Introduction

In Chapter 5, we discussed the normal responses to a variety of noxious stimuli and their modulation by peripheral and central neural mechanisms. This review showed that noxious stimuli preferentially and most commonly activate a set of interconnected structures, namely the insula and secondary (SII) somatosensory cortices, anterior cingulate gyrus and thalamus. Several additional structures are also activated during normal acute pain although somewhat less frequently: the primary (SI) somatosensory cortex, components of the striatum, the cerebellum, premotor cortex, dorsolateral and orbitofrontal regions of the prefrontal cortex, and the medial midbrain in the region of the periaqueductal gray matter.

In this chapter we review the evidence that chronically painful conditions, whether of peripheral or central origin, may alter the nociceptive processing that normally follows the application of noxious or innocuous stimuli (see Chapter 7). In clinical practice and in the interpretation of the results of pain research, the assumption is often made that the perceptual abnormalities sometimes associated with chronic pain states are attributable only to changes occurring at the peripheral or spinal level. Although this assumption may be correct in most instances, functional imaging studies provide evidence to the contrary in some cases. We cannot assume that, in pathological or chronically painful conditions, information ascending through the spinothalamic tract will be processed by the same mechanisms used for acute pain; this has important clinical implications for the management of chronic pain.

The term “chronic pain” is seldom defined.

Neuropathic pain is pain following a disease or injury to the nervous system, and can be categorized by the location of the causative injury. Chronic pain following injury of the peripheral nervous system, distal to the oligodendroglial cell – Schwann cell junction, can be termed deafferentation pain or peripheral neuropathic pain. Chronic pain “associated with lesions of the CNS” is termed central pain syndrome (Merskey, 1986; Bonica, 1991). There are many situations in which there is injury of both the peripheral and central nervous system, particularly with injuries of the conus medullaris. In this chapter we will consider primate neuropathic pain states, beginning with peripheral neuropathic or deafferentation syndromes, and concluding with central pain syndromes.

In general terms, both central and peripheral chronic pain syndromes have similar characteristics. These include evidence of sensory loss, ongoing pain and pain evoked by stimuli that are not normally painful (allodynia or hyperalgesia). The sensory loss and hypersensitivity are demonstrated by quantitative sensory testing (QST). In addition, a number of primate models have been developed which mimic the sensory abnormalities in patients with neuropathic pain.

Clinical characteristics of peripheral neuropathic pain

The cause of most neuropathies is based on the medical history, supported by laboratory investigations (Casey et al., 1996b). Diabetes is the most common cause of painful neuropathy. Generally, a progressive course suggests an inherited, metabolic or recurrent toxic etiology.

Inputs from nociceptors

The nature of nociceptors

Nociceptors are sensory receptors that respond to stimuli that are damaging or potentially damaging to tissues (Sherrington,1906). The thresholds for activation of many nociceptors can be reached when stimuli of only moderate or non-damaging intensities are applied, but responses continue to increase as stimulus intensity is progressively increased to a level that produces overt damage. By contrast, other nociceptors respond only to intense stimuli and some may not respond at all, even to the strongest mechanical stimuli, unless they are first sensitized (Lynn and Carpenter, 1982; Meyer et al., 1991; Kress et al., 1992; Davis et al., 1993; Treede et al., 1998). The last mentioned have been called “silent nociceptors” (Schaible and Schmidt, 1985, 1988a, 1988b; Schmidt et al., 1995, 2000). Overall, if we include receptors responding to innocuous warming and cooling of the skin, there may be as many as six receptor classes specific for cooling, warming, noxious heat or cold, destructive mechanical or mixed noxious stimuli in humans and other animals.

Types of nociceptors

Nociceptors can be subdivided according to the tissue in which they are found, the size or conduction velocity of the afferent fiber supplying them and the type of stimulus that activates them. Most experimental studies of nociceptors have been performed on common laboratory animals, especially rodents and cats. Some of the most informative, however, have been made during recordings from peripheral nerves of monkeys or human subjects (reviewed in Willis and Coggeshall, 2004).

The clinical descriptions of cordotomy played a major role in elucidating the function and the anatomy of the human spinothalamic tract (STT) (Chapter 1). There are a number of other examples of surgical interventions which have informed our understanding of the pain system. In particular, the pain-related role of the cingulate gyrus is suggested by imaging studies and by the effect of cingulotomy on experimental pain (Rainville et al., 1997; Gildenberg, 2004). Similarly the role of the motor cortex in these systems has suggested the effects of stimulation on activity throughout the pain system (Brown and Barbaro, 2003; Brown, 2004; Peyron et al., 2007). The purpose of this chapter is to examine these surgical interventions in terms of the anatomy and function of structures involved in these interventions. The inclusion of procedures in this chapter is arbitrary and many other such procedures which might have been included have been excluded.

Cordotomy and myelotomy

Percutaneous cordotomy produces relief of pain by interrupting the transmission of signals in the STT from below the level of intervention (Tasker, 1988; Tasker, 2004). The anterolateral quadrant of the spinal cord has long been recognized as the location of the STT (Chapter 1). Recent findings indicate that the dorsal column system also has an important role in visceral nociception (Nauta et al., 1997; Willis et al., 1999). The STT terminates in the primate thalamus, brainstem and other structures such as the hypothalamus and amygdala whereas the dorsal column system terminates in the dorsal column nuclei (Newman et al., 1996).

Introduction

Before the introduction of computerized tomographic (CT) brain imaging, studying human brain mechanisms of pain was largely limited to clinical reports and the post-mortem analysis of brain lesions. Although this approach provided important information and established the background for current investigations, these studies were usually limited by clinical descriptions of each patient's condition. Somatosensory psychophysics seldom included studies of pain and even then it was not possible to relate these observations to brain function or physiology. Because the living brain was invisible (except in the neurosurgery operating suite), research on pain mechanisms focused almost exclusively on the peripheral nervous system.

Brain CT scans introduced the opportunity to apply quantitative sensory testing to the study of living patients with visible, localized brain lesions and to begin to test hypotheses about functional localization and brain mechanisms of pain. The introduction of functional imaging by positron emission tomography (PET) and magnetic resonance imaging (MRI; fMRI) launched a new investigational paradigm into the study of pain mechanisms. Now it is possible to go well beyond the lesion analysis method and to relate human experience, in this case using somatosensory psychophysics, directly to a surrogate measure of activity in groups of neurons at the level of visible, localized brain structure. Since the early 1990s, the number and technical sophistication of functional brain imaging studies, including those related to pain, has increased at a rate that makes it almost impossible to incorporate the results into a conceptual framework.

Introduction

It is well known that much of the sensory input to the central nervous system can be modulated by centrifugally organized control systems that originate in the central nervous system (Head and Holmes, 1911; Hagbarth, 1960). The control mechanisms can be excitatory or inhibitory processes that may occur in the periphery or within the central nervous system. Inhibition can be at pre- and/or postsynaptic sites (Fig. 6.1(I)). Presynaptic inhibition at the first central synapse of a sensory pathway has the potential advantage of being able to reduce sensory input prior to wide dissemination of that sensory input within the central nervous system through the activation of interneuronal networks and multiple ascending pathways, for example, in the spinal cord (Schmidt, 1973; see Chapter 3).

Pre- and postsynaptic inhibition can have somewhat different effects on the stimulus-response curves of second-order sensory neurons, as shown in Fig. 6.1(II). Postsynaptic inhibition involves inhibitory postsynaptic potentials that sum with excitatory postsynaptic potentials (Fig. 6.1(IIA)). If there is a linear summation, the stimulus-response curve will be shifted to the right in a parallel fashion (Carstens et al., 1980). However, if the IPSP is generated in a membrane area near that in which the EPSP is generated, the excitatory current may be shunted and the slope of the stimulus-response curve reduced, causing a reduction in the gain of synaptic transmission (Fig. 6.1(IIB)). A similar reduction in gain can be produced by presynaptic inhibition.

Introduction

On January 19 1911, persuaded by his colleague, the neurologist William Spiller, a Philadelphia surgeon named Edward Martin made a small transverse cut in the spinal cord of a patient suffering from severe pain caused by a tumor affecting the lower end of the spinal column. The cut, made with a thin cataract knife, was no more than 2 mm deep or wide and entered the cord some 3 mm ventral to the entry of a dorsal root in the middle thoracic region. The patient experienced much relief from what had until then been intractable pain (Spiller and Martin,1912). The operation of “chordotomie” or section of the anterolateral tracts of the spinal cord had been introduced in 1910 by Schüller in work on monkeys in which he was exploring the possibility of using the operation for the alleviation of spastic paralysis and tabetic crises in humans. Spiller argued for the procedure on the basis of clinico-pathological observations that appeared to implicate the anterolateral tracts as pathways for conduction of impulses related to pain and temperature through the spinal cord (Müller, 1871; Gowers, 1879; Spiller, 1905; Petrén, 1910). Reports of other successful cases quickly followed (Beer, 1913; Foerster, 1913) and soon, at the hands of Foerster (1913, 1927; Foerster and Gagel, 1932) in Germany and Frazier (1920) in the United States, cordotomy was to become for a time the surgical method of choice in dealing with intractable pain.

Introduction

It is well understood that there are different components to the sensation of pain (Melzack and Casey,1968). The sensory-discriminative aspect of pain refers to the location, intensity and quality of the sensory experience of pain. The affective-motivational aspect of pain refers to the unpleasantness of the pain and how likely it is that it will motivate the animal to escape the pain. We refer to these different components of the pain sensation throughout this review as we examine the possibility that these different components are mediated by different structures in the brain.

The spinothalamic tract (STT) is the spinal tract projecting toward the brain which is most often associated with the sensation of pain (Price and Dubner, 1977; Willis, 1985; Price et al., 2003). Cells of origin of the STT can be divided into those which respond to low-threshold stimuli (LT cells), those which respond to stimuli across the intensive continuum into the noxious range (wide dynamic range, WDR), and those that respond only to noxious stimuli (nociceptive specific, NS). Evidence that any structure mediates the sensory aspect of pain is grouped into four lines: that the structure is connected to other structures known to demonstrate pain-related activity; that neural elements in that structure respond to noxious stimuli; that stimulation of that structure produces pain; and that interventions which interfere with the function of that structure interfere with the sensation of pain evoked by noxious stimuli (Price and Dubner, 1977).

Introduction

As discussed in Chapter 2, several of the sensory pathways that ascend from the spinal cord or brainstem to higher levels of the monkey central nervous system have a nociceptive component and thus may contribute to pain sensation. Spinal cord projections with nociceptive components that ascend to the brain in the anterolateral quadrant of the spinal cord include the spinothalamic, spinoreticular, spinomesencephalic and spinohypothalamic tracts; nociceptive projections that ascend in the dorsolateral or dorsal funiculus are the spinocervical tract and the postsynaptic dorsal column pathway (see Willis and Coggeshall, 2004). Brainstem projections include the trigeminothalamic tract (Price et al., 1976).

To investigate the physiology of an individual spinal cord or brainstem neuron that belongs to one of the ascending nociceptive pathways, it is important to “identify” the neuron by showing that the axon of the individual neuron under investigation actually projects to the appropriate target (Willis and Coggeshall, 2004). Recordings from a neuron unidentified in terms of its projection can be misleading, since many unidentified neurons are likely to be interneurons, and these could be excitatory or inhibitory and might or might not influence the activity of sensory projection neurons. For instance, many spinal cord interneurons belong to neural circuits that function to control motor output (Jankowska et al., 1981; Rudomin et al., 1987).

Identification of a projection neuron is typically accomplished by demonstrating that the neuron can be activated antidromically in response to electrical stimulation in a region in which the axon of that projection neuron synapses (Trevino et al., 1973; Bryan et al., 1974; Haber et al., 1982; see Willis and Coggeshall, 2004).