Bnuih Mtthcal BidUlm (1991) Vol. 47, No 3, pp.

631-643
© The Bntilh Council 1991

Neuropathic pain and

injured nerve: Central
mechanisms
PDWaU
Department of Anatomy & Developmental Biology, University College London,
London, UK

A satisfactory explanation of neuropathic pain must

include mechanisms capable of generating three types of
pain: ongoing, episodic and allodynic. It must explain why
many such pains develop very soon after injury while
others occur after long delays. It must take into account
the many painless neuropathies and the unpredictable
relationship of the pain to the pathology in the painful
neuropathies. While these diseases clearly start in the
periphery and peripheral changes must contribute to the
pain, there are also three types of central change. First
changes in the afferent impulse barrage can induce long
term shifts of central synaptic excitability. Second, changes
of the chemical substances transported from the periphery
to the cord produce alterations of cord cell excitability.
Third, central control mechanisms can change into a
pathological state permitting hyperexcitability. The
combined peripheral and central pathology offers more
than explanation since each factor could be a target for
prevention as well as cure.

If the classical specificity theory of pain mechanisms was correct,

this chapter would be very short because there are no relevant
central mechanisms in that scheme. Classically the sensation of
pain was considered as being created by the presence of nerve
impulses in nociceptors which are conducted to the spinal cord
over specialized fibres which excite nociceptive specific cells which
in turn project to the pain centre by way of the spinothalamic
tract. Since these central mechanisms were considered no more
632 PAIN

than a reliable transmission relay system, central mechanisms were

simple slaves to the information generated in peripheral nerve
fibres. If patients were sufficiently injudicious as to report painful
sensations which appeared not to match the activity of peripheral
nerves they were assigned to a mental limbo reserved for neurotics,
hypochondriacs, or hallucinators. The specificity theory is not cor-
rect however and it is now widely appreciated that central mechan-
isms contribute to the painful consequences of diseases which are
clearly initially limited to the periphery.
In considering the origin of these pains in a more subtle way
than implied by the classical specificity theory, it will be necessary
to move beyond the presence or absence of pain. Neuropathic
pains have six characteristics which require an explanation in the
periphery or in the central nervous system or by a combination of
both. The characteristics of neuropathic pains are:
1. Some pains are ongoing and little influenced by peripheral
stimuli.
2. Some pains are lancinating spontaneous stabs.
3. Some pains are provoked by normally innocuous stimuli
(allodynia).
4. All pains except trigeminal neuralgia are associated widi
hypaesthesia.
5. Provoked pains require summation by prolongation, movement
and repetition of the stimulus.
6. Provoked pains frequendy demonstrate a delay and build up
after the beginning of the stimulus.
7. Some pain states appear immediately after nerve injury, others
after delays.
All seven of these characteristics could be used to indicate the
likely existence of an abnormal central component in the pain.
Devor, in the preceding chapter, has described the nature of
ectopic sources of impulses which appear in damaged nerve. Each
example which he has given of a source of ongoing activity is
accompanied by an increase of mechanical sensitivity. Therefore
one would expect to be able to locate die peripheral source of an
ongoing pain by mechanical probing, die Tinel sign. This is fre-
quently not apparent. In postherpetic neuralgia, for example
where allodynia has not appeared or has been alleviated by TENS
or sympathectomy, a deep ongoing pain which is not exaggerated
by palpation is frequently present. Patients with postherpetic
neuralgia show a mixture of three sensory signs, allodynia, anaes-
NEUROPATHIC PAIN AND INJURED NERVE: CENTRAL MECHANISMS 633

thesia and ongoing deep pain. The latter is the particular charac-
teristic of patients with brachial plexus avulsions where their lesion
lies in roots central to the dorsal root ganglia. The pain of such
patients can only have a central origin since the input from the
periphery has been entirely destroyed. The storms of lancinating
stabs of pain from which many amputees suffer can similarly not
be easily attributed to the periphery since they can rarely be pro-
voked and no reports exist of such synchronous outbursts in ani-
mal model ectopic sources. The obvious explanation for allodynia
is that the disease has sensitized previously high threshold periph-
eral fibres. However, as reviewed by Campbell et al.,1 much recent
research points to a central component in which normal low
threshold sensory afferents excite hyperexcitable spinal cord cells
as described by Woolf (this issue). The most obvious examples of
pain produced by gentle stimulation of normal afferents are the
allodynic areas always found on the stumps of amputees at a dis-
tance from the area of damage created by the injury or the sur-
gery.2 The next three characteristics, hypaesthesia with allodynia
associated with summation and delay were first emphasized by
Noordenbos.3 Here pain is frequently not evoked by a single brief
innocuous stimulus. However, if a light mechanical stimulus is
repeated or a mild thermal stimulus is maintained, the patient
often escalates from no response to feeling a minor sensation to
excrutiating pain taking tens of seconds to reach maximum quite
unlike normal responses. As he pointed out, it would be extremely
difficult to explain this syndrome by pathological peripheral
characteristics and he proposed central disinhibition as a likely
alternative. Finally a severe limitation on possible mechanisms is
provided by the observation that neuropathic pain states may have
an extremely rapid onset. Obviously in certain painful neuro-
pathies where the disease is of insidious onset such as diabetes or
alcoholic neuropathies, it is not possible to identify the time of
disease onset and of pain onset. However, accidental or surgical
nerve damage is frequently followed by a very rapid and abrupt
onset of severe pain often apparent after amputation or thora-
cotomy as the patient recovers from the anaesthetic.4 It is true
that there is a subsequent slow evolution of the nature and distri-
bution of the pain but since the major problem is immediately
apparent and since no animal models of peripheral change have
such rapid onset, our attention is again directed to possible central
explanations.
In addition to incorporating the nature of the symptomatology
634 PAIN

in any satisfactory explanation, it is equally necessary to include

an explanation of the absence of pain in certain nerve lesions such
as:
Analgesic polyneuropathy5
Tangier disease6
Friedrich's ataxia7
Chronic renal failure8
Crush lesions9
A possible explanation for the first three is that they are in fact
congenital disorders and that their tendency to produce pain by
deafferentation may be overwhelmed by the restorative abilities of
the developing nervous system. The absence of pain in the massive
destruction of peripheral nerves in uraemic patients may be
excused by the obvious existence of a general systemic disorder
which could, without any evidence, prohibit the development of
hyperexcitable tissue either in the periphery or centrally. The lack
of pain following crush lesions of peripheral nerves is agreed by
all, with such patients suffering paraesthesias at worst. Clearly this
fact must be incorporated in any satisfactory explanation. Perhaps
the most surprising pain free neuropathy is so common that it is
never mentioned, tooth extraction. Here a branch of the fifth nerve
is crudely avulsed and left to lie untended in the debris of the
frequently infected tooth socket. Obviously chronic complaints
must be very rare indeed or diis procedure would not have per-
sisted with such abandon over the centuries. Only rare or very
rare or non-sequitur sequalae are mentioned, phantom tooth
pain10 dry socket syndrome, a possible causalgia like phenomenon
normally attributed to infection11 and atypical facial neuralgia.12
Many assign the latter diagnosis to a psychiatric category as is
apparent by such sentences as 'almost all are young women with
significant psychopathology'.13 The evident absence of chronic
pain in the vast majority of those who suffer tooth extraction is
particularly surprising since the trigeminal system has a highly
specific form of painful fragility which results in trigeminal neural-
gia, a condition which has no equivalent in other segments of the
body. Furthermore 15% of cases of postherpetic neuralgia suffer
pain in the trigeminal area and facial trauma with nerve damage
may result in chronic neuropathic pain.
Clearly the existence of pain free neuropathies as well as painful
ones should provide a clue as to the mechanism of the pain. In
addition we are faced with a similar challenging fact that chronic
NEUROPATHIC PAIN AND INJURED NERVE: CENTRAL MECHANISMS 635

pain is never present in 100% of the cases with the pathology to

which the pain is commonly attributed. This variable relation of
pathology with pain is not a monopoly of neuropathies since it
plagues an understanding of the pain in much more common dis-
eases such as cardiac ischaemia and osteoarthritis. Amputation is
an example of major nerve damage with painful consequences
occurring with great severity in only a fraction of the cases.14 In
our series of military casualties seen within months of their injury
and again 10-15 years later, 67% were in the same pain on both
occasions.2 These fractions are explained away without evidence
as attributable either to the variations in the nature of the lesion
or more commonly are attributed to central factors particularly
attitude.15 The variation of the peripheral damage cannot be so
easily used as an explanation where the results of the same sur-
geons' elective leg amputations were examined after 6 and 12
months with 50% of the patients in severe pain.16 If the cause was
the precise nature of the nerve lesions then either none or all of
these patients should have been in pain. A genetic predisposition
is proposed as a possible explanation based on the demonstration
of a single recessive gene involved in the reaction of rats to sciatic
section.17 However, I have urged great caution in the premature
attribution of single peripheral or central factors as the explanation
of the puzzling variability of the relation of pain to nerve damage.18
A major reason for this caution is a study of patients with intrac-
table chronic causalgia following partial nerve lesions.19 These
patients had the site of the original partial nerve damage recon-
structed by total nerve resection of the damaged area followed by
cable grafting of the sural nerve across the gap of the excised nerve.
All patients successfully regenerated their nerves across the graft
but relapsed into the precise state and location of the pain they
had experienced before the graft. The relevance of that finding to
this chapter is two fold. These patients were not peculiarly suscep-
tible to nerve lesions since the sural lesions produced no sequelae.
Second since the surgeons cut the entire nerve but the symptoms
recurred only in that fraction of the nerves' territory involved in
the original trauma, there must have been something special about
that incident which originated from those particular damaged
nerves, but was maintained even when the original area of damage
was excised. This observation among many drives one to seek
some mechanism initiated by local damage but located central to
the lesion.
In seeking peripheral and central mechanisms responsible for
636 PAIN

neuropathic pain, we must incorporate mechanisms that are cap-

able of generating:
A. Three types of pain: (1) ongoing; (2) episodic; and (3) provoked
with the spatial and temporal characteristics of allodynia.
B. Pains associated with some types of nerve lesion and not with
others.
C. Pains associated with only a fraction of apparently identical
lesions.
Three quite separate classes of central mechanism must be con-
sidered:
1. Central changes generated by arriving nerve impulses or their
absence.
2. Central changes generated by substances transported from the
periphery or their absence.
3. Central changes generated by instability of central control
mechanisms.
Just as central and peripheral explanations are not contradictory
and mutually exclusive these three different central mechanisms
could all be operative. It is to be hoped that one mechanism will
predominate in a particular patient at a particular time since ther-
apy would thereby be easier.

AFFERENT IMPULSE INDUCED CENTRAL CHANGES

Impulse triggered central hyperexcitability
We now know, thanks to the work of Woolf and his collaborators
as described in this issue, that an afferent barrage, particularly in
the unmyelinated afferents originating from deep tissue, triggers
a long lasting change of the excitability of dorsal horn cells. This
increased excitability is provoked by the arriving impulses but
maintained by central mechanisms. It may be that die trigger is
the emission of peptides from the terminals of unmyelinated
afferents. The brief input is followed by a series of prolonged
alterations in die dorsal horn cells in which the membrane and
chemistry of the cells change. These changes are now the subject
of widespread research including the manipulation of NMDA
receptors, and activity and calcium dependent changes of enzyme
activity within the cell. The onset of such changes is strongly
influenced by NMDA receptor blockers and by narcotics. The
latency of onset occurs within seconds to minutes of the arriving
NEUROPATHIC PAIN AND INJURED NERVE: CENTRAL MECHANISMS 637

barrage. They are therefore candidates for explaining the rapidity

of onset of some of the pains with which we are concerned. The
duration of these central changes after an abrupt brief input volley
has so far been shown in animal models to last for hours. In order
to explain chronic pains triggered by a brief episode, it would be
necessary to demonstrate persistance of central hyperexcitability
for which there is so far no evidence. However, the animal exper-
imental test inputs and the observation periods have been restric-
ted in time and intensity. It is possible that the observed changes
may themselves set off more fundamental and irreversible changes
in the cells and this is now being investigated. It is also possible
that once triggered by an initial large input, the prolonged changes
might be maintained by a trickle of abnormal afferent inputs. This
would result in an abnormal central component maintained in part
by a continuous low level of pathological peripheral input.
If the classical picture of dorsal horn pain mechanisms was true,
an increase of excitability would not explain the origin of the
symptoms which interest us here. If pain was only to be attributed
to the firing of nociceptive specific cells, an increase of their excit-
ability would produce only hyperalgesia, that is to say an increase
of the amount of pain produced by normally painful stimuli and
would not produce allodynia. However, these nociceptive specific
cells are the minority of the cells which signal noxious events in
the periphery. By far the commonest cells respond not only to
injury but also to gentle stimuli and are called convergent or wide
dynamic range cells.20"22 In comparing the onset of aversive
behaviour in monkeys with the onset of firing of convergent versus
nociceptive specific cells, it appears that 'pain behaviour' is best
related to the response of the convergent cells.23 Furthermore,
Cook et al.,24 showed that under conditions of afferent induced
central hyperexcitability, nociceptive specific cells could convert
to the convergent variety. Taken together these results support the
suggestion that a hyperexcitable response of such cells could be
responsible for the allodynic aspects of neuropathic pain where
normally innocuous stimuli provoke pain. Furthermore, since
these cells gradually increase their output with repeated inputs,
the 'wind up' phenomenon,25 they could be the arbitrators of the
temporal and spatial summation which is a critical feature of the
pain evoked in the patients we consider here. As the excitability
of central cells rises so does their ongoing activity. Therefore a
sufficient increase of excitability produced by the mechanism dis-
cussed here and by others to be discussed could result in a high
638 PAIN

enough level of ongoing activity to form the basis of the continuous

ongoing pains.

Impulse triggered central degeneration

The basis of this possible mechanism depends on the extensive
work of Sugimoto most recently reported in Sugimoto et al.26
When nerves are cut a brief maximal frequency volley is generated
in all types of afferents, the injury discharge. It is probable that in
normal spinal cord, there are limiting mechanisms which prevent
the postsynaptic arrival of such massive volleys. These depend
partly on the impulse carrying ability of the terminal arborizations
of afferents and partly on the presence of inhibitory mechanisms,
both of which limit the maximal effective size of inputs. Sugim-
oto's original observations showed that central cells degenerated
following section of trigeminal afferents if the inhibitory mechan-
isms had been disabled by strychnine. Now it has been observed
that such degenerating neurons occur in the dorsal horn without
strychnine but in the presence of a particularly evocative partial
nerve lesion in rats developed by Bennett & Xie.27 This partial
nerve lesion produced by constriction results in a very prolonged
hyperpathic state. It is accompanied by a massive ongoing afferent
barrage originating from the dorsal root ganglia presumably by
the mechanism described in the chapter by Devor (this issue).
Furthermore, it is likely that such a lesion may simultaneously
abolish the normal presynaptic and postsynaptic inhibitions in
dorsal horn as was shown to occur after sciatic lesions.2829 There-
fore a combination of increased afferent barrage with decreased
central inhibition may permit such a catastrophic increase of syn-
aptic excitation that cells die by the now well established mechan-
ism of amino acid excitotoxicity. It could be that this effect is the
extreme end result of Woolf's hyperexcitability process described
in the section above. Induced degeneration would be one way in
which temporary hyperexcitability could be converted to perma-
nent and irreversible changes.

The absence of afferent impulses

Arriving sensory impulses over dorsal roots to the dorsal horn
produce both excitations and inhibitions. A reduction of sensory
input therefore may reduce both. This would result in a partial
disinhibition which is presumably the explanation for the immedi-
NEUROPATHIC PAIN AND INJURED NERVE: CENTRAL MECHANISMS 639

ate appearance of a phantom limb after plexus local anaesthesia.30

Local anaesthesia of peripheral nerves or epidurally or intrathec-
ally frequently unmasks pain in addition to the expected anaes-
thesia.31 Partial blockade was the basis of Noordenbos'3 theory of
hyperpathic states. Some support is provided by the success of
stimulation in normalizing sensation after partial nerve injuries32
and of TENS 33 where the intention is to generate an afferent
barrage which has been prevented by the disease. This section is
introduced to emphasize the fact that pain may be the consequence
of an increase of excitation or of a decrease of inhibition.

CENTRAL CHANGES INDUCED BY CHEMICAL

TRANSPORT
All nerve cells transport crucial chemicals for their survival from
the cell body to the far reaches of the cell and back from a distance
to the cell body. If an axon is cut in the periphery, the dorsal root
ganglia change chemically, physiologically and pharmacologically
as a consequence of the changed transport.34'35 Interruption of
impulse traffic alone does not produce these changes.36 They are
produced by a failure of normal transport of substances, such as
nerve growth factor,37 or by the leaking in of foreign substances
in the area of damage. Crushing a nerve does not produce nearly
such severe changes as cutting the nerve38 an observation which
is clearly evocative of the paradox that nerve section has much
more severe pain sequelae than crushing the nerve. The expla-
nation for this difference may be that the sprouts which grow from
crushed axons are rapidly in contact with Schwann cells which
provide a substitute supply of nerve growth factor whereas in the
neuroma sprouts growing from cut axons contact many unusual
cells. After the cutting of axons and the failure of regeneration,
the changes in dorsal root ganglia may be so severe that some cells
die.39 When rat sciatic is cut and ligated in the periphery, 20-30%
of the dorsal root ganglion die after a delay of 3-4 weeks. In this
way a peripheral lesion is slowly converted into a partial spinal
cord deafferentation by the degeneration of dorsal root afferents.
Short of frank degeneration, there are marked changes in the
chemistry of the central terminals of sensory axons which have
been cut rather than crushed in the periphery.35 These centrally
moving changes proceed trans-synaptically. The spinal cord cells
on which the peripherally damaged afferents end lose both pre-
and postsynaptic inhibitions, increase their excitability, expand
640 PAIN

their receptive fields and become spontaneously active. Central

homeostatic mechanisms detect the failure of the input and raise
the excitability of central cells in an attempt to compensate for the
diminished input. These changes depend on transport mechanisms
which are very slow by comparison with the conduction of nerve
impulses. Even with the relatively short distances for transport
after sciatic nerve section in the rat, the central changes do not
begin for three days and reach their maximum after two weeks.
Once established they are permanent. Therefore these mechanisms
cannot be involved in the rapid onset phase of some neuropathic
problems but they could play a part in the delayed onset pains and
in the slow evolution and extension which is characteristic of both
rapid and slow onset neuropathic pain.

CHANGES IN CENTRAL CONTROL MECHANISMS

Up to this point we have discussed pathological pain mechanisms
as though they were a straight line of relay stations from input to
output. Furthermore it has been proposed that there are control
mechanisms set around each relay. Lastly it has been shown that
a centrally moving wave originates from the point of damage and
changes excitability successively at one station after another. This
is a conceptually simple plan even though the details are quite
complicated. However, the central nervous system is more than a
collection of input-output lines. It certainly incorporates at a mini-
mum, feedback systems capable not only of gain control but also
of filtering and selecting the type of information flowing over the
input-output lines.
An example of such a system highly relevant to neuropathic pain
is mainly due to the work of McMahon.40 The fine A delta and C
afferents particularly concerned with injury detection terminate
selectively on cells in the marginal layer, lamina I of the dorsal
horn. These cells project mainly in the controlateral dorsolateral
white matter of the spinal cord to terminate in the caudal midbrain
with other targets in medulla and thalamus. The midbrain area is
the indirect origin of powerful descending systems which return
to the dorsal horn by way of the ipsilateral dorsolateral white
matter. The effect in the cord is to facilitate the lamina I system,
a positive feedback, and to inhibit the deep cells, negative
feedback.
On any simple theory, this cord-brainstem-cord system would
be expected to affect pain. Therefore it has been extensively inves-
NEUROPATHIC PAIN AND INJURED NERVE: CENTRAL MECHANISMS 641

tigated by many groups who have completely failed to show any

effect on responses to acute noxious mechanical, thermal or chemi-
cal stimuli. A possible solution to this paradox is contained in the
discovery that the rat behavioural response to peripheral nerve
lesions, autotomy, is grossly exaggerated by lesions placed along
the ascending or descending loops of this circuit.41 Here we see
not the expected analgesia but an increased response. In addition
further work has shown that the expected decrease is indeed pro-
duced by ventral lateral white matter lesions in contrast to the
dorsolateral lesions which increase the response.42
The existence of such long range dynamic loops may have con-
siderable relevance to neuropathic pain in two ways. Their overall
activity is likely to be set not only by the input but by the atten-
tional set of the CNS as a whole. They are more than simple
automatic volume amplification limiting feedback controls. There-
fore on occasions, they could permit penetrance of pathological
inputs and on other occasions they could forbid it. They are conse-
quently a possible source of the observed variance in which the
CNS has options for its reaction to inputs and may on occasions
be caught unaware of a catastrophic event. The second relevance
is that such circuits are famously capable of instability, the topic
of Wiener's 'Cybernetics'. They have a limited normal operating
range restricted by their power and temporal response character-
istics. When pushed outside this range, they oscillate or flip to a
fixed extreme from which they cannot restore normality. They
therefore provide a possible mechanism for the observed storms
of episodic pain which are such a puzzling feature of some neuro-
pathic pains. This likelihood is increased for the lamina I-brain-
stem-cord loop by the finding that the circuit has a long time
constant being unable to respond to acute brief inputs.

REFERENCES
1 Campbell JN, Raja SN, Cohen RH, Manning DC, Khan AA, Mayer RA.
Peripheral neural mechanism of nociception. In: Melzack R, Wall PD, eds.
Textbook of Pain, 2nd edn. Edinburgh: Churchill Livingstone, 1989
2 Carlen PL, Wall PD, Nadvoma MD, Steinbach T. Phantom limbs and related
phenomena in recent traumatic amputations. Neurology 1978; 28: 211—217
3 Noordenbos W. Pain. Amsterdam: Elscvier, 1959
4 Tasker RR, Dostrovsky JO. Deafferentation and central pain. In: Melzack R,
Wall PD, eds. Textbook of Pain, 2nd edn. Edinburgh: Churchill Livingstone,
1989
5 Sweet W. Animal models of chronic pain. Pain 1981; 10: 275-295
6 Pleasure DE. Tangier Disease. In: Dyck PJ et al. eds. Peripheral Neuropathy,
Philadelphia: Saunders, 1975
642 PAIN

7 Dyck PJ, Thomas PK, Lambert EH, eds. Peripheral Neuropathy. Philadelphia:
Saunders, 1974
8 Thomas PK, Hollinrake K, Lascelles RG. The polyneuropathy of chronic renal
failure. Brain 1971; 94: 761-780
9 Scadding JW. Peripheral neuropathies. In: Melzack R, Wall PD, eds. Textbook
of Pain, 2nd edn. Edinburgh: Churchill Livingstone, 1989
10 Marbach JJ. Phantom tooth pain. J Endodontics 1978; 4: 362-370
11 Sharav Y. Orofacial pain. In: Melzack R, Wall PD, eds. Textbook of Pain, 2nd
cdn. Edinburgh: Churchill Livingstone, 1989
12 Loeser J. Atypical facial pain. In: Bonica JJ, ed. The Management of Pain, 2nd
edn. Philadelphia: Lea & Febiger, 1990
13 Weddington WN, Blazer D. Atypical facial pain and trigeminal neuralgia.
Psychosomatics 1979; 20: 348-359
14 Jensen TS, Rasmussen P. Pain after amputation. In: Melzack R, Wall PD, eds.
Textbook of Pain, 2nd edn. Churchill Livingstone, Edinburgh, 1989
15 Sherman RA, Sherman CJ. A comparison of phantom sensations among ampu-
tees. Pain 1985; 21: 91-97
16 Bach S, Noreng MF; Tjellden NU. Phantom limb pains in amputees. Pain
1988; 33: 297-301
17 Devor M, Raber P. Heritability of symptoms in an experimental animal model
of neuropathic pain. Pain 1990; 42: 51-68
18 Wall PD. Editorial. A genetic factor in the reaction of rats to peripheral nerve
injury. Pain 1990; 42: 49-50
19 Noordenbos W, Wall PD. Implications of the failure of nerve resection and
graft to cure chronic pain produced by nerve lesions. J Neurol Neurosurg
Psychiatr 1981; 44: 1068-1073
20 Wall PD. Cord cells responding to touch damage and temperature of skin. J
Neurophysiol 1960; 23: 197-210
21 Mendell LM, Wall PD. Responses of single dorsal cord cells to peripheral
cutaneous unmyelinated fibres. Nature 1965; 206: 97-99
22 Hillman P, Wall PD. Inhibitory and excitatory factors controlling lamina 5
cells. Exp Brain Res 1969; 9: 284-306
23 Dubner R. Specialization in nociceptive pathways. Adv Pain Res Ther 1985;
9: 111-138
24 Cook AJ, Woolf CJ, Wall PD, McMahon SB. Expansion of cutaneous receptive
fields of dorsal horn neurones following C-primary afferent fibre inputs. Nature
1987; 325: 151-153
25 Mendell LM. Physiological properties of unmyelinated fibre projections to the
spinal cord. Exp Neurol 1966; 16: 316-332
26 Sugimoto T, Bennett GJ, Kajander KC. Transynaptic degeneration in the
superficial dorsal horn. Pain 1990; 42: 205-214
27 Bennett GJ, Xie YK. A peripheral mononeuropathy in rat. Pain 1988; 33:
87-108
28 Wall PD, Devor M. The effect on peripheral nerve injury on dorsal root poten-
tials and on transmission of afferent signals into the spinal cord. Brain Res
1981; 209: 95-111
29 Woolf CJ, Wall PD. Chronic peripheral nerve section diminishes the primary
afferent A-fibre mediated inhibition of rat dorsal horn neurones. Brain Res
1982; 242: 77-85
30 Bromagc PR, Melzack R. Phantom limbs and the body scheme. Canad Anacsth
Soc J 1974; 21: 267-274
31 Cousins MJ, Bridenbaugh PO. Neural Blockade, 2nd edn. Philadelphia: Lip-
pincott, 1988
32 Lindblom U, Meyerson BA. Influence on touch, vibration and cutaneous pain
of dorsal column stimulation. Pain 1975; 1: 257-270
NEUROPATHIC PAIN AND INJURED NERVE: CENTRAL MECHANISMS 643

33 Wall PD, Sweet WH. Temporary abolition of pain in man. Science 1967; 155:
108-109
34 Devor M. The pathophysiology ofg damaged peripheral nerve. In: Melzack R,
Wall PD, eds. Textbook of Pain, 2nd edn. Edinburgh: Churchill Livingstone,
1989
35 Fitzgerald M. The course and termination of primary afferent fibres. In: Melz-
ack R, Wall PD, eds. Textbook of Pain, 2nd edn. Edinburgh: Churchill Liv-
ingstone, 1989
36 Wall PD, Mills R, Fitzgerald M, Gibson SJ. Chronic blockade of sciatic nerve
transmission by tetrodotoxin does not produce central changes in the dorsal
horn of the spinal cord of the rat. Neurosci Lett 1982; 30: 315-320
37 Fitzgerald M, Wall PD, Gocdert M, Emson PC. Nerve growth factor counter-
acts the neurophysiological and neurochemical effects of chronic sciatic nerve
injury. Brain Res 1985; 332: 131-141
38 Devor M, Wall PD. Plasticity in the spinal cord sensory map following periph-
eral nerve injury in rats. J Neurosci 1981; 1 (No. 7): 679-684
39 Ygge J. Intercostal nerve transection and its effect on the dorsal root ganglion.
J Comp Neurol 1989; 279: 199-211
40 McMahon SB, Wall PD. The significance of plastic changes in lamina I sys-
tems. In: Cervero F, Bennett GJ, Headley PM, eds. Processing of Sensory
Information in the Superficial Dorsal Horn of the Spinal Cord. New York:
Plenum Press, 1989: pp. 249-272
41 Wall PD, Bery J, Saade NE. Effects of lesions to rat spinal cord lamina I cell
projection pathways on reactions to acute and chronic noxious stimuli. Pain
1988; 35: 327-339
42 Saade NE, Atweh SF, Jabbur SJ, Wall PD. Effects of lesions in anterolateral
columns and dorsolateral funiculi on autotomy and reactions to acute noxious
stimuli in rats. Pain 1990; 42: 313-321