerve damage (Brinkhus & Zimmermann, 1983). It may also be responsible for the abnormally enhanced pain perception (hyperpathia) in some patients with neuropathy (Cervera & Laird, 1996). 27 1.4.3.2. Changes in neurotransmitter systems The different pain phases not only involve different neurophysiological mechanisms, but also changes in pharmacology as different transmitter systems become involved. Opiates, for example, are extremely effective in inhibiting both Phase 1 and Phase 2 pain states. Although there is no obvious theoretical reason why they should not be effective in treating Phase 3 pain, their usefulness in neuropathic pain is subject to some debate (see e.g. , Arner & Meyerson, 1988; Portenoy, Foley, & lnturrisi , 1990). The opioid inhibitory system appears to be less sensitive following deafferentation than following inflammation (Lombard & Besson, 1989; Mao, Price, & Mayer, 1995; Mayer, Mao, & Price, 1995). Neuropathic pain can be very severe and, for extraneous opioids to be effective, their dosage may need to be so high that side effects become intolerable. The extent of changes in the opiate system following nerve damage may also depend on the time elapsed after injury. Phase 2 pain resulting from tissue damage secondary to inflammation is activated by the increased formation and release of prostaglandins due to the de nova synthesis of the cyclo-oxygenase (COX-2) enzyme (Cervera & Laird , 1996). Cyclo-oxygenase enzymes produce and release prostaglandins, which sensitise nociceptors and stimulate the production of other substances that activate nociceptors such as bradykinins, SP, histamines, and acetylcholine (Stimmel , 1997). This facilitative effect is cumulative and may persist for relatively long periods. The COX-1 enzyme is present in most tissues were it produces prostaglandins as part of normal healthy tissue regulation . The COX-2 enzyme is absent from healthy tissues, but is rapidly induced under conditions of inflammation, resulting in an accumulation of prostaglandins at the injury site (Cervera & Laird, 1996). The effectiveness of non­ steroidal anti-inflammatory drugs (NSAIDs) results from their ability to inhibit the activity of, in particular, the COX-2 enzyme. Although NSAIDs can inhibit Phase-1- type pain, they are more effective in conditions of inflammation. As noted earlier (p. 7), glutamate is a major transmitter in the spinal cord, and the NMDA glutamate receptor subtype is suggested to have an important role in mediating persistent pain and hyperalgesia (Dickenson, 1996a, b ). 28 NMDA receptors are thought to be responsible for the hyperexcitability seen in chronic pain states through their ability to increase the discharge frequency of dorsal­ horn neurons (Stimmel, 1997; Woolf, 1983). Severe and persistent injury results in the repetitive firing of C fibres and a progressivly increasing response of dorsal horn neurons referred to as wind-up. This wind-up phenomenon is dependent on the release of glutamate from C fibres and the consequent opening of postsynaptic ion channels by the NMDA-type glutamate receptor. These long-term changes in the excitability of dorsal horn neurons constitute a memory of the C-fibre input that has an important influence on chronic pain states. NMDA receptor antagonist are not very effective in Phase 1 type pain, but can reverse hyperalgesia evoked by local inflammation and some, but not all, types of abnormal pain behaviours. There is some indication that NMDA receptor antagonists can reduce abnormal neurogenic "wind-up" pain in humans (Kristensen, Svensson, & Gordh, 1992). The development of new NMDA antagonists that are active at the glycine modulatory site of NMDA receptors, thereby greatly reducing the potential for adverse CNS side-effects present in currently available agents, could provide an alternative therapy for Phase 2 and Phase 3 type pain (Cervera & Laird, 1996). Substance P is another intermediary that is important in the transmission and modulation of neural activity in the peripheral and central nervous system. It binds at the NK1 tachykinin receptor, to which it has the highest chemical attraction of all endogenous molecules. Recently discovered selective NK1 -receptor antagonists do not affect responses of dorsal-horn neurons or spinal reflexes to brief noxious stimuli (De Konick & Henry, 1991 ). This suggests that NK1 receptors are not involved in the transmission of Phase-1-type pain. They do, however, appear to play a role in the processing of Phase 2 pain as they inhibit responses in dorsal-horn neurons to prolonged or intense stimulation and enhanced responding evoked by inflammation (De Konick & Henry, 1991 ). Furthermore, peripheral inflammation increases the expression of both SP and NK1 receptors in the spinal cord (Cervera & Laird, 1996). Preliminary evidence from animal studies indicates that NK1-receptor antagonists may be effect