PAIN
Department of Oral and Cruniofacial Biological Sciences. Univrrsity of Marylnnd Dental School, Baltimore, MD 21201. USA
Kewvordst
E-mail address:
0304-3959/99/$20.00 0 1999

s.46 R. Duhnrr. K. Ren /Pain Supplement 6 (I 999) S4.C25.3
normal function of the brain. Evidence for a pain descending
facilitatory system came later (see below).
The above concepts, the varying relationship between
tissue damage and pain sensation, the presence of endogen-
ous descending pain modulatory mechanisms, and the
recognition that the response to tissue damage includes
cognitive, emotional, attentional and motoric consequences,
form the basis of present day hypotheses of forebrain modu-
lation of the pain experience (Casey, 1999). The large
volume of the human forebrain in relation to that of the
spinal cord (and the corresponding relationship between
corticospinal and spinothalamic pathways) suggests that
descending direct corticospinal modulatory infuences may
assume greater importance in humans than in other animals.
There is also now considerable evidence that forebrain
structures including the medial preoptic area, the amygdala,
the insular cortex and the anterior cingulate cortex, modu-
late activity at brain stem levels through projections to the
midbrain periaqueductal gray (PAG) and the rostra1 ventro-
medial medulla (RVM) (Rizvi et al., 1991; Bandler and
Shipley, 1994) and that these brainstem structures project
to the spinal and medullary dorsal horns (Basbaum and
Fields, 1984. for review).
The interest of one of us in descending modulation of
ascending sensory input was ignited by an early study in
the trigeminal principal sensory nucleus (Dubner, 1967). All
single neurons studied were localized to the rostra1 compo-
nent of the trigeminal brain stem nucleus and had restricted
receptive fields activated by hair movement, light touch or
pressure. Interestingly, over 20% of these cells were excited
by stimulation of somatosensory cortical areas I/II and over
50% were inhibited. Almost all cells that responded to cort-
cal stimulation also responded to visual stimuli. These find-
ings suggested that cerebral cortex modulation of
somatosensory input occurred at the first central relay
nucleus in the trigeminal system and that other sensory
modalities had access to these descending pathways. This
modulation did not involve nociceptive input since none of
the neurons studied responded to tissue-damaging stimuli.
Research on pain mechanisms followed this study, includ-
ing a year well spent with Pat Wall in London. The oppor-
tunity to work with Pat provided a new and important
perspective on the function of the nervous system that has
influenced all of my (R.D.) subsequent research.
The interest of the other one of us (K.R.) in descending
modulation of ascending sensory input began with an obser-
vation that responses of spinal dorsal horn neurons to
noxious stimuli were unmasked by a low dose of pentobar-
bital in intact cats (Collins and Ren, 1987). This effect is
likely due to a release of inhibitory systems that regulate
spinal dorsal horn activity, an idea that is consistent with the
pioneer work by Pat Wall (1967). He suggested that the
removal of descending modulatory systems by blocking of
the rostra1 spinal cord allows dorsal horn neurons to respond
to noxious stimuli that they would not have responded to at
all or only poorly before the block. Further studies were
devoted to a phenomenon that activation of visceral vagal
afferents would produce inhibition of spinothalamic neuro-
nal activity (Ammons et al., 1983) and behavioral antinoci-
ception (Randich et al., 1984). After a systematic analysis, it
became apparent that vagal afferent activation modulates
nociceptive activity via descending pain modulation circui-
try, including relays in the nucleus raphe magnus and locus
coeruleus/subcoeruleus (Ren et al., 1990). Under certain
circumstances, vagal afferent stimulation also produces
facilitation of the nociceptive tail flick reflex as well as
inhibition of dorsal horn nociceptive neuronal activity
(Ren et al., 1988; Ren et al., 1989). The neural mechanisms
involved in vagal afferent-produced facilitation and inhibi-
tion appear to be different. Forebrain input or modulation by
the forebrain of a brain stem site is required for nociceptive
facilitation, but not inhibition, by vagal afferents. Thus.
evidence that spinal nociceptive processing was subject to
bi-directional control from supraspinal sites was revealed in
these earlier studies as well as those of others (Fields, 1988;
McMahon and Wall, 1988).
The above studies demonstrated that neural networks
were available by which unnatural electrical stimulation
of descending or primary afferent input could modulate
the transmission of ascending signals. Missing was evidence
that these pathways could modulate natural behaviors in
humans and other animals. There is still a paucity of direct
evidence of such modulatory influences. Furthermore, most
previous studies have focused on responses to transient
noxious stimuli with little information on the influence of
descending modulation on behavioral responses to persis-
tent pain and hyperalgesia after tissue or nerve injury. We
will provide evidence below that 1) behavioral context
modulates neuronal activity in nociceptive and non-noci-
ceptive somatosensory pathways, supporting the hypothesis
that responses in these pathways are not immutable; 2)
descending modulation influences behavior and neuronal
activity at spinal cord levels after inflammation and persis-
tent pain; and 3) there are descending facilitatory as well as
inhibitory influences on behavior and spinal cord neuronal
activity that may impact on persistent pain particularly of
deep muscle and visceral origin.
2. Behavioral modulation of sensory processing
The approach we have adopted to study pain mechanisms
is twofold. First, develop an animal model which mimics
experimental or clinical pain and utilize quantitative inde-
pendent and dependent variables to demonstrate nocifensive
behavior. Second, study the nervous system in this model. In
earlier studies, we trained monkeys to detect or discriminate
transient noxious thermal stimulation and then recorded
single unit activity while the animals performed what is
commonly referred to as a delayed-response task (Evarts
et al., 1984). The monkey is given an instructional sensory
cue indicating that a response to a second, trigger stimulus