Content uploaded by Robert Gatchel
Author content
All content in this area was uploaded by Robert Gatchel
Content may be subject to copyright.
The Biopsychosocial Approach to Chronic Pain: Scientific Advances and
Future Directions
Robert J. Gatchel and Yuan Bo Peng
The University of Texas at Arlington Madelon L. Peters
Maastricht University
Perry N. Fuchs
The University of Texas at Arlington Dennis C. Turk
University of Washington, Seattle
The prevalence and cost of chronic pain is a major physical and mental health care problem in the United
States today. As a result, there has been a recent explosion of research on chronic pain, with significant
advances in better understanding its etiology, assessment, and treatment. The purpose of the present
article is to provide a review of the most noteworthy developments in the field. The biopsychosocial
model is now widely accepted as the most heuristic approach to chronic pain. With this model in mind,
a review of the basic neuroscience processes of pain (the bio part of biopsychosocial), as well as the
psychosocial factors, is presented. This spans research on how psychological and social factors can
interact with brain processes to influence health and illness as well as on the development of new
technologies, such as brain imaging, that provide new insights into brain-pain mechanisms.
Keywords: biopsychosocial, chronic pain, neuroscience of pain, pain and cognition, pain and emotion
During the past decade, there has been an explosion of research
on chronic pain, with significant advances in understanding its
etiology, assessment, and treatment (Gatchel, 2004a, 2004b, 2005;
Turk & Monarch, 2002). This research has important health care
implications. Epidemiological research has shown that chronic
pain (loosely defined as prolonged and persistant pain of at least 3
months in duration) and chronic recurrent pain (recurrent episodes
of pain interspersed with pain-free periods extending over months
or years) affects 10%–20% of adults in the general population
(Blyth et al., 2001; Gureje, Von Korff, Simon, & Gater, 1998;
Verhaak, Kerssens, Dekker, Sorbi, & Bensing, 1998). For exam-
ple, in a large-scale epidemiological study, Von Korff et al. (2005)
estimated a 19% prevalence for chronic spinal pain (neck and
back) in the United States in the previous year and a 29% lifetime
rate. The American Academy of Pain Management (2003) asserted
that approximately 57% of all adult Americans reported experi-
encing recurrent or chronic pain in the past year. About 62% of
those individuals reported being in pain for more than 1 year, and
40% noted that they were constantly in pain. Indeed, as Gatchel
(2004a, 2004b) has summarized, pain is a pervasive medical
problem: It affects over 50 million Americans and costs more than
$70 billion annually in health care costs and lost productivity; it
accounts for more than 80% of all physician visits. Moreover,
chronic pain is often associated with major comorbid psychiatric
disorders and emotional suffering.
As the above factors attest, the prevalence and cost of chronic
pain is a major physical and mental health care problem in the
United States. Moreover, individuals 50 years of age and older are
twice as likely to have been diagnosed with chronic pain (Gatchel,
2004, 2005). Currently, there are approximately 35 million Amer-
icans aged 65 years or older, accounting for 12.4% of the total
population. The proportion of the population aged 65 and over is
expected to increase by 57% by the year 2030, with Americans
now having an average life expectancy of 77 years (Social Security
Administration, n.d.). Awareness of these population trends has
contributed to an increased concern about health care issues of
older Americans, including chronic pain problems. With these
estimates in mind, it is not surprising that the U.S. Congress
designated 2001–2010 as the Decade of Pain Control and Research
and that the Joint Commission on Accreditation of Healthcare
Organizations now requires physicians to consider pain as the fifth
vital sign (added to the other vital signs of pulse, blood pressure,
core temperature, and respiration).
The statistics cited above and population trends have fueled
a great deal of research on chronic pain. The purpose of the
present article is to provide a review of some of the most
noteworthy scientific advances in this area. As is initially
discussed, the biopsychosocial model has proved to be the most
widely accepted and most heuristic perspective to the under-
standing and treatment of chronic pain. Subsequently, reviews
Robert J. Gatchel and Yuan Bo Peng, Department of Psychology, The
University of Texas at Arlington; Madelon L. Peters, Department of
Medical, Clinical and Experimental Psychology, Maastricht University,
The Netherlands; Perry N. Fuchs, Department of Psychology and Depart-
ment of Biology, The University of Texas at Arlington; Dennis C. Turk,
Department of Anesthesiology, University of Washington, Seattle.
Preparation of this article was supported by National Institutes of Health
Grants 1K05-MH071892, 2R01 DE 010713, and 2R01 MH 046452; De-
partment of Defense Grant DAMD17-03-1-0055; and National Institute of
Arthritis and Musculoskeletal and Skin Diseases Grants AR44724 and
AR47298.
Correspondence concerning this article should be addressed to Robert J.
Gatchel, Department of Psychology, University of Texas at Arlington, Box
19528, Arlington, TX 76063. E-mail: gatchel@uta.edu
Psychological Bulletin Copyright 2007 by the American Psychological Association
2007, Vol. 133, No. 4, 581–624 0033-2909/07/$12.00 DOI: 10.1037/0033-2909.133.4.581
581
of important biobehavioral mechanisms and psychosocial fac-
tors are provided.
The Biopsychosocial Model of Chronic Pain
The traditional approach embraced a dualistic viewpoint that
conceptualized the mind and body as functioning separately and
independently. The inadequacy of the dualistic model contributed
to a growing recognition that psychosocial factors, such as emo-
tional stress, could impact the reporting of symptoms, medical
disorders, and response to treatment. George Engel (1977) is
credited as one of the first to call for the need of a new approach
to the traditional biomedical reductionistic philosophy that domi-
nated the field of medicine since the Renaissance. This subse-
quently led to the growth of the field of behavioral medicine and
health psychology (Gatchel & Baum, 1983). A major outgrowth, in
turn, was the development and evolution of the biopsychosocial
model. This model has been especially influential in the area of
chronic pain.
The biopsychosocial model focuses on both disease and illness,
with illness being viewed as the complex interaction of biological,
psychological, and social factors (Gatchel, 2005). As succinctly
summarized by several authors (e.g., Gatchel, 2004a, 2004b; Turk
& Monarch, 2002), disease is defined as an objective biological
event involving the disruption of specific body structures or organ
systems caused by either anatomical, pathological, or physiologi-
cal changes. In contrast, illness refers to a subjective experience or
self-attribution that a disease is present. Thus, illness refers to how
a sick person and members of his or her family live with, and
respond to, symptoms of disability.
The distinction between disease and illness is analogous to the
distinction that can be made between nociception and pain. Noci-
ception involves the stimulation of nerves that convey information
about potential tissue damage to the brain. In contrast, pain is the
subjective perception that results from the transduction, transmis-
sion, and modulation of sensory information. This input may be
filtered through an individual’s genetic composition, prior learning
history, current psychological status, and sociocultural influences.
For pain to be registered, the organism must be conscious. To the
best of our knowledge, completely anesthetized patients do not
perceive pain; however, nociception can be detected following a
surgical incision even in the absence of any subjective report.
Loeser (1982) originally formulated a general model that delin-
eated four dimensions associated with the concept of pain: the
dimensions of nociception and pain reviewed above, suffering (the
emotional responses that are triggered by nociception or some
other aversive event associated with it, such as fear or depression),
and pain behavior (those things that people say or do when they
are suffering or in pain, such as avoiding activities or exercise for
fear of reinjury). Pain behaviors are overt communications of pain,
distress, and suffering.
Waddell (1987) has emphasized that pain cannot be comprehen-
sively evaluated without an understanding of the individual who is
exposed to the nociception. Waddell also made a comparison
between Loeser’s (1982) model of pain and the biopsychosocial
model put forth by Engel (1977). In particular, Engel proposed the
important dimensions of the physical problem, distress, illness
behavior, and the sick role, which corresponded to Loeser’s di-
mensions of nociception, pain, suffering, and pain behavior, re-
spectively. In order to fully understand a person’s perception and
response to pain and illness, the interrelationships among biolog-
ical changes, psychological status, and the sociocultural context all
need to be considered (see Figure 1). Any model that focuses on
only one of these dimensions will be incomplete and inadequate.
Many of the individual dimensions depicted in Figure 1, and the
complexities involved with their interactions, are discussed in
subsequent sections of this article, particularly the neurobiology of
the nociception process and other basic neuroscience processes of
pain (the bio part of biopsychosocial), as well as psychological and
social factors. The psychosocial factors involve both emotion and
cognition. Emotion is the more immediate reaction to nociception
and is more midbrain based. Cognitions then attach meaning to the
emotional experience and can then trigger additional emotional
reactions and thereby amplify the experience of pain, thus perpet-
uating a vicious circle of nociception, pain, distress, and disability.
We then review the implications of the new insights for better
understanding the etiology, assessment, treatment, and prevention
of chronic disability.
The Nociceptive Process
Early Biomedical Models
Historically, 19th and 20th century models of nociceptive pro-
cessing followed the traditional biomedical model of disease. The
ideas followed a Cartesian view that there was an isomorphic
relationship between pain and tissue injury. The early biomedical
models can, in general, be divided into two general perspectives.
One perspective, “specificity theory,” generally stated that there
were unique receptor mechanisms and pathways that transduced
and transmitted specific painful information from the periphery to
the spinal cord and then to the brain. This direct transmission line
model can be traced back to views expressed by the ancient
Greeks. One of the earliest and best known of the modern speci-
ficity theorists was von Frey (see Finger, 1994). His work revolved
around the identification and description of mechanical and ther-
mal receptive fields on the skin. On the basis of his work, it was
suggested that specialized nerve endings were involved in the
transduction and transmission of painful information.
Another general theoretical perspective has been referred to as
the “pattern response” (Nafe, 1934; Sinclair, 1955; Weddell,
1955). According to this perspective, nociceptive information was
not primarily due to activation of specific receptors and pathways
but rather was due to the pattern of responses in afferent systems.
It was the stimulus intensity and the processing of the pattern of
responses that determined the perceptual response to the nocicep-
tive input, namely, pain. Although these two general perspectives
explained much of the literature and prompted a wealth of scien-
tific literature, both perspectives had limitations, and many issues
and potential explanations related to pain and suffering remained
elusive.
Another perspective, harkening back to Aristotle and challeng-
ing the pure sensory models described, conceptualized pain as a
“quality of the soul”—an emotion in contrast to a pure sensory
event. This competing viewpoint was carried forward to more
recent times. For instance, Livingston (1943, 1998) was one of the
first to expose the weaknesses of specificity theory and argue for
pain as a subjective state that arises from activation of aversive
582 GATCHEL, PENG, PETERS, FUCHS, AND TURK
networks in the brain. His concept of “appetites,” with pleasure
and pain as the motivating attributes, was a dramatic shift in
thought and reflected Hebb’s (1949) belief that pain was a factor
motivating behavior. The failure of these unidimensional sensory
and affective models to explain much of what was observed
experimentally and clinically (Beecher, 1959) and the inadequacy
of treatments based on these models served as the impetus for a
more complex, integrative model. In particular, the seminal gate
control theory of pain postulated by Melzack and colleagues
(Melzack & Casey, 1968; Melzack & Wall, 1996).
The Gate Control Theory of Pain
The initial framework for the gate control theory of pain, which
built on the ideas of the Dutch surgeon Nordenbras (1959), was
developed to ensure that the known properties of clinical pain
conditions at the time were explained. Melzack and Wall (1965)
sought to combine the properties of the specificity theories with
the best features of the pattern response theories and the affective-
motivational view in order to generate the more inclusive gate
control theory of pain. They recognized that there was a certain
degree of specificity for peripheral nerve function. They also
realized that there was a certain degree of pattern recognition that
was responsible for the underlying peripheral and central process-
ing of noxious information. Moreover, they acknowledged that a
comprehensive model must take into consideration the amplifiying
effects of emotion and the interpretive role of cognitive evaluation.
As outlined by Melzack and Wall (1996), the gate control theory
of pain had to account for a number of facts such as the following:
(1) the variable relationship between injury and pain; (2) non-noxious
stimuli can sometimes produce pain; (3) the location of pain and tissue
damage is sometimes different; (4) pain can persist long after tissue
healing; (5) the nature of the pain and sometimes the location can
change over time; (6) pain is a multi-dimensional experience; and (7)
there is a lack of adequate pain treatments. (p. 165)
It is precisely these facts that no theory at the time could explain.
The initial formulation proposed that there are five stages that
compose the mechanism by which noxious signals enter the spinal
cord from the periphery and then proceed to higher level brain
areas. The first stage consisted of the small diameter peripheral
nerve fiber transmission of signals to cells in the spinal cord. The
second stage included facilitatory interneurons in the region of the
spinal cord to account for the fact that cells in the spinal cord can
show prolonged afterdischarge following the arrival of a signal
from the peripheral nerve (Wall, 1960). The afterdischarge was
accounted for by an excitatory interneuron. The third stage incor-
porated a large-fiber, low-threshold input. This third stage focused
attention on a group of additional peripheral fiber inputs to the
spinal cord that could be involved in pain processing. As they
indicated, most research prior to the gate control theory focused on
nociceptive specific neurons, or those cells that responded only to
high-threshold peripheral stimulation. The fourth stage included
inhibitory interneurons to account for the fact that postsynaptic
Figure 1. A conceptual model of the biopsychosocial interactive processes involved in health and illness. From
“Comorbidity of Chronic Mental and Physical Health Conditions: The Biopsychosocial Perspective,” by R. J.
Gatchel, American Psychologist, 59, 792–805. Copyright 2004 by the American Psychological Association.
583
BIOPSYCHOSOCIAL APPROACH TO CHRONIC PAIN
inhibition was likely to occur in the spinal cord. The fifth stage was
the inclusion of a descending modulatory system to account for the
finding that there was an inhibitory influence from the brainstem to
the spinal cord (Wall, 1967). The final stage was the inclusion of
a loop system, with the assumption that ascending signals to the
brain engage and influence descending modulatory systems.
Therefore, Figure 2 illustrates the final diagram of the gate control
theory. There is little doubt that the gate control theory, with a
focus on the multidimensional and variable relationship between
pain and tissue damage, was a major advancement in the field of
pain research and management, prompting much research and
ultimately increasing researchers’ understanding of pain mecha-
nisms. As the field of pain research and management evolves, the
adequacy of the gate control theory of pain to explain different
factors has also continued to be examined.
The Neuromatrix Theory of Pain
The neuromatrix theory of pain proposes that pain is a multi-
faceted experience that is produced by a characteristic neurosig-
nature of a widely distributed brain neural network, called the
body–self neuromatrix (Melzack, 2001, 2005; see Figure 3). The
body–self neuromatrix integrates cognitive–evaluative, sensory–
discriminative, and motivational–affective components proposed
by Melzack and Casey (1968). The theory proposes that the output
patterns of the neuromatrix engage perceptual, behavioral, and
homeostatic systems in response to injury and chronic stress. A
critical component of the neuromatrix theory of pain is the recog-
nition that pain is the consequence of the output of the widely
distributed brain neural network rather than a direct response to
sensory input following tissue injury, inflammation, and other
pathologies (Melzack, 2001).
The development of such a hypothetical system stems primarily
from reports and research in patients with spinal cord injuries and
in patients that experience phantom limb and phantom limb pain.
In a large number of cases, paraplegics will continue to experience
body sensations and pain below the level of the spinal section. A
significant proportion of individuals who lose a limb or sensation
in other body regions will continue to experience the presence of
the limb or the otherwise anesthetic area (e.g., below a spinal
lesion). Although the experience of the phantom limb might be in
some cases maintained by altered peripheral nerve activity in the
region of the stump (Hunter, Katz, & Davis, 2005), there is
sufficient evidence indicating that peripheral mechanisms do not
fully account for such phenomena (Katz & Melzack, 1990; Ram-
achandran, 1998; Ramachandran & Hirstein, 1998; Wu et al.,
2002). Traditional specificity and pattern theories of pain, in
particular, have difficultly accounting for these phenomena. The
body–self neuromatrix, however, requires no actual sensory input
to produce experiences of the body.
Homeostasis, Allostatic Load, and Hypothalamic-
Pituitary-Adrenal (HPA) Axis Dysregulation
It is becoming clear that the pain experience is determined by a
multitude of factors. Although the focus has historically been
directed at sensory mechanisms, more attention is being placed on
factors related to cognitive, affective, behavioral, and homeostatic
factors. The primary basis for including discussions of homeostatic
factors is that chronic pain threatens the organism and produces a
cascade of events that eventually contributes to the maintenance of
such conditions even after the original tissue damage has been
resolved or in the absence of any objectively determined pathol-
ogy. If one views pain as a primary threat to the organism, similar
to the stress of extreme hunger and thirst, then mechanisms should
be present to engage and motivate the organism to restore basic
homeostatic function (LaGraize, Borzan, Rinker, Kopp, & Fuchs,
2004). The major consequence of homeostatic imbalance is stress.
Regardless of the source, stressors activate numerous systems such
as the autonomic nervous system and the HPA axis. Prolonged
activation of the stress system has disastrous effects on the body
(cf. Selye, 1950; Korte, Koolhaas, Wingfield, & McEwen, 2005)
and sets up a condition of a feedback loop between pain and stress
reactivity.
During periods of short-term stress and homeostatic imbalance,
the hypothalamus activates the pituitary gland to secrete adreno-
corticotropic hormone, which acts on the adrenal cortex to secrete
cortisol. Secretion of cortisol elevates blood sugar levels and
enhances metabolism, an adaptive response that allows the organ-
ism to mobilize energy resources to deal with the threat and restore
homeostatic balance (i.e., fight or flight response). The situation is
much more serious during prolonged periods of stress and homeo-
static imbalance that is associated with long-term psychological
stress, including chronic pain and other pathological conditions.
Prolonged, elevated levels of cortisol are related to the exhaustion
phase of Selye’s (1950) General Adaptation Syndrome. The neg-
ative effects of this stage of the adaptation syndrome include
atrophy of muscle tissue, impairment of growth and tissue repair,
immune system suppression, and morphological alterations of
brain structures, which together might set up conditions for the
development and maintenance of a variety of chronic pain condi-
tions (Chrousos & Gold, 1992; McBeth et al., 2005; McEwen,
2001; McLean et al., 2005). The concept of allostatic load, and the
factors that contribute to physiological burden, is becoming in-
creasingly recognized as an important component of disease and
disabilities (Seng, Graham-Bermann, Clark, McCarthy, & Ronis,
Figure 2. L⫽large diameter fibers; S ⫽small diameter fibers; SG ⫽
substantia gelatinosa; T ⫽first central transmission cells; ⫹⫽excitation;
⫺⫽inhibition. Melzack and Wall’s gate control theory of pain. From
“Pain Mechanisms: A New Theory,” by R. Melzack and P. D. Wall, 1965,
Science, 150, p. 975. Copyright 1965 by the American Association for the
Advancement of Science. Reprinted with permission.
584 GATCHEL, PENG, PETERS, FUCHS, AND TURK
2005; Singer, Friedman, Seeman, Fava, & Ryff, 2005; Tucker,
2005).
According to Melzack (2005), psychological stress, as well as
sensory and cognitive events, modulates the neurosignature of the
body–self neuromatrix which, as a consequence of altered neu-
romatrix output, is associated with chronic pain conditions. The
concept of the neuromatrix has potentially important explanatory
implications for brain function in general and, together with the
concepts of allostasis and homeostasis, provides a theoretical
framework for the biopsychosocial perspective of chronic pain. As
is discussed later, there is a growing literature demonstrating the
importance of psychosocial factors (emotion and cognition) in this
neuromatrix conceptualization.
The Neuroscience of Pain
The field of neuroscience has contributed to a better delineation
of basic mechanisms in pain processing by conducting carefully
controlled experimental studies. In doing do, several experimental
pain models have been proposed involving inflammatory pain,
neuropathic pain, and cancer pain. These models, in turn, have led
to important clinical applications, such as the development of
analgesic agents for improved management of chronic pain. A
summary of research related to these three models is presented in
Table 1. The purpose of this section is to simply provide a general
overview of the wide breadth of neuroscience research of chronic
pain that is ongoing. Research on inflammatory, neuropathic, and
cancer pain models is reviewed.
Inflammatory mediators and their action on pain pathways have
a direct effect on pain states through stimulation or potentiation of
nociceptive transduction at peripheral terminals and central
changes contributing to hypersensitivity (Levine & Reichling,
1999; Raja, Meyer, Ringkamp, & Campbell, 1999). As noted in
Table 1, inflammatory pain models have been tested with a number
of different techniques, such as experimental arthritis by intra-
articular injections of certain substances, cutaneous inflammation
by application of certain extracts, and so forth. In addition to the
somatic pain models, several visceral pain models have been
developed, including writhing test by intraperitoneal injection of
phenylquinone or acetic acid or injection of formalin into the colon
wall. Other methods used have been intracolonic application of
mustard oil or capsaicin, colorectal distention, and intrabladder
injection of xylene, to mention some of the more commonly used
ones.
Neuropathic pain results from damage to the nervous system,
including peripheral nerves, spinal cord, and certain central ner-
vous system (CNS) regions. As seen with many other clinical pain
conditions, the clinical symptoms of neuropathic pain include
spontaneous pain, allodynia (i.e., pain due to a stimulus that does
not normally produce pain, such as soft touch), and hyperalgesia
(i.e., an exaggerated response to a stimulus that is normally some-
what painful). It may spread to the neighboring cutaneous distri-
bution of the injured nerve or develop bilaterally in mirror image
sites with the quality of burning, shooting, stabbing, piercing, and
electric shock.
Following trauma, inflammation or infection causes almost half
of human neuropathies. Sciatic inflammatory neuritis models have
been developed to address this issue by injection of zymosan
around the sciatic nerve. As a further model for two of the major
human diseases that cause peripheral neuropathy, there has been
the development of the postherpetic neuralgia model, involving
reactivation of a primary infection with varicella-zoster virus, and
the diabetic neuropathic pain model, involving injection of strep-
tozocin or use of animal strains (such as insulin-deficient rats and
mice, insulin resistant mice, and Mongolian gerbil).
Cancer pain is an increasingly devastating problem affecting the
quality of life for patients undergoing active treatment and ad-
vanced cancer stages. Cancer-related pain can be caused directly
by tumor infiltration or compression of peripheral nerve, plexus, or
Figure 3. Melzack’s body-self neuromatrix model of pain. From “Pain and the Neuromatrix in the Brain,” by
R. Melzack, 2001, Journal of Dental Education, 65, p. 1382. Copyright 2001 by the American Dental Education
Association. Reprinted with permission.
585
BIOPSYCHOSOCIAL APPROACH TO CHRONIC PAIN
Table 1
Summary of Experimental Pain Models
Category Models Procedures References
Inflammatory
pain
Experimental arthritis Intra-articular injection of kaolin and
carrageenan
Schaible & Schmidt (1985); Schaible & Schmidt (1988a, 1988b);
Neugebauer & Schaible (1990); Schaible & Grubb (1993)
Intra-articular injection of Freund’s
adjuvant
Butler et al. (1992); Grubb et al. (1993); L. F. Donaldson et al. (1993)
Cutaneous
inflammation
Topical application of mustard oil Woolf & King (1990); Koltzenburg et al. (1992); Koltzenburg &
Handwerker (1994)
Injection of carrageenan Traub (1996); Ren, Williams, et al. (1992); Meller et al. (1994)
Injection of complete Freund’s
adjuvant
Ren, Hylden, et al. (1992); Ren & Dubner (1993); Ren et al. (1994);
Ren & Dubner (1996); Ruda et al. (2000)
Injection of formalin Dubuisson & Dennis (1977); Dickenson & Sullivan (1987); Coderre et
al. (1990); Coderre & Melzack (1992a, 1992b); Coderre et al. (1993);
Abbott et al. (1995)
Injection of bee venom Lariviere & Melzack (1996); J. Chen et al. (1998); J. Chen et al. (1999);
Lariviere & Melzack (2000)
Injection or topical application of
capsaicin
Bodnar et al. (1983); Simone et al. (1989); LaMotte et al. (1991, 1992);
Torebjo¨rk et al. (1992); Sluka et al. (1997)
Myofascial pain Injection of acidic saline Sluka et al. (2001); Hoeger-Bement & Sluka (2003); Radhakrishnan et
al. (2003); Skyba, et al. (2005)
Injection of carrageenan Radhakrishnan et al. (2003); Skyba et al. (2005)
Visceral pain Intraperitoneal injection of
phenylquinone
Hendershot & Forsaith (1959); Taber et al. (1964)
Intraperitoneal injection of acetic
acid
Taber et al. (1969)
Injection of formalin into the colon
wall
Miampamba et al. (1994)
Intracolonic application of mustard
oil or capsaicin
Laird, Martinez-Caro, et al. (2001)
Colorectal distention Ness et al. (1991); Al-Chaer et al. (1996)
Intrabladder injection of xylene Abelli et al. (1988, 1989)
Intrauterine injection of mustard oil Wesselmann et al. (1998)
Vaginal hyperalgesia through
endometriosis
Berkley et al. (2001)
Artificial ureter stone Giamberardino et al. (1995)
Neuropathic
pain
Deafferentation Spared root paradigm Liu & Chambers (1958); Loeser & Ward (1967); Goldberger & Murray
(1974); Basbaum & Wall (1976); Wall et al. (1979); Pubols &
Goldberger (1980); Brinkhus & Zimmerman (1983)
Nerve transaction Devor & Wall (1981a, 1981b); Woolf & Wall (1982); Hylden et al.
(1987)
Peripheral
neuropathic pain
(Kim et al., 1997)
Bennett model Mosconi & Kruger (1996); Bennett & Xie (1988); Maves et al. (1993)
Seltzer model Seltzer et al. (1990)
Chung model S. H. Kim & Chung (1992); Palecek et al. (1992); Carlton et al.
(1994); Yoon et al. (1996); Ali et al. (1999)
Sciatic cryoneurolysis Freezing of the sciatic nerve DeLeo et al. (1994); Imamura & Bennett (1995); Willenbring et al.
(1995)
Sciatic demyelination Topical application of lysolecithin Wallace et al. (2003)
Central neuropathic
pain (Boivie et al.,
1989; Leijon et al.,
1989; Boivie,
1990; Willis,
2002)
Cordotomy White et al. (1950); Levitt & Levitt (1981); Lenz et al. (1987, 1989);
Vierck et al. (1990); Ovelmen-Levitt et al. (1995); Vierck & Light
(1999); Weng et al. (2000)
Contusion Siddall et al. (1995); Basso et al. (1995); Hulsebosch et al. (2000)
Spinal cord hemisection Christensen et al. (1996); Christensen & Hulsebosch (1997a, 1997b)
Injection of quisqualic acid Yezierski & Park (1993); Yezierski et al. (1993, 1998)
Injection of kainate LaBuda et al. (2000)
Ischemia Hao, Xu, Yu, et al. (1991); Hao, Xu, Aldskogius, et al. (1991); Hao,
Xu, Yu, et al. (1992); Xu et al. (1992); Hao, Xu, Aldskogius, et al.
(1992)
Sciatic Injection of zymosan Gazda et al. (2001); Chacur et al. (2001)
inflammatory neuritis around the sciatic nerve
Placing proinflammatory gut suture Maves et al. (1993)
Placing dead bacteria or carrageenan Eliav et al. (1999)
Postherpetic
neuralgia model
Infection with varicella-zoster virus Sadzot-Delvaux et al. (1990, 1995); Fleetwood-Walker et al. (1999)
586 GATCHEL, PENG, PETERS, FUCHS, AND TURK
roots; indirectly by immunoreactive and pronociceptive substances
released from tumors; or by treatment (chemotherapy, radiation, or
surgery). In order to model human cancer pain, several cancer-
related pain models have been developed, including the
chemotherapy-induced peripheral neuropathy model by injection
of vincristine, taxol, and cisplatin; the cancer invasion pain model
by implantation of Meth A sarcoma cells around the sciatic nerve;
and the bone cancer pain model by injection of osteolytic mouse
sarcoma NCTC2472 cells into the femur bone marrow or by
injection of MRMT-1 rat mammary gland carcinoma cells into the
tibia bone marrow of rats.
With the establishment of these pain models, the biological
mechanisms of pain can be further studied by application of
various techniques. For example, following the L5 spinal nerve
ligation, in combination with behavioral observation, electrophys-
iological techniques can be applied to study the peripheral single
fiber response properties and change of excitability of central
dorsal horn neurons; genetic expression of target proteins (such as
different voltage sensitive sodium channels, opioid receptors, early
response genes) can be evaluated during the acute phase and
chronic phase of the neuropathic pain. In addition to anatomical,
behavioral, psychophysical, and computational neural modeling
tools, genetic, electrophysiological, and imaging techniques can be
widely used in the investigation of these three models of pain.
They are discussed next.
Genetics
With the rapid advances in molecular biology and genetics, the
human genome was mapped out in 2001 (Jasny & Kennedy, 2001).
Biological functions of every system, organ, and each individual
cell depend on genetic expression to produce peptides or proteins,
which either contribute to the structure of the cell or participate in
metabolism through various enzymes. Overexpression or elimina-
tion of a gene results in functional changes. The neuronal activities
involved in pain transmission can be influenced by activities of
immediate, early genes as well as transcriptional factors, all of
which may result in changes in gene expression. With the under-
standing of the gene expression in response to noxious stimuli,
genetic engineering can be applied in experimental or potential
clinical conditions. Examples of this are “knock-out” mice or
antisense oligonucleotides and viral transfection of neurons locally
at various levels along the ascending or descending pathways of
noxious signal transmission (Mogil, Yu, & Basbaum, 2000).
A recent new technique, ribonucleic acid interference (RNAi),
has been introduced in studying the effect of delta opioid receptor
in the spinal cord and dorsal root ganglion (Luo et al., 2005).
Double-stranded, short-interfering RNAs (siRNAs) of 21–22 nu-
cleotides in length initiate a sequence-specific, post-trancriptional
gene silencing in animals and plants known as RNAi. siRNA has
been found to selectively silence the delta opioid receptor, but not
mu opioid receptors. The antinocicetive effects of the correspond-
ing agonists are dose-dependently and reversibly blocked (Luo et
al., 2005). A brief summary of genetic manipulation is listed in
Table 2.
In contrast to an increased sensitivity along the somatosensory
system to nociceptive signal transmission, a rare opposite condition,
congenital insensitivity to pain (usually associated with anhidrosis),
has been reported in the literature (see Table 2). As opposed to
increased pain sensitivity, it is characterized by recurrent episodic
fevers, anhidrosis (inability to sweat), absence of reaction to noxious
(or painful) stimuli, self-mutilating behavior and mental retardation. It
is explained by several genetic mechanisms:
1. Developmentally, a combined defect in sensory and au-
Table 1 (continued)
Category Models Procedures References
Diabetic neuropathic
pain model
Injection of streptozocin Wuarin-Bierman et al. (1987); Courteix et al. (1993, 1994)
Insulin deficient BB rat NOD mice Sima (1980)
Mosseri et al. (2000); Pieper et al. (2000)
Insulin resistant ob/ob and db/db
mice
Meyerovitch et al. (1991); Takeshita & Yamaguchi (1998)
Mongolian gerbil Vincent et al. (1979); Shafrir et al. (2001)
Cancer pain Chemotherapy-
induced peripheral
neuropathy model
Injection of vincristine Aley et al. (1996); Nozaki-Taguchi et al. (2001); Tanner et al. (1998)
Injection of taxol Apfel et al. (1991); Cavaletti et al. (1995); Cliffer et al. (1998); Boyle et
al. (1999); Authier et al. (2000); Dina et al. (2001); Polomano et al.
(2001)
Injection of cisplatin de Koning et al. (1987a, 1987b); Verdu et al. (1999); ter Laak et al.
(2000)
Cancer invasion pain
model
Implantation of Meth A sarcoma
cells around the sciatic nerve in
BALB/c mice
Shimoyama et al. (2002, 2005)
Bone cancer pain
model
Injection of osteolytic mouse
sarcoma NCTC2472 cells into the
femur bone marrow
Schwei et al. (1999); Honore et al. (2000); Mantyh et al. (2002); Luger
et al. (2005)
Injection of MRMT-1 rat mammary
gland carcinoma cells into the tibia
bone marrow of Sprague–Dawley
rats
Medhurst et al. (2002); Walker et al. (2002)
587
BIOPSYCHOSOCIAL APPROACH TO CHRONIC PAIN
Table 2
Brief Summary of Genetic Modulation of Pain
Category Causes Symptoms and Signs References
Congenital insensitivity
to pain
A combined defect in sensory
and autonomic neurons
derived from the neural
crest
A reduced evoked potential
Lack of pain experience following
electrical shock;
Self-mutilation and fractures
Chatrian et al. (1975); Shorey & Lobo (1990)
Manfredi et al. (1981)
Itoh et al. (1986); Chatrian et al. (1975);
Matsuo et al. (1981); Sweet (1981);
Derwin et al. (1994); Nolano et al. (2000);
Schulman et al. (2001)
Lack of flare response to histamine
injection
Manfredi et al. (1981); Nolano et al. (2000)
Lack of temperature regulation Itoh et al., (1986); Matsuo et al. (1981); Vital
et al. (1998); Sztriha et al. (2001)
Overexpression of
endogenous opioids
Dehen et al. (1977, 1978)
Reduced number of primary
afferent nociceptors
Larner et al. (1994)
Loss of neurons in
sympathetic ganglia
Dyck et al. (1983); Derwin et al. (1994);
Shorer et al. (2001); Sztriha et al. (2001)
Loss of trkA function
(receptor for nerve growth
factor) as the result of
mutations of the trkA
receptor gene
Indo et al. (1996); Mardy et al. (1999);
Yotsumoto et al. (1999); Shatzky et al.
(2000); Toscano et al. (2000); Greco et al.
(2000); Miura, Hiura, et al. (2000); Miura,
Mardy, et al. (2000); Indo (2001); Toscano
& Andria (2001); Bodzioch et al. (2001);
Miranda et al. (2002); Barone et al. (2005)
Rat strain differences Carrageenan to induce
inflammatory pain
Inbred Lewis (LEW), Fischer 344
(FIS), and outbred Sprague-
Dawley (SD) rat strains differ in
their pain sensitivity to
mechanical and thermal stimuli
Fecho et al. (2005)
Modulation at sensory
receptor
Mechanoreceptor BNC1 (a non-voltage-dependent
sodium channel) DRASIC
M. P. Price et al. (1996); Drummond et al.
(2000); M. P. Price et al. (2000);
Drummond et al. (2001); Welsh et al.
(2002)
DRASIC M. P. Price et al. (2001)
TRPV1-deficient mice Essential for selective modalities of
pain sensation and for thermal
hyperalgesia
Numazaki & Tominaga (2004)
Modulation at
membrane receptor
Null mutants for nerve
growth factor
Loss of primary afferent and
sympathetic neurons
Crowley et al. (1994)
Deletion of neurokinin-1
receptors
Reduction in response to
intradermal injection of
capsaicin
Laird, Roza et al. (2001)
Reduction in response to second
phase of formalin test
De Felipe et al. (1998)
Deletion of the CGRP gene Fail to develop secondary heat
hyperalgesia by kaolin and
carrageenan
Zhang et al. (2001)
Deletion of mu opioid
receptor gene
Oprm1 gene Matthes et al. (1996); Sora et al. (1997);
Lotsch & Geisslinger (2005)
Genetic modulation at
intracellular
molecules
Deletion of the R1subunit
of protein kinase A (PKA)
Reduction of allodynia by tissue
damage, a reduction of the
responses to the second phase of
formalin test, and central
sensitization caused by
intrathecal injection of PGE2
Malmberg, Brandon, et al. (1997)
Deletion of the gamma
isoform of protein kinase C
(PKCg)
Fail to develop neuropathic pain
after partial sciatic nerve injury,
but show normal responses to
acute noxious stimuli
Malmberg, Chen, et al. (1997)
Mitogen-activated protein
kinase (MAPK)
Regulation of central sensitization Ji & Woolf (2001)
RNA interference
(RNAi)
Short interfering RNAs
(siRNA) of 21–22
nucleotide
Selectively silence the delta opioid
receptor, but not mu opioid
receptors
Luo et al. (2005)
588 GATCHEL, PENG, PETERS, FUCHS, AND TURK
tonomic neurons derived from the neural crest. This has
been supported by the following findings: (a) a reduced
evoked potential and lack of pain experience following
electrical shock, (b) self-mutilation and fractures, (c) lack
of flare response to histamine injection, and (d) lack of
temperature regulation.
2. Overexpression of endogenous opioids leads to suppres-
sion of nociceptive transmission either peripherally or
centrally.
3. Reduced number of primary afferent nociceptors elimi-
nates the ability to initiate a nociceptive signal in the
periphery, whereas loss of neurons in sympathetic gan-
glia contributes to anhidrosis.
4. Loss of trkA function (receptor tyrosine kinase type A for
nerve growth factor [NGF]) results from mutations of the
trkA receptor gene, which is located on chromosome 1. It
has also been demonstrated that the presence of a trkA
mutation in B lymphocytes results in a lymphocyte sig-
naling defect, which could contribute to recurrent epi-
sodes of fever.
Other genes involved in various sensory transmissions have
been identified. Mechanoreceptor plays a role in transduction of
mechanical force by opening ion channels that link to extracellular
matrix and the cytoskeleton. Opening of these channels leads to
excitation of mechanoreceptors. Subunits of these ion channels
have been demonstrated in cutaneous mechanoreceptors, which are
known as BNC1 (brain type Na channel), a nonvoltage-dependent
sodium channel, and DRASIC (dorsal root acid sensitive ion
channel), both belonging to the DEG/ENaC (epithelial Na⫹chan-
nel/nematode degenerins) family. In DRASIC-knock-out mice, the
sensitivity to light touch is increased, but the sensitivity to noxious
pinch is reduced.
NGF has been playing critical roles in developing and main-
taining the survival of the nerves (especially sympathetic nerves)
and contributing to increased nociception. A loss of primary af-
ferent and sympathetic neurons has also been found in null mutants
for NGF. These mice also show depletion of immunoreactivity for
trkA receptors and calcitonin gene-related peptide (CGRP) and
substance P (SP). Both CGRP and SP are important neuropeptides
found in A-delta and C fibers, the ones responsible for transmis-
sion of nociceptive and thermal signals. Most of the animals die
within a week. Those who survive show almost no response to
noxious mechanical and thermal stimuli. On the other hand, there
is an increased sympathetic innervation of dorsal root ganglion
cells in mice with an overexpressed NGF. They show exaggerated
responses to noxious mechanical and thermal stimuli. In addition,
deletion of neurokinin-1 receptors (receptors for SP) in mice
shows normal response to brief noxious mechanical stimuli, but a
reduction in response to intradermal injection of capsaicin and the
second phase of a formalin test. Mice with deletion of the CGRP
gene have normal responses to noxious stimuli but fail to develop
secondary heat hyperalgesia by kaolin and carrageenan. When
carrageenan is used to induce inflammatory pain, inbred and
outbred rat strains differ in their pain sensitivity, as tested by
mechanical stimulation (the von Frey monofilament test) and
noxious heat pain (the Hargreaves radiant heat test), also suggest-
ing a genetic basis for differential sensitivity to pain.
A specific block of the morphine effect in mice with deletion of
the mu-opioid receptor gene has also been found. Recent studies
with inbred and knockout mice have revealed that the mu-opioid
peptide receptor encoded by the Oprm1 gene has a crucial role in
the analgesic and addictive properties of opiate drugs. Differences
in Oprm1 gene sequences affect the amount of Oprm1 messenger
RNA and sensitivity to opiates, and more than 100 polymorphisms
have been identified in the human OPRM1 gene, some of which
are related to vulnerability to drug dependence in some popula-
tions.
Genetic modulation of intracellular signal transduction mole-
cules has played a significant role in pain transmission. Deletion of
the R1subunit of protein kinase A in mice shows a reduction of
allodynia by tissue damage, a reduction of the responses to the
second phase of formalin test, and central sensitization caused by
intrathecal injection of PGE2. In contrast, mice with deletion of the
gamma isoform of protein kinase C show normal responses to
acute noxious stimuli but fail to develop neuropathic pain after
partial sciatic nerve injury. Tissue injury-induced inflammatory
and nerve injury-induced neuropathic pain (expressed as neuronal
plasticity) is generated by injury and intense noxious stimuli to
trigger an increased excitability of nociceptive neurons in the
spinal cord. This central sensitization is an activity-dependent
functional plasticity that results from activation of different intra-
cellular kinase cascades, leading to the phosphorylation of key
membrane receptors and channels and increasing synaptic effi-
cacy. Several different intracellular signal transduction cascades
converge on mitogen-activated protein kinase (MAPK). The acti-
vation of MAPK appears to be a master switch or gate for the
regulation of central sensitization. In addition to posttranslational
regulation, the MAPK pathway may also regulate long-term pain
hypersensitivity via transcriptional regulation of key gene prod-
ucts. Furthermore, activated microglia is a key cellular intermedi-
ate step in the pathogenesis of nerve injury-induced pain hyper-
sensitivity. This is supported by the observation that p38 MAPK,
together with P2X
4
purinoceptors, are present in activated micro-
glia and are required molecular mediators.
Similar to other neural transmission, transmission of pain sig-
nals requires a variety of molecules, including neurotransmitters,
neuromodulators, neurotransmitter receptors, signal transduction
molecules, and enzymes involved in protein synthesis. To ensure a
normal synaptic transmission, it is crucial to have a normal process
of protein synthesis (transcription and translation from the genetic
code), neurotransmitter transportation, storage, release, receptor
binding, and breakdown or reuptake. To accomplish these com-
plicated processes, various proteins or peptides are playing either
vital or supporting roles. Any malfunction of each individual step
will cause either elevated or reduced transmission of pain signals.
Some important molecules for pain processing include substances
that act on (a) neurotransmitters and neuromodulators (e.g., bra-
dykinin, capsaicin, CGRP, glutamate, histamine, serotonin, nor-
epinephrine, neuropeptide Y, prostaglanding E
2
, and SP); (b)
membrane receptors (mu and delta opiate receptors, purinergic
receptor P2X3, tyrosine kinase receptor A, and vanilloid receptor
1); (c) ion channels (e.g., Na
⫹
,K
⫹
, and Ca
⫹⫹
channels,
tetrodotoxin-resistance Na
⫹
channels); (d) intracellular signal
transduction molecules (e.g., R1subunit of protein kinase A and
589
BIOPSYCHOSOCIAL APPROACH TO CHRONIC PAIN
gamma isoform of protein kinase C); and (e) enzymes (e.g.,
fluoride-resistant acid phosphatase).
In summary, with all this aforementioned evidence of how gene
expression can modulate the sensitivity of pain, with individual
variation, a new direction for screening individual patients for
genetic susceptibility will provide a potential targeted treatment of
pain in the future. Indeed, three genetic haplotypes of the gene
encoding catecholamine-O-methyltransferase is significantly asso-
ciated with variation in sensitivity to experimental pain and is also
correlated to the risk of developing temporomandibular joint dis-
order (Diatchenko et al., 2005). The serotonin transporter gene is
also a promising candidate locus for the genetic susceptibility of
migraine (Szilagyi et al., 2006). Eventually it may become possible
to “turn on” or “turn off” a single gene or batch of gene expression
to relieve patient suffering from various types of pain.
Electrophysiology
Since the perception of pain is mainly dependent on the neuro-
nal activities along the axis of the somatosensory system through
signal reception, transduction, action potential generation, and
action potential propagation, it makes electrophysiological record-
ing the most direct measurement to study pain nociceptive pro-
cessing. It provides the most accurate temporal responses of the
nervous system in response to external stimuli (mechanical, ther-
mal, chemical, and electrical). In general, there are five electro-
physiological approaches to study the peripheral and central neu-
ronal activities involved in pain processing at various levels: (1)
extracellular recording in vivo from axon tracts, individual axons,
or cell body of neurons; (2) intracellular recording in vivo; (3)
intracellular recording from neurons in intact ganglia or tissue
slices in vitro; (4) intracellular recording from dissociated neurons
in vitro; and (5) patch clamp recording in vitro and in vivo. A
simple illustration is presented in Figure 4.
Extracellular recording in vivo. Extracellular recording in
vivo has been widely used in the primary afferent neurons, spinal
cord dorsal horn neurons, brainstem, thalamus, and the cortex. The
advantages of extracellular recording include (a) a complete char-
acterization of receptive fields, response properties, and conduc-
tion velocity by primary afferent recording in vivo; (b) minimizing
the amount of tissue injury to gain the access to the afferents; (c)
the ability to study changes in peripheral terminals of sensory
neurons; and (d) the ability to activate brain regions to study the
central descending modulation of the primary afferent inputs.
Population responses can be recorded by cord dorsum potential
and intraspinal field potentials extracellularly, which are the dis-
tributions of activity that are evoked in large populations of spinal
cord neurons by stimulation of primary afferent fibers. The poten-
tials reflect, in large part, the depolarization of interneurons or of
primary afferent fibers in the dorsal horn. The cord dorsum po-
tentials can be recorded from the dorsal surface of the spinal cord
in response to electrical stimulation of myelinated cutaneous af-
ferent fibers in a peripheral nerve, which include an afferent
volley, one or more negative (N) waves, and a positive wave
(Beall, Applebaum, Foreman, & Willis, 1977; Gasser & Graham,
1933; Hughes & Gasser, 1934a, 1934b; Lindblom & Ottosson,
1953a, 1953b; Willis, Weir, Skinner, & Bryan, 1973). The nega-
tive potentials can be subsequently named N1 (evoked by A␣
fibers), N2 (evoked by A␣and A ␦fibers), and N3 (evoked by
A␦fibers). The maximal response of these negative potentials can
be recorded within the spinal cord (Beall et al., 1977). The nega-
tive potential recorded in the extracellular space is due to the
moving of positively charged ions into dorsal horn neurons that
occurs during excitatory postsynaptic potentials and action poten-
tials. The positive wave that follows the N waves, evoked by
stimulation of cutaneous nerve, reflects a long-lasting depolariza-
tion of primary afferent fibers. This part of the cord dorsum
potential corresponds to the negative dorsal root potential, which
can be recorded from a disconnected filament of dorsal root
(Barron & Matthews, 1938; Eccles & Krnjevic, 1959; Eccles,
Magni, & Willis, 1963; Eccles, Schmidt, & Willis, 1963; Lloyd,
1952; Lloyd & McIntyre, 1949). Primary afferent fibers are con-
sidered to be one of the mechanisms responsible for the inhibitory
process known as presynaptic inhibition (Eccles, 1964; Rudomin
& Schmidt, 1999; R. F. Schmidt, 1971; Willis, 1999).
In the spinal cord, ascending tract neurons or motor neurons can
be distinguished from interneurons by antidromic activation fol-
lowing stimulation of their axons near projection targets in the
brain or of motor axons in a ventral root or peripheral nerve.
Criteria for antidromic activation include (a) the action potential
follows the stimulus at a constant latency, (b) collision between
orthodromic and antidromic action potentials, and (c) the anti-
dromic action potential can follow high frequencies of stimulation
(Trevino, Coulter, & Willis, 1973).
Some of the most important discoveries about the nature of pain
and nociception were determined with extracellular recordings.
The finding of the superficial laminae of the spinal cord for
nociception demonstrated that these neurons responded to mechan-
ical and thermal nociceptive inputs in lamina I (B. N. Christensen
& Perl, 1970) and II (Kumazawa & Perl, 1978). Another important
finding was that the plasticity of neuronal responses is located
deeper in the dorsal horn. An enhanced response (“windup”) was
demonstrated when peripheral nerves were stimulated at C-fiber
intensities (Mendell & Wall, 1965; Woolf, 1996). Clinically,
windup has been reported in fibromyalgia patients compared to
normal controls, suggesting that central sensitization contributes to
processes underlying hyperalgesia and persistant pain states (Price
et al., 2002; Staud, Price, Robinson, Mauderli, & Vierck, 2004).
One type of extracellular recordings is the compound action
potential recording from nerves and fiber tracts, which is a record
of the various peaks related to the conduction velocity of various
axon population in the peripheral nerve (Clark, Hughes, & Gasser,
1935; Gasser, 1941). Recording of compound action potentials in
humans is crucial to determine the impulse conduction in the
slowest fibers, which is correlated to the sensation of pain (Collins,
Nulsen, & Randt, 1960; Heinbecker, Bishop, & O’Leary, 1933).
Field potential in the CNS tracts is also a valuable approach in
determining the rostrocaudal distribution of nociceptive primary
afferent axons and their terminal arborizations (Traub & Mendell,
1988; Traub, Sedivec, & Mendell, 1986).
Microneurography (Hagbarth & Vallbo, 1967) in humans is
another extracellular recording technique, which led to many hu-
man studies that clearly defined the involvement of unmyelinated
C-fibers in pain sensation and some pathological conditions (Hag-
barth, 1979; Hallin & Wu, 1998; Ochoa & Torebjo¨rk, 1980;
Ochoa, Torebjo¨rk, Culp, & Schady, 1982; Torebjo¨rk & Hallin,
1970; Torebjo¨rk, Ochoa, & McCann, 1979; Vallbo, Hagbarth,
Torebjo¨rk, & Wallin, 1979; Van Hees & Gybels, 1972). The
590 GATCHEL, PENG, PETERS, FUCHS, AND TURK
advantage of microneurography is that recorded axons can be
stimulated relatively selectively following isolation of a single unit
(Simone, Marchettini, Caputi, & Ochoa, 1994; Torebjo¨rk, Vallbo,
& Ochoa, 1987), while allowing the human subject to describe
accurately the quality and intensity of pain (Marchettini, Simone,
Caputi, & Ochoa, 1996; Ochoa & Torebjo¨rk, 1989; Torebjo¨rk et
al., 1987) as well as of an itch (Schmelz et al., 2003; Schmelz,
Schmidt, Bickel, Handwerker, & Torebjo¨rk, 1997).
Single unit recording from axon fibers of the peripheral nerve
has also been tested on their own peripheral nerves by some
neuroscientists (Hensel & Boman, 1960) and later were mostly
used in animals studies (LaMotte & Campbell, 1978). It is a
relatively simple technique, but it is crucial for determining the
spontaneous discharge in sensory fibers after peripheral nerve
injury, especially when ectopic spikes are generated from the
ganglion (Kajander, Wakisaka, & Bennett, 1992; Xie, Zhang,
Petersen, & LaMotte, 1995).
Intracellular recording. Intracellular recording can directly
measure the membrane potential change, and one can inject a dye
into the recording neuron for labeling purposes. Technically, it is
more difficult than extracellular recording and may not be acces-
sible to small diameter fibers. In addition, it may cause damage of
the neuron because of the nature of the technique (puncture of the
cell membrane by a sharp electrode). By directly monitoring the
membrane potential, intracellular recording in vivo contributes
important information of the classification of sensory neurons in
response to peripheral receptive properties (Djouhri, Bleazard, &
Lawson, 1998; Giesler, Gerhart, Yezierski, Wilcox, & Willis,
1981; Koerber, Druzinsky, & Mendell, 1988; Lawson, Crepps, &
Perl, 1997; Ritter & Mendell, 1992; Willis, Trevino, Coulter, &
Maunz, 1974), such as inflammation or nerve injury (Czeh, Kudo,
& Kuno, 1977; Djouhri & Lawson, 1999). Intracellular recording
of substantia gelatinosa neurons has demonstrated direct modula-
tion of their activity through stimulation of brainstem structures
(NRM and PAG; Bennett, Hayashi, Abdelmoumene, & Dubner,
1979; Light, Casale, & Menetrey, 1986; Steedman, Molony, &
Iggo, 1985).
When neurons in a part of the nervous system are isolated and
set in a recording chamber, intracellular recording from a tissue
slice in vitro has several advantages over in vivo. It has the
following properties: better control of the extracellular milieu (e.g.,
absence of blood-brain barrier), some degree of electrical control
of the soma membrane, possible identification of primary afferents
and their receptive field properties, a possible observation of
injury-induced increase in excitability, and a condition without
enzymatic or mechanical treatment prior to recording. However,
intracellular recording in vitro also suffers from several disadvan-
tages. For example, it may not be able to determine the change in
response properties due to a direct change in the neuron or an
indirect change caused by surrounding cells. In an isolated envi-
ronment, there is an absence or a low level of the proteins neces-
sary for the transduction of stimuli.
Figure 4. An illustration of an electrophysiological setup for extracellular, intracellular, and patch-clamp
recording in either in vivo or in vitro preparations.
591
BIOPSYCHOSOCIAL APPROACH TO CHRONIC PAIN
Intracellular recording from dissociated neurons in vitro has
several advantages as compared with slice preparation. It allows
for complete control of the extracellular milieu, as well as the
intracellular milieu, when used with patch-clamp techniques. Re-
distribution of proteins to the plasma membrane that are normally
presenting afferent terminals has been observed. For example, a
proton receptor/ion channel complex that is usually present in
terminals has been demonstrated in the isolated cell body (Bevan
& Yeats, 1991; Steen, Issberner, & Reeh, 1995). However, this
approach suffers from several disadvantages. It is impossible to
identify the primary afferents with respect to conduction velocity
or receptive field properties. There is a potential for damage of the
membrane properties due to enzymatic treatment. There is also the
potential for alteration of neuron properties because of a lack of
unknown important factors. The results cannot be applied directly
to the conditions in the behaving animals because of lack of
supporting cells and other neurons.
Patch-clamp recording in vitro. First described by Neher and
Sakmann (1976), patch-clamp recording in vitro is now a powerful
method for studying electrophysiological properties and chemo-
sensitivity of neurons involved in the transduction and transmis-
sion of nociceptive stimuli. It is widely used to study the primary
afferent terminals (Brock, McLachlan, & Belmonte, 1998; Reid,
Scholz, Bostock, & Vogel, 1999; Scholz, Reid, Vogel, & Bostock,
1993), soma of the sensory ganglion (Huang & Neher, 1996; L.
Liu & Simon, 1996; L. Liu, Wang, & Simon, 1996; Todorovic &
Anderson, 1990), dissociated central neurons (Reichling, Kyrozis,
Wang, & MacDermott, 1994; Rusin, Jiang, Cerne, & Randic,
1993), slice preparations (Baba et al., 1998; Bao, Li, & Perl, 1998;
Pan, 1998; Pan & Fields, 1996; Pan, Tershner, & Fields, 1997;
Schneider, Eckert, & Light, 1998; Yoshimura & Nishi, 1993), and
in vivo (Light & Willcockson, 1999). Even though cell-attached
recording of an afferent terminal of corneal afferents was reported
(Brock et al., 1998), this approach has been used to record ion
channel activity from C-fiber axons, as well as the afferent cell
body (Reid et al., 1999; Scholz et al., 1993). One of the advantages
of this technique is that it can generate the most detailed informa-
tion of the biophysical properties of the ion channel. It is also
possible to record from specific sites on a neuron to obtain infor-
mation of the relative distribution of ion channels. However, the
disadvantage is that it is the most technically difficult and labor
intensive electrophysiological approach.
A relative easier target is the cell body of the primary afferent
neuron. It has been suggested that the cell body of acutely isolated
sensory neurons in vitro is a valid model for the afferent terminal
in vivo. Receptors and ion channels in the peripheral or central
terminals of sensory neurons are present and are functional in the
plasma membrane of the cell body in vitro (Huang & Neher, 1996;
L. Liu & Simon, 1996; L. Liu et al., 1996; Todorovic & Anderson,
1990). Pharmacologically, these receptors on the cell body show
properties similar to those near the peripheral and central terminals
(Carlton & Coggeshall, 1997; Carlton, Zhou, & Coggeshall, 1999;
X. Chen, Belmonte, & Rang, 1997; X. Chen, Gallar, & Belmonte,
1997; Coggeshall & Carlton, 1998; H. Liu, Wang, Sheng, Jan, &
Basbaum, 1994). It is also possible to induce a similar change in
the excitability of the cell body in vitro with the same manipula-
tions that induce changes in the peripheral terminals in vivo. For
example, PGE2 can induce sensitization of the cell body in vitro
(Baccaglini & Hogan, 1983; Fowler, Wonderlin, & Weinreich,
1985; Gold, Dastmalchi, & Levine, 1996; Nicol & Cui, 1994;
Vasko, Campbell, & Waite, 1994; Weinreich & Wonderlin, 1987).
Furthermore, the sensory neuron cell body in vitro can be induced
to release neurotransmitters (Gu, Albuquerque, Lee, & MacDer-
mott, 1996; Gu & MacDermott, 1997; Lee, Engelman, & Mac-
Dermott, 1999; MacDermott, Role, & Siegelbaum, 1999), which is
aCa
2⫹
-dependent process (Huang & Neher, 1996).
Patch-clamp recording in vivo. Following the successful ap-
plication of patch-clamp recording in vivo in other systems
(Covey, Kauer, & Casseday, 1996; C. I. Moore & Nelson, 1998),
use of whole-cell recording techniques in the nociceptive systems
of the spinal cord of the rat in vivo was reported (Furue, Narikawa,
Kumamoto, & Yoshimura, 1999; Graham, Brichta, & Callister,
2004; Light & Willcockson, 1999; Weng & Dougherty, 2002;
Yoshimura, Doi, Mizuno, Furue, & Katafuchi, 2005). The obvious
advantages of this technique include better control over the elec-
trical properties of the neuron, a more robust technique than sharp
electrode intracellular recording, the ability to observe single chan-
nel activity in the native milieu, and easier control of both intra-
cellular and extracellular medium for drug application. However,
this technique suffers from difficulties in stabilization of animals
and obtaining adequate seals due to movement and covering glial
cells. It is not the best choice for obtaining large samples in a
study.
Imaging
A variety of imaging techniques have been developed and used
to study pain and nociception. Functional imaging techniques have
played a crucial role because of the advantage of correlating the
brain activity with human perception.
Positron emission tomography (PET). Since its development
in the 1970s, PET has been used for imaging human brain func-
tion. A PET image is created by the detection of positrons emitted
from an intravenously injected radionuclide (i.e., the tracer).
Through blood circulation, the tracer is distributed to the brain. As
the tracer decays, it emits a positron which travels a few millime-
ters, collides with an electron, and releases two photons (gamma
rays) in opposite directions. A series of PET detectors over the
head detects the signals, which are used to create a tomographic
image. PET images can be overlapped with a subject’s own fMRI
to fit onto a standardized atlas to be visualized. Depending on the
half-life, different radionuclides can be used for different purposes.
For example,
15
O is used to measure cerebral blood flow in
activation studies because of its half-life (2 min);
18
F with a
half-life of 110 min can be used to measure cerebral glucose
metabolism;
11
C with a half-life of 20 min can be used to study
receptor binding of dopamine, benzodiazepine, and opiates (Keg-
eles & Mann, 1997; Phelps & Mazziotta, 1985; Slifstein & Laru-
elle, 2001; Tai & Piccini, 2004). PET can be used in three major
ways. The receptor density and binding properties of ligands in the
brain can be identified by injecting a radioactive receptor antago-
nist or agonist (Sadzot et al., 1991). It can also be used to measure
regional cerebral blood flow (rCBF) in the resting state to detect
neurological abnormalities in disease or injury (Hsieh, Belfrage,
Stone-Elander, Hansson, & Ingvar, 1995; Iadarola et al., 1995;
Peyron et al., 1998). Finally, in activation studies, [
15
O]H
2
Ois
injected to identify task-related changes in blood flow. The advan-
tages of PET include a relatively open, noise-free environment that
592 GATCHEL, PENG, PETERS, FUCHS, AND TURK
can accommodate most experimental or internal devices. The
disadvantages include that it is relatively expensive, injection of
radioactive tracer is invasive, the time frame is relatively restricted
because of the half-life of the tracer, and it has a moderate to poor
spatial resolution and poor temporal resolution.
The PET scan was first applied to study acute pain (Talbot et al.,
1991), which identified four cortical regions of activation by
noxious heat stimuli: primary and secondary somatosensory cor-
tex, anterior insula, and anterior cingulated cortex (ACC). Subse-
quent studies have confirmed and extended cortical areas in ther-
mal, mechanical, and laser-evoked pain, such as prefrontal cortex,
supplemental motor cortex, basal ganglia, cerebellum, and the
hypothalamus and periaqueductal gray (Aziz, 1997; Casey, 1999;
Casey et al., 1994; Casey, Minoshima, Morrow, & Koeppe, 1996;
Coghill et al., 1994; Derbyshire et al., 1997; Hsieh et al., 1996;
Jones, Brown, Friston, Qi, & Frackowiak, 1991; Svensson,
Minoshima, Beydoun, Morrow, & Casey, 1997; Xu et al., 1997).
An example is provided in Figure 5. The widespread cortical
activations identified in pain studies have been implicated in
affective, cognitive, and reflexive responses to a painful stimulus,
which demonstrate that there is a distributed network of many
brain areas that are recruited by a painful stimulus that contribute
to the multidimensional experience (Coghill et al., 1994; Coghill,
Sang, Maisog, & Iadarola, 1999). In a study that used hypnosis to
manipulate pain unpleasantness independent of pain intensity, a
relationship between unpleasantness and ACC activation was iden-
tified (P. Rainville, Duncan, Price, Carrier, & Bushnell, 1997). The
ACC was uniquely activated both in real pain and during an
illusion of pain evoked by simultaneous warm and cool stimuli
(Craig, Reiman, Evans, & Bushnell, 1996). PET scans obtained
during motor cortex stimulation for chronic pain revealed activa-
tion of the thalamus, ACC, anterior insula, and frontal cortex
(Garcia-Larrea et al., 1999; Peyron et al., 1995). Thalamic stimu-
lation also activates the ACC (Davis et al., 2000) and the anterior
insula, which is accompanied by thermal sensations (Duncan et al.,
1998).
Single-photon emission computerized tomography (SPECT).
Similar to PET, the SPECT scanner has a gamma camera that
detects emissions from decaying isotopes, which have long half-
lives (often in the order of hours). Cerebral blood flow can be
measured with the inhalation of Xenon-133 or technetium 99mTc
HM-PAO (99m-hexamethylpropyleneamineoxime; Canavero et
al., 1993; Di Piero, Pantano, Ricci, & Lenzi, 1993). The long
retention of SPECT tracers allows for flexibility in the timing of
data acquisition, after injection, of the overall effect of an injury,
disease, stimulus, treatment, or other manipulation (Prichard &
Brass, 1992). However, the SPECT scanner is more expensive and
suffers from poor spatial and temporal resolution.
Through use of SPECT, different types of pain have been
examined. A decrease in cortical rCBF in the SI region associated
Figure 5. A positron emission tomography scan image of rCBF responses of 10 males (M) and 10 females (F)
to repetitive noxious heat stimulation (50°C) of the left volar forearm. Significant activation of the contralateral
anterior cingulate cortex, premotor, insular cortex, ipsilateral insula, and bilateral cerebellar vermis has been
identified. From “Gender Differences in Pain Perception and Patterns of Cerebral Activation During Noxious
Heat Stimulation in Humans,” by P. E. Paulson, S. Minoshima, T. J. Morrow, and K. L. Casey, Pain, 76, 1998,
p. 227. Copyright 1998 by the International Association for the Study of Pain. Reprinted with permission.
593
BIOPSYCHOSOCIAL APPROACH TO CHRONIC PAIN
with a long (3 min) sustained contralateral heat pain stimulus was
found (Apkarian et al., 1992), whereas an increased rCBF in SI
contralateral to a tonic cold-pain stimulus was also reported in a
cold pressor test (Di Piero et al., 1994) and among cluster head-
ache patients (Di Piero, Fiacco, Tombari, & Pantano, 1997). In a
chronic pain state, such as painful restless legs syndrome and
spinal cord injury, there is an increase in blood flow in the
contralateral SI, ACC, and thalamus (see Figure 6; Ness et al.,
1998; San Pedro et al., 1998). However, hypoperfusion was also
found in the caudate of a patient with spinal cord injury pain (Ness
et al., 1998), in the frontoparietal region (Ogawa, Lee, Nayak, &
Glynn, 1990), and in the thalamus of patients with central pain
(Pagni & Canavero, 1995; Tanaka et al., 1997).
Functional magnetic resonance imaging (fMRI). fMRI is now
widely used in the medical field to obtain normal or pathological
anatomical changes. It was first used to study brain function by
intravenous injection of gadolinium as a contrast agent to enhance
the magnetic resonance signals in the visual cortex evoked by
flashing light (Belliveau et al., 1991). It was soon realized that
visualization of these signals did not require injection of a contrast
agent because the body has its own natural contrast agent, deox-
ygenated hemoglobin (Ogawa et al., 1992; Ogawa, Lee, Kay, &
Tank, 1990; Ogawa, Lee, Nayak, & Glynn, 1990). Most fMRIs
now rely on the blood oxygenation level dependent effect, which
is based on the increased neuronal firing in response to a stimulus
that will induce hemodynamic changes and ultimately modify the
magnetic field to increase the fMRI signals (Porro, Lui, Facchin,
Maieron, & Baraldi, 2004). It is thought that an increased meta-
bolic demand, due to increased neuronal activity, results in an
increase in blood flow beyond metabolic needs, such that the final
ration of deoxyHb/oxyHb actually is reduced. It is the reduction in
deoxyHb that alters the magnetic field properties and produces the
increased fMRI signal (DeYoe, Bandettini, Neitz, Miller, &
Winans, 1994). The advantages of fMRI include its noninvasive-
ness, good spatial (down to 1–2 mm), and temporal (hundreds of
ms is possible) resolution. The disadvantages include its expense,
limited availability due to time-sharing on a clinical scanner,
restriction to metallic devices, and loud noise.
A large body of literature, nevertheless, has been conducted
investigating brain mechanisms underlying both chronic and acute
pain (see one example in Figure 7). fMRI has been used to study
stimulus-related responses, such as noxious electrical stimulation
of the skin or peripheral nerve (Davis, Taylor, Crawley, Wood, &
Mikulis, 1997; Davis, Wood, Crawley, & Mikulis, 1995; Oshiro et
al., 1998); noxious heat or cold (Apkarian, Darbar, Krauss, Gelnar,
& Szeverenyi, 1999; Becerra et al., 1999; Davis, Kwan, Crawley,
& Mikulis, 1998; Gelnar, Krauss, Sheehe, Szeverenyi, & Apkar-
ian, 1999; Ploghaus et al., 1999; Raij, Forss, Stancak, & Hari,
2005); and mechanical (Disbrow, Buonocore, Antognini, Carstens,
& Rowley, 1998), chemical (Maiho¨fner & Handwerker, 2005;
Porro, Cettolo, Francescato, & Baraldi, 1998), or visceral stimuli
(Binkofski et al., 1998). Brain areas that can be activated include
the primary somatosensory cortex (SI), secondary somatosensory
cortex (SII), anterior insula, ACC, thalamus, and cerebellum. With
the combination of psychophysical assessment and fMRI, pain-
related activations have been obtained in parallel psychophysical
sessions (Apkarian et al., 1999) or during the imaging sessions
(Davis et al., 1997; Porro et al., 1998; Raij et al., 2005) in order to
separate those activations that are due to the mere presence of a
stimulus (i.e., due to attention) from those that are related to the
subjects’ actual sensory experiences. On the other hand, the rela-
tionship and interaction of pain, attention, and anticipation has
been demonstrated to activate slightly different areas of the brain
(Davis et al., 1997; Ploghaus et al., 1999). For example, anticipa-
tion of pain activated the anterior ACC, whereas the pain itself
activated the posterior ACC. Further dissection on this line dem-
onstrated that the posterior insula/secondary somatosensory cortex,
the sensorimotor cortex, and the caudal ACC were specific to
receiving pain, whereas the anterior insula and rostral ACC acti-
vation correlated with individual empathy scores when the subjects
watched their loved ones receiving pain stimuli (Singer et al.,
2004). The underlying modulatory effect of expectation on pain
transmission might involve activation of descending modulatory
systems (Keltner et al., 2006).
Recently, a caution has been raised about fMRI (Nair, 2005;
Savoy, 2005), especially when it is used in dissecting the cognitive
and emotional mechanisms, because cognitive function is a mov-
ing target. It is difficult to design a good fMRI study and to analyze
and interpret the data. The notion of whether fMRI is a modern
phrenology is under debate (D. I. Donaldson, 2004; Terrazas &
McNaughton, 2000; Uttal, 2001). Regardless, fMRI has been used
to image allodynia in complex regional pain syndrome (Maiho¨fner,
Handwerker, & Birklein, 2006) and patients suffering from neu-
ropathic pain (Schweinhardt et al., 2006).
Magnetoencephalography (MEG). MEG is a technique that
detects weak magnetic fields within the human brain. MEG is the
most sensitive for cortical neuronal activity because electrical
currents generated by neurons induce a perpendicularly oriented
magnetic field, which can be detected directly outside the head by
MEG detectors (Hari & Forss, 1999; Naatanen, Ilmoniemi, &
Alho, 1994). The major advantage of MEG is that it is noninvasive
and has excellent temporal resolution (down to milliseconds) to
directly measure neuronal activity. It can also be superimposed
onto a high-resolution fMRI to provide good spatial localization.
Figure 6. A single-photon emission computerized tomography image
showing baseline scan (top row) and postacupuncture scan (bottom row) of
the thalamic activity. From “Cerebral Blood Flow Effects of Pain and
Acupuncture: A Preliminary Single-Photon Emission Computed Tomo-
graphy Imaging Study,” by A. B. Newberg, P. J. LaRiccia, B. Y. Lee, J. T.
Farrar, L. Lee, A. Alavi, 2005, Journal of Neuroimaging, 15, p. 45.
Copyright 2005 by the American Society of Neuroimaging. Reprinted with
permission.
594 GATCHEL, PENG, PETERS, FUCHS, AND TURK
The current major disadvantage is the high cost of the device and
the necessity for a magnetically shielded space.
In Figure 8, a sample MEG image indicates activation of the
primary motor cortex while stimulation is applied at A-delta and
C-fiber intensities (Raij et al., 2005). Use of MEG in the study of pain
demonstrated responses to electrical stimulation of the digit at short
latencies in the contralateral SI (at 40–60 ms), followed by longer
latencies in the ipsilateral and bilateral SII and insula (100–250 ms)
(Howland, Wakai, Mjaanes, Balog, & Cleeland, 1995) as well as the
ACC (Kitamura et al., 1995, 1997). A study of painful laser-evoked
responses reported that both the contralateral SI and SII dipoles
occurred at around 130 ms, suggesting a parallel processing of
thalamocortical inputs to these two cortical regions (Ploner, Schmitz,
Freund, & Schnitzler, 1999). MEG has also been used to follow the
extent of cortical plasticity in phantom limb pain in traumatic or
congenital amputees (Flor et al., 1995, 1998). It has also been found
that focally applied brief painful stimulus generates a global suppres-
sion of spontaneous oscillations in somatosensory, motor, and visual
areas, indicating that pain induces a widespread change in cortical
function and excitability (Ploner, Gross, Timmermann, Pollok, &
Schnitzler, 2005; Raij et al., 2005).
Intrinsic optical signals (IOS) and intrinsic optical imaging
(IOI). Changes in optical signals of transmitted or reflected light
through brain tissue can indicate regional differences in brain
activity. The transmitted or reflected light through brain tissue can
be detected or imaged without using dyes or florescent markers.
Use of IOS and IOI to monitor and understand neural activities and
physiological changes in vitro and in vivo becomes more recog-
nized (Asai, Kusudo, Ikeda, & Murase, 2002; Ikeda, Terakawa,
Murota, Morita, & Hirakawa, 1996; Johnson, Hanley, & Thakor,
2000; Kristal & Dubinsky, 1997; Lemasters, Nieminen, Qian,
Trost, & Herman, 1997; Miller, Petrozzino, Mahanty, & Connor,
1993; Scarfone, McComas, Pape, & Newberry, 1999; Uchino,
Elmer, Uchino, Lindvall, & Siesjo, 1995). It is now popularly
utilized in the field of neuroscience (Aitken, Fayuk, Somjen, &
Turner, 1999; Andrew, Jarvis, & Obeidat, 1999; Haller, Mironov,
& Richter, 2001; Nomura, Fujii, Sato, Nemoto, & Tamura, 2000).
Subcellular organelles, such as nuclei (Ikeda et al., 1996), and
mitochondria (Kristal & Dubinsky, 1997; Lemasters et al., 1997),
are known to change size with different levels of tissue activity and
injury (Johnson et al., 2000). It is also known that changes of
particle size within tissue will result in changes of light scattering.
Thus, IOS and IOI of brain tissue have been used to investigate a
variety of brain physiology models for more than 25 years (Lipton,
1973), although the causes of changes in light scattering and
birefringence (initially studied 30 years ago by L. B. Cohen,
Keynes, & Hille, 1968, in giant axons of squid) are still not
completely understood (Johnson et al., 2000; Nomura et al., 2000).
Figure 7. An image from functional magnetic resonance imaging showing key areas of activations in primary
somatosensory cortex (S1), primary motor cortex (M1), secondary somatosensory cortex (S2), parietal associ-
ation cortex (PA), inferior parietal lobule (IPL), superior frontal cortex (SFC), middle frontal cortex (MFC),
inferior frontal cortex (IFC), and cingulate cortex (GC) when pin-prick (A, B) or thermal (C, D) stimuli were
applied before (A, C) or after (B, D) capsaicin injection. From “Differential Coding of Hyperalgesia in the
Human Brain: A Functional MRI Study,” by C. Maiho¨fner and H. O. Handwerker, 2005, NeuroImage, p. 1000.
Copyright 2005 by Elsevier. Reprinted with permission.
595
BIOPSYCHOSOCIAL APPROACH TO CHRONIC PAIN
However, both IOS and IOI are highly affected by light scattering
and absorption of the measured neural tissue of the brain, mainly
as a result of morphological structures and hemodynamic (such as
blood concentrations, blood oxygenation levels, and blood flow)
aspects of the brain, respectively. So far, it is difficult to separate
light scattering and absorption effects within the measured data of
IOS and IOI. One common practice in neuroscience research using
optical brain imaging is to define a practical index, such as an
intensity index, to associate either the IOS or IOI with the neural
activity, without being able to decouple the effects from morpho-
logical and hemodynamic aspects of the brain. To date, there is no
direct evidence using this technique to address pain problems.
However, optical imaging has been used in studying the somato-
sensory systems (Berwick et al., 2005; Sasaki et al., 2002; Tom-
merdahl, Simons, Chiu, Favorov, & Whitsel, 2005; see example in
Figure 9) and the visual cortex (Blasdel, 1989, 1992). When a
sinusoidal mechanical stimulation is applied to the contralateral or
bilateral skin in the cat, there is an increase in absorption in the
primary and secondary somatosensory cortices (Tommerdahl et
al., 2005), whereas ipsilateral stimulation only elicits an increase
of absorption in the secondary somatosensory cortex.
Nanotechnology: Quantum dots. Although histological local-
ization has been used extensively through a variety of staining
techniques that can be examined under light or an electron micro-
scope, a recent development in nanotechnology, the use of quan-
tum dots, has advanced the field further. In brief, quantum dots are
fluorescent semiconductor nanocrystals (i.e., cadmium selenide)
that can be conjugated with antibodies of interested targets (i.e.,
any proteins and peptides), such as variety of ion channels, neu-
rotransmitters and their receptors, enzymes involved in neurotrans-
mitter synthesis and metabolism, and molecules involved in intra-
cellular cascade. The sizes of the quantum dots can be different
(2–9.5 nm), which causes emission of different colors (emission
wavelength from 400 to 1,350 nm) under microscope. It enables
one to label multiple targets in the same tissue to examine cellular
or subcellular structures (see Figure 10; Giepmans, Deerinck,
Smarr, Jones, & Ellisman, 2005; Michalet et al., 2005). The major
advantages of quantum dots over currently widely used fluoro-
phores include their brightness, distinguishable emission spectra,
and resistance to photobleach, which make quantum dots espe-
cially valuable for imaging anatomical structure and track physi-
ological events in in vivo (Jaiswal, Mattoussi, Mauro, & Simon,
2003; Voura, Jaiswal, Mattoussi, & Simon, 2004) or in vitro
(Goldman et al., 2004) preparations (Alivisatos, Gu, & Larabell,
2005; Jaiswal & Simon, 2004) without detectable toxicity for
weeks to months (Jaiswal et al., 2003) in noninvasive imaging
(Ballou, Lagerholm, Ernst, Bruchez, & Waggoner, 2004). Poten-
tial applications of quantum dots include bioanalytical assays,
fixed cell imaging, biosensors, in vivo animal targeting, and ex
vivo live cell imaging (Michalet et al., 2005). Although there is no
report of using quantum dots in the study of pain mechanisms, its
use is expected to explode in the near future.
Summary of Neuroscience Research
The overwhelming experimental data generated by basic neu-
roscience studies will continue to lead to a better understanding of
chronic pain. These techniques or tools include molecular biology,
anatomy, physiology, behavior, and imaging at cellular, organic,
and systemic levels. Although not yet fully developed, the current
data have the following implications for dealing with chronic pain.
1. Genetic factors may play a crucial role in the susceptibility,
initiation, maintenance, and aggravation of chronic pain.
2. Imbalance of a variety of neurotransmitters, neuromodulators,
and their various types and subtypes of receptors, may contribute
to the chronic pain state. An overproduction and release of an
excitatory neurotransmitter, for example, may increase the mem-
brane excitability, thus leading to an increased sensitivity of neu-
rons that are part of the pain transmission system. On the other
Figure 8. A magnetoencephalography image in the contralateral primary motor cortex (a) while A␦- and
C-fiber stimuli were applied (b). From “Modulation of Motor-Cortex Oscillatory Activity by Painful A␦- and
C-fiber Stimuli,” by T. T. Raij, N. Forss, A. Stancak, and R. Hari, 2004, NeuroImage, p. 571. Copyright 2004
by Elsevier. Reprinted with permission.
596 GATCHEL, PENG, PETERS, FUCHS, AND TURK
hand, lowered production of an inhibitory neurotransmitter will
play a similar function.
3. Neurons in the somatosensory system are “wired” in certain
patterns. They are dynamic and subject to constant modification
depending on incoming signals from various connections. Al-
though different neurotransmitters may act on their specific recep-
tors, they may share a similar intracellular cascade pathway or
interact with different intracellular pathway. Some of these intra-
cellular events are critical to modifying genetic expression that
may have a long-term effect, like in chronic pain.
4. Chronic pain is also phasic depending on the psychosocial
status of the patient; it may change within hours, days, or weeks,
possibly related to various hormones and their concentrations in
the system.
5. Noninvasive imaging tools (CT, fMRI) have greatly advanced
our knowledge of the anatomical and pathological conditions of
the nervous system. PET, SPECT, and MEG techniques have
added a further step in understanding the dynamic changes of the
brain in response to certain stimuli or tasks. However, interpreta-
tion of results from these relatively new technologies should be
made with some caution, especially in dealing with chronic pain.
A newly developed, much less expensive tool—optical imag-
ing—is another new potential technique, but it may suffer from
less resolution in terms of noninvasiveness. A critical component
of future research is to examine tonic neuronal activity, since some
chronic pain conditions may possibly be associated with abnor-
malities in tonic levels of activity, in noxious evoked brain acti-
vation, or both.
6. Recent developments in nanotechnology could also contribute
to the understanding of basic mechanisms (by using quantum dots
labeling) and may provide a potential therapeutical mean (by
targeted drug delivery).
Pain and Emotion
Historically, pain has been viewed as a symptom secondary to
the presence of tissue pathology and, thus, as being of secondary
importance. From this perspective, the amount of pain experienced
and reported should be directly proportional to the amount of
tissue pathology. Once the physical pathology has resolved, the
pain should subside. Emphasis then should be on treating the cause
of pain. Conversely, as noted previously, pain has also been
viewed as being outside the senses and among the emotions (e.g.,
Aristotle). A new era in thinking about pain was ushered in by the
conceptual model underlying the gate control theory by Melzack
and Casey (1968) who, as we reviewed earlier in this article,
suggested that the end experience of pain was a composite of
Figure 9. An example of the surface plots of absorbance evoked in somatosensory cortex by contralateral,
ipsilateral, and bilateral flutter stimulation in the cat. From “Response of SI Cortex to Ipsilateral, Contralateral
and Bilateral Flutter Stimulation in the Cat,” by M. Tommerdahl, S. B. Simons, J. S. Chiu, O. Favorov, and B.
Whitsel, 2005, BMC Neuroscience, 6, p. 32.
597
BIOPSYCHOSOCIAL APPROACH TO CHRONIC PAIN
sensory-discriminative, cognitive-evaluative, and motivational
features. In this view, although the three components may be
disentangled and assessed separately, they are interdependent. As
noted, the integrative model postulated by Melzack and Wall
(1965), and expanded by Melzack and Casey, has become the
dominant paradigm in the specialized field of pain and pain man-
agement; however, there continue to be vestiges of mind-body,
dualistic views in research on pain and clinical pain management.
Pain is ultimately a subjective, private experience, but it is
invariably described in terms of sensory and affective proper-
ties. As defined by the International Association for the Study
of Pain, “[pain] is unquestionably a sensation in a part or parts
of the body but it is also always unpleasant and therefore also
an emotional experience [emphasis added]” (Merskey, 1986, p.
1). The central and interactive roles of sensory information and
affective state are supported by an overwhelming amount of
evidence (Fernandez, 2002; Robinson & Riley, 1999; Smeets,
Vlaeyen, Kester & Knottnerus, 2006; Keogh & Asmundson,
2004; Turk & Monarch, 2002). The affective component of pain
incorporates many different emotions, but they are primarily
negative. Depression and anxiety have received the greatest
amount of attention in chronic pain patients; however, anger has
Figure 10. The optical properties of quantum dots (a–d) and potential applications in imaging cellular
structures (e) and lymph nodes or prostate tumor in live animal (f). From “In vivo Molecular and Cellular
Imaging With Quantum Dots,” by X. Gao, L. Yang, J. A. Petros, F. F. Marshall, J. W. Simons, and S. Nie, 2005,
Current Opinion in Biotechnology, 16,p. 67. Copyright 2005 by Elsevier. Reprinted with permission.
598 GATCHEL, PENG, PETERS, FUCHS, AND TURK
recently received considerable interest as a significant emotion
in chronic pain patients.
In addition to affect being one of the three interconnected
components of pain, pain and emotions interact in a number of
ways. Emotional distress may predispose people to experience
pain, be a precipitant of symptoms, be a modulating factor ampli-
fying or inhibiting the severity of pain, be a consequence of
persistent pain, or be a perpetuating factor. Moreover, these po-
tential roles are not mutually exclusive, and any number of them
may be involved in a particular circumstance interacting with
cognitive appraisals. For example, the literature is replete with
studies demonstrating that current mood state modulates reports of
pain as well as tolerance for acute pain (e.g., Fernandez & Turk,
1992; Turk & Monarch, 2002). Levels of anxiety have been shown
to influence not only pain severity but also complications follow-
ing surgery and number of days of hospitalization (e.g., DeGroot
et al., 1997; Pavlin, Rapp, & Pollisar, 1998). Individual difference
variables, such as anxiety sensitivity (discussed later in this arti-
cle), have also been shown to play an important predisposing and
augmenting role in the experience of pain (Asmundson, 1999;
Asmundson, Wright, & Hadjistavropoulos, 2000). Level of depres-
sion has been observed to be closely tied to chronic pain (Gatchel,
2005) and to play a significant role in premature termination from
pain rehabilitation programs (Kerns & Haythornthwaite, 1988).
Emotional distress is commonly observed in people with
chronic pain. People with chronic and recurrent (episodic) acute
pain often feel rejected by the medical system, believing that they
are blamed or labeled as symptom magnifiers and complainers by
their physicians, family members, friends, and employers when
their pain condition does not respond to treatment. They may see
multiple physicians and undergo numerous laboratory tests and
imaging procedures in an effort to have their pain diagnosed and
successfully treated. As treatments expected to alleviate pain are
proven ineffective, pain sufferers may lose faith and become
frustrated and irritated with the medical system. As their pain
persists, they may be unable to work or have financial difficulties,
difficulty performing everyday activities, sleep disturbance, or
treatment-related complications. They may be fearful and have
inadequate or maladaptive support systems and other coping re-
sources available to them. They may feel hostility toward the
health care system in its inability to eliminate their pain. They may
also feel resentment toward their significant others who they may
perceive as providing inadequate support. And, they may even be
angry with themselves for allowing their pain to take over their
lives. These consequences of chronic pain can result in depression,
anger, anxiety, self-preoccupation, and isolation—an overall sense
of demoralization. Because chronic pain persists for long periods
of time, affective state will continue to play a role as the impact of
pain comes to influence all aspects of the pain sufferers’ lives.
Although we provide an overview of research on the predomi-
nant emotions—anxiety, depression, and anger—associated with
pain individually, it is important to acknowledge that these emo-
tions are not as distinct when it comes to the experience of pain.
They interact and augment each other over time.
Anxiety
It is common for patients with symptoms of pain to be anxious
and worried. This is especially true when the symptoms are unex-
plained, as is often the case for chronic pain syndromes. For
example, in a large-scale, multicentered study of fibromyalgia
syndrome patients, between 44% and 51% of patients acknowl-
edged that they were anxious (Wolfe et al., 1990). People with
persistent pain may be anxious about the meaning of their symp-
toms and for their futures—will their pain increase, will their
physical capacity diminish, will their symptoms result in progres-
sive disability where they ultimately end in a wheelchair or bed-
ridden? In addition to these sources of fear, pain sufferers may be
worried that, on the one hand, people will not believe that they are
suffering and, on the other, they may be told that they are beyond
help and will just have to learn to live with it. Fear and anxiety also
relate to activities that people with pain anticipate will increase
their pain or exacerbate whatever physical factors might be con-
tributing to the pain. These fears may contribute to avoidance,
motivate inactivity, and ultimately greater disability (Boersma &
Linton, 2006). Continual vigilance and monitoring of noxious
stimulation and the belief that it signifies disease progression may
render even low-intensity aversive sensations less bearable. In
addition, such fears will contribute to increased muscle tension and
physiological arousal that may exacerbate and maintain pain
(Gatchel, 2005; Robinson & Riley, 1999).
Threat of intense pain captures attention in such a way that
individuals have difficulty disengaging from it. The experience of
pain may initiate a set of extremely negative thoughts, as noted
previously, and arouse fears—fears of inciting more pain and
injury or fear of their future impact (see Vlaeyen & Linton, 2000).
Fear and anticipation of pain are cognitive-perceptual processes
that are not driven exclusively by the actual sensory experience of
pain and can exert a significant impact on the level of function and
pain tolerance (Feuerstein & Beattie, 1995; Vlaeyen et al., 1999;
Vlaeyen & Linton, 2000). People are motivated to avoid and
escape from unpleasant consequences; they learn that avoidance of
situations and activities in which they have experienced acute
episodes of pain will reduce the likelihood of reexperiencing pain
or causing further physical damage. They may become hypervigi-
lant to their environment as a way of preventing the occurrence of
pain.
Investigators (e.g., Lethem, Slade, Troup, & Bentley, 1983;
Vlaeyen, Kole-Snijders, Rotteveel, et al., 1995) have suggested
that fear of pain, driven by the anticipation of pain and not by the
sensory experience of pain itself, produce strong negative rein-
forcement for the persistence of avoidance behavior and the puta-
tive functional disability in pain patients. Avoidance behavior is
reinforced in the short term through the reduction of suffering
associated with noxious stimulation (McCracken, Gross, Sorg, &
Edmands, 1993). Avoidance, however, can be a maladaptive re-
sponse if it persists and leads to increased fear, limited activity,
and other physical and psychological consequences that contribute
to disability and persistence of pain.
Studies have demonstrated that fear of movement and fear of
(re)injury are better predictors of functional limitations than bio-
medical parameters or even pain severity and duration (e.g., Crom-
bez, Vlaeyen, & Heuts, 1999; Turk, Robinson, & Burwinkle, 2004;
Vlaeyen, Kole-Snijders, Rotteveel, et al., 1995). For example,
Crombez et al. (1999) showed that pain-related fear was the best
predictor of behavioral performance in trunk-extension, flexion,
and weight-lifting tasks, even after the effects of pain intensity are
partialled out. Moreover, Vlaeyen, Kole-Snijders, Rotteveel, Rue-
599
BIOPSYCHOSOCIAL APPROACH TO CHRONIC PAIN
sink, and Heuts (1995) found that fear of movement and (re)injury
was the best predictor of self-reported disability among chronic
back-pain patients and that physiological sensory perception of
pain and biomedical findings did not add any predictive value. The
importance of fear of activity appears to generalize to daily activ-
ities as well as the clinical experimental context. Approximately
two thirds of chronic nonspecific low back-pain sufferers avoid
back straining activities because of fear of (re)injury (Crombez et
al., 1999). For example, fear-avoidance beliefs about physical
demands of a job are strongly related to disability and work lost
during the previous year, even more so than pain severity or other
pain variables (Asmundson, Norton, & Norton, 1999; Vlaeyen &
Crombez, 1999; Vlaeyen, Kole-Snijders, Boeren, & van Eek,
1995; Vlaeyen, Kole-Snijders, Rotteveel, et al., 1995). Interest-
ingly, reduction in pain-related anxiety predicts improvement in
functioning, affective distress, pain, and pain-related interference
with activity (McCracken & Gross, 1998). Clearly, fear, pain-
related anxiety, and concerns about harm-avoidance all play im-
portant roles in chronic pain and need to be assessed and addressed
in treatment.
Pain-related fear and concerns about harm avoidance all appear
to exacerbate symptoms (Vlaeyen, Kole-Snijders, Boeren, & van
Eek, 1995). Anxiety is an affective state that is greatly influenced
by appraisal processes