A brief history of the Australasian Neuroscience Society.

The collective efforts of Australasian neuroscientists over the past 50 years to forge a binational presence are reviewed in this article. The events in the 1970s leading to the formation of an informal Australian Neurosciences Society are discussed in the context of the international emergence of neuroscience as an interdisciplinary science. Thereafter, the establishment in 1980 of the Australian Neuroscience Society, subsequently renamed as the Australasian Neuroscience Society (ANS), is described. The achievements of ANS-including its active role in developing national, regional, and global cooperation to promote neuroscience-are chronicled over successive decades, followed by a discussion of the future challenges facing the society and its associated neuroscience institutions.

Transparency in Research involving Animals: The Basel Declaration and new principles for reporting research in BJP manuscripts.

This article discusses the background to the need for change in the reporting of experiments involving animals, including a report of a consensus meeting organised by the Basel Declaration Society and Understanding Animal Research UK that sought to Internationalise guidelines for reporting experiments involving animals. A commentary on the evolution of BJP's attempts to implement the ARRIVE guidelines and details of our new guidance for authors is published separately (McGrath, 2014). This is one of a series of editorials discussing updates to the BJP Instructions to Authors LINKED EDITORIALS: This Editorial is the first in a series. The other Editorials in this series will be published in the forthcoming issues. To view them, visit: http://onlinelibrary.wiley.com/journal/10.1111/(ISSN)1476-5381.

Inflammation in dorsal root ganglia after peripheral nerve injury: effects of the sympathetic innervation.

Following a peripheral nerve injury, a sterile inflammation develops in sympathetic and dorsal root ganglia (DRGs) with axons that project in the damaged nerve trunk. Macrophages and T-lymphocytes invade these ganglia where they are believed to release cytokines that lead to hyperexcitability and ectopic discharge, possibly contributing to neuropathic pain. Here, we examined the role of the sympathetic innervation in the inflammation of L5 DRGs of Wistar rats following transection of the sciatic nerve, comparing the effects of specific surgical interventions 10-14 days prior to the nerve lesion with those of chronic administration of adrenoceptor antagonists. Immunohistochemistry was used to define the invading immune cell populations 7 days after sciatic transection. Removal of sympathetic activity in the hind limb by transecting the preganglionic input to the relevant lumbar sympathetic ganglia (ipsi- or bilateral decentralization) or by ipsilateral removal of these ganglia with degeneration of postganglionic axons (denervation), caused less DRG inflammation than occurred after a sham sympathectomy. By contrast, denervation of the lymph node draining the lesion site potentiated T-cell influx. Systemic treatment with antagonists of α1-adrenoceptors (prazosin) or ß-adrenoceptors (propranolol) led to opposite but unexpected effects on infiltration of DRGs after sciatic transection. Prazosin potentiated the influx of macrophages and CD4(+) T-lymphocytes whereas propranolol tended to reduce immune cell invasion. These data are hard to reconcile with many in vitro studies in which catecholamines acting mainly via ß2-adrenoceptors have inhibited the activation and proliferation of immune cells following an inflammatory challenge.

Local and remote immune-mediated inflammation after mild peripheral nerve compression in rats.

After experimental nerve injuries that extensively disrupt axons, such as chronic constriction injury, immune cells invade the nerve, related dorsal root ganglia (DRGs), and spinal cord, leading to hyperexcitability, raised sensitivity, and pain. Entrapment neuropathies, such as carpal tunnel syndrome, involve minimal axon damage, but patients often report widespread symptoms. To understand the underlying pathology, a tube was placed around the sciatic nerve in 8-week-old rats, leading to progressive mild compression as the animals grew. Immunofluorescence was used to examine myelin and axonal integrity, glia, macrophages, and T lymphocytes in the nerve, L5 DRGs, and spinal cord after 12 weeks. Tubes that did not constrict the nerve when applied caused extensive and ongoing loss of myelin, together with compromise of small-, but not large-, diameter axons. Macrophages and T lymphocytes infiltrated the nerve and DRGs. Activated glia proliferated in DRGs but not in spinal cord. Histologic findings were supported by clinical hyperalgesia to blunt pressure and cold allodynia. Tubes that did not compress the nerve induced only minor local inflammation. Thus, progressive mild nerve compression resulted in chronic local and remote immune-mediated inflammation depending on the degree of compression. Such neuroinflammation may explain the widespread symptoms in patients with entrapment neuropathies.

Removal of half the sympathetic innervation does not reduce vasoconstrictor responses in rat tail artery.

Following reinnervation of denervated rat tail arteries, nerve-evoked contractions are at least as large as those evoked in normally innervated arteries despite a much lower nerve terminal density. Here nerve-evoked contractions have been investigated after transection of half the sympathetic innervation of normal tail arteries. After 1 week, the noradrenergic plexus 50-70 mm along the tail was about half as dense as control. Excitatory junction potentials recorded in smooth muscle cells of arterial segments isolated in vitro were half their normal amplitude. Surprisingly, nerve-evoked contractions of isometrically mounted segments were not reduced in amplitude, as was also the case after only 3 days. After 1 week, enhancement of nerve-evoked contractions by blocking either neuronal re-uptake of noradrenaline with desmethylimipramine or prejunctional α2-adrenoceptors with idazoxan was similar to control, suggesting that these mechanisms are matched to the number of innervating axons. The relative contribution of postjunctional α2-adrenoceptors to contractions evoked by long trains of stimuli was enhanced but that of α1-adrenoceptors was unchanged. Transiently, sensitivity to the α1-adrenoceptor agonist phenylephrine was slightly increased. After 7 weeks, amplitudes of nerve-evoked contractions remained similar to control, and sensitivity to phenylephrine had recovered but that to the α2-adrenoceptor agonist clonidine was slightly raised. The normal amplitude of nerve-evoked contractions after partial denervation is only partly explained by the greater contribution of α2-adrenoceptors. While the post-receptor mechanisms activated by nerve-released transmitter may be modified to amplify the contractions after partial denervation, our findings suggest that these mechanisms are normally saturated, at least in this artery.

Nerve-evoked constriction of rat tail veins is potentiated and venous diameter is reduced after chronic spinal cord transection.

Despite reduced sympathetic activity below the level of a spinal cord injury (SCI), venoconstriction during autonomic dysreflexia increases venous return to the heart. Here, contractions of isometrically mounted tail veins from rats with spinal transection at T4 performed 8 - 10 weeks earlier are compared with those from sham-operated rats. After SCI, lumen diameter was reduced by ∼30% and the contractions evoked by electrical stimulation of the perivascular axons were larger than control. This augmentation of neurovascular transmission was not associated with enhanced sensitivity to α-adrenoceptor agonists or to adenosine-5'-triphosphate (ATP) although contractions to depolarization with K(+) were larger after SCI. The percentage reduction in nerve-evoked contraction after SCI produced by the α(1)-adrenoceptor antagonist prazosin (10 nM) was unchanged but that by the α(2)-adrenoceptor antagonist rauwolscine (0.1 µM) was reduced. The relative contribution of P2-purinoceptors to nerve-evoked contractions after α-adrenoceptor blockade, revealed by adding suramin (0.1 mM), was unchanged. The greater depolarization-induced contraction and the reduced contribution of α(2)-adrenoceptors to nerve-evoked contraction suggest that changes in the venous smooth muscle underlie the potentiation of neurovascular transmission after SCI. Furthermore, the smaller lumen diameter after SCI will increase the pressure that the veins exert on the luminal contents when they are neurally activated.

Slow and incomplete sympathetic reinnervation of rat tail artery restores the amplitude of nerve-evoked contractions provided a perivascular plexus is present.

We have investigated the recovery of sympathetic control following reinnervation of denervated rat tail arteries by relating the reappearance of noradrenergic terminals to the amplitude of nerve-evoked contractions of isometrically mounted artery segments in vitro. We have also assessed reactivity to vasoconstrictor agonists. Freezing the collector nerves near the base of the tail in adult rats denervated the artery from ∼40 mm along the tail. Restoration of the perivascular plexus declined along the length of the tail, remaining incomplete for >6 mo. After 4 mo, nerve-evoked contractions were prolonged but of comparable amplitude to control at ∼60 mm along the tail; they were smaller at ∼110 mm. At ∼60 mm, facilitation of contractions to short trains of stimuli by the norepinephrine transporter blocker, desmethylimipramine, and by the α2-adrenoceptor antagonist, idazoxan, was reduced in reinnervated arteries. Blockade of nerve-evoked contractions by the α1-adrenoceptor antagonist, prazosin, was less and by idazoxan greater than control after 8 wk but similar to control after 16 wk. Sensitivity of reinnervated arteries to the α1-adrenoceptor agonist, phenylephrine, was raised in the absence but not in the presence of desmethylimipramine. Sensitivity to the α2-adrenoceptor agonist, clonidine, was maintained in 16-wk reinnervated arteries when it had declined in controls. Thus regenerating sympathetic axons have a limited capacity to reinnervate the rat tail artery, but nerve-evoked contractions match control once a relatively sparse perivascular plexus is reestablished. Functional recovery involves prolongation of contractions and deficits in both clearance of released norepinephrine and autoinhibition of norepinephrine release.

Sympathetic vasoconstriction is potentiated in arteries caudal but not rostral to a spinal cord transection in rats.

Sympathetic nerve-mediated contractions of mesenteric and tail arteries controlled by preganglionic neurones decentralized by a spinal cord injury (SCI) are potentiated, and likely contribute to autonomic dysreflexia. However, reactivity to the α(1)-adrenoceptor agonist phenylephrine has been reported to be enhanced in vascular beds controlled by preganglionic neurones lying both rostral and caudal to an SCI in vivo. Here responses of isometrically-mounted median and saphenous arteries isolated from rats 2 and 8 weeks after transection of the T4 spinal cord have been compared with those from sham-operated rats. After SCI, contractions of median arteries to perivascular nerve stimulation, to α-adrenoceptor agonists (phenylephrine and clonidine), to the P2X-purinoceptor agonist α,ß-methylene ATP, and to 60 mM K(+) were unchanged. Blockade of nerve-evoked contractions by α-adrenoceptor antagonists (prazosin and idazoxan) was not affected by SCI in either the median or saphenous arteries. In contrast, at 2 and 8 weeks after SCI, nerve-evoked contractions of saphenous arteries were potentiated. Saphenous arteries were less sensitive to phenylephrine 8 weeks after SCI, and their contractions to 60 mM K(+) were reduced. However, the sensitivity of saphenous arteries to clonidine was unchanged by SCI. Eight weeks after SCI, the reactivity of saphenous arteries to α,ß-methylene ATP was unchanged, but the P2-antagonist suramin produced more blockade of nerve-evoked contractions. These findings demonstrate that neurovascular transmission is enhanced in arteries located caudal, but not rostral, to a spinal transection. In the saphenous artery, the most likely explanation seems to be an increase in neurotransmitter release, as may occur in other inactive sympathetic pathways caudal to the lesion.

Transient supersensitivity to alpha-adrenoceptor agonists, and distinct hyper-reactivity to vasopressin and angiotensin II after denervation of rat tail artery.

BACKGROUND AND PURPOSE: Vascular 'denervation' hyper-reactivity has generally been investigated 1-2 weeks after administration of chemicals that temporarily prevent transmitter release, but do not necessarily inactivate the neuronal noradrenaline transporters (NETs). We have investigated the reactivity of rat tail arteries over longer periods after removing the terminals by surgical denervation. EXPERIMENTAL APPROACH: Two and 7 weeks after denervation, myography was used to assess contractions of isolated arterial segments to phenylephrine, methoxamine, clonidine, vasopressin and angiotensin II (AII). Denervation was confirmed by lack of tyrosine hydroxylase immunoreactive nerve terminals. KEY RESULTS: The NET inhibitor, desmethylimipramine, increased the pEC(50) for phenylephrine in control, but not denervated arteries after both 2 and 7 weeks. Relative to controls, pEC(50)s for phenylephrine (with desmethylimipramine), methoxamine, clonidine and vasopressin were increased at 2 but not 7 weeks after denervation. The pEC(50) for phenylephrine in the absence of desmethylimipramine was greater than control after both 2 and 7 weeks' denervation. The maximum contraction to vasopressin was larger than in controls at 2 but not 7 weeks after denervation, whereas contractions to AII were markedly enhanced at both time points. CONCLUSIONS AND IMPLICATIONS: Increased vascular reactivity to alpha(1)- and alpha(2)-adrenoceptor agonists, and vasopressin is transient following denervation. After 7 weeks, increased reactivity to phenylephrine can be entirely accounted for by the loss of NETs. Maintained supersensitivity to AII indicates that denervation differentially and selectively affects vascular reactivity to circulating vasoconstrictor agents. This might explain persistent vasoconstriction in denervated skin of humans after nerve injuries.

The effect of menthol on cold allodynia in patients with neuropathic pain.

OBJECTIVE: Cutaneous application of menthol in healthy subjects induces cold allodynia via sensitization of cold-sensitive nociceptors. We investigated the effects of menthol on preexisting cold allodynia in patients to test whether the allodynia was exacerbated. DESIGN: In eight neuropathic pain patients (six of peripheral, two of central origin), 40% menthol was applied topically to an area of preexisting cold allodynia. Mirror-image skin areas and aged-matched healthy subjects served as controls in patients with unilateral and bilateral neuropathic pain, respectively. Prior to and after menthol, cold pain thresholds were measured using a thermotest device. RESULTS: Menthol induced significant cold allodynia in control areas. However, in neuropathic areas, results were more heterogeneous. Overall, preexisting cold allodynia was not aggravated by topical menthol and was attenuated in 6/8 patients. CONCLUSIONS: These results suggest that, unlike in controls, menthol is not more hyperalgesic, but may be analgesic in some patients with peripheral and central neuropathic pain.

Distinct types of microglial activation in white and grey matter of rat lumbosacral cord after mid-thoracic spinal transection.

The inflammatory response has been characterized in the lumbosacral segments (L4-S1) of rats after spinal transection at T8. Immune cells were identified immunohistochemically using antibodies to complement type 3 receptor, CD11b (OX-42), the macrophage lysosomal antigen, CD68 (ED1), major histocompatibility complex class II (MHC II), and CD163 (ED2), a marker of perivascular cells. One week after cord transection, OX-42+ microglial density had nearly doubled. In the white matter, microglia became enlarged, often with retracted processes. In contrast, microglia in the grey matter remained ramified although nearly half of those lying medially contained clusters of ED1+ granules. After 8 weeks, ED1+ (+/-MHC II) macrophages were prominent in regions of Wallerian degeneration extending from dorsolateral to ventral funiculi. Microglial density remained raised in grey matter, particularly in the ventral horns of L4/5. Ramified microglia expressing MHC II+ (+/-ED1) extended from deep in the dorsal columns and around the central canal to the ventral columns. More ED2+ (+/-MHC II) perivascular and meningeal cells were recruited and expressed ED1. Thus, distinct from their conversion into macrophages in the white matter, the activation of ramified microglia after degeneration in the grey matter involves expression of ED1 without morphologic transformation.

Diversity of sympathetic vasoconstrictor pathways and their plasticity after spinal cord injury.

Sympathetic vasoconstrictor pathways pass through paravertebral ganglia carrying ongoing and reflex activity arising within the central nervous system to their vascular targets. The pattern of reflex activity is selective for particular vascular beds and appropriate for the physiological outcome (vasoconstriction or vasodilation). The preganglionic signals are distributed to most postganglionic neurones in ganglia via synapses that are always suprathreshold for action potential initiation (like skeletal neuromuscular junctions). Most postganglionic neurones receive only one of these "strong" inputs, other preganglionic connections being ineffective. Pre- and postganglionic neurones discharge normally at frequencies of 0.5-1 Hz and maximally in short bursts at <10 Hz. Animal experiments have revealed unexpected changes in these pathways following spinal cord injury. (1) After destruction of preganglionic neurones or axons, surviving terminals in ganglia sprout and rapidly re-establish strong connections, probably even to inappropriate postganglionic neurones. This could explain aberrant reflexes after spinal cord injury. (2) Cutaneous (tail) and splanchnic (mesenteric) arteries taken from below a spinal transection show dramatically enhanced responses in vitro to norepinephrine released from perivascular nerves. However the mechanisms that are modified differ between the two vessels, being mostly postjunctional in the tail artery and mostly prejunctional in the mesenteric artery. The changes are mimicked when postganglionic neurones are silenced by removal of their preganglionic input. Whether or not other arteries are also hyperresponsive to reflex activation, these observations suggest that the greatest contribution to raised peripheral resistance in autonomic dysreflexia follows the modifications of neurovascular transmission.

Immune cell involvement in dorsal root ganglia and spinal cord after chronic constriction or transection of the rat sciatic nerve.

Chronic constriction injury (CCI) of the sciatic nerve in rodents produces mechanical and thermal hyperalgesia and is a common model of neuropathic pain. Here we compare the inflammatory responses in L4/5 dorsal root ganglia (DRGs) and spinal segments after CCI with those after transection and ligation at the same site. Expression of ATF3 after one week implied that 75% of sensory and 100% of motor neurones had been axotomized after CCI. Macrophage invasion of DRGs and microglial and astrocytic activation in the spinal cord were qualitatively similar but quantitatively distinct between the lesions. The macrophage and glial reactions around neurone somata in DRGs and ventral horn were slightly greater after transection than CCI while, in the dorsal horn, microglial activation (using markers OX-42(for CD11b) and ED1(for CD68)) was greater after CCI. In DRGs, macrophages positive for OX-42(CD11b), CD4, MHC II and ED1(CD68) more frequently formed perineuronal rings beneath the glial sheath of ATF3+ medium to large neurone somata after CCI. There were more invading MHC II+ macrophages lacking OX-42(CD11b)/CD4/ED1(CD68) after transection. MHC I was expressed in DRGs and in spinal sciatic territories to a similar extent after both lesions. CD8+ T-lymphocytes aggregated to a greater extent both in DRGs and the dorsal horn after CCI, but in the ventral horn after transection. This occurred mainly by migration, additional T-cells being recruited only after CCI. Some of these were probably CD4+. It appears that inflammation of the peripheral nerve trunk after CCI triggers an adaptive immune response not seen after axotomy.

Enhanced neurally evoked responses and inhibition of norepinephrine reuptake in rat mesenteric arteries after spinal transection.

In patients with high thoracic spinal lesions that remove most of the central drive to splanchnic preganglionic neurons, visceral or nociceptive stimuli below the lesion can provoke large increases in blood pressure (autonomic dysreflexia). We have examined the effects of T4 spinal transection on isometric contractions of mesenteric arteries isolated from spinalized rats. Nerve-evoked contractions involved synergistic roles for norepinephrine and ATP. At 7 wk after spinal transection, responses to perivascular stimulation at 1-5 Hz were enhanced fivefold, whereas the alpha1-adrenoceptor antagonist prazosin (10 nM) produced a twofold larger reduction in contraction (to 20 pulses at 10 Hz) than in unoperated controls. In contrast, the reduction in nerve-evoked contractions by the P2-purinoceptor antagonist suramin (0.1 mM) and the responses to the P2-purinoceptor agonist alpha,beta-methylene ATP or to high K+ concentration did not greatly differ between groups, indicating that arteries from spinalized rats were not generally hyperreactive. Sensitivity to the alpha1-adrenoceptor agonist phenylephrine was enhanced in arteries from spinalized rats, and the difference from controls was abolished by the norepinephrine uptake blocker desmethylimipramine. Sensitivity to the alpha1-adrenoceptor agonist methoxamine, which is not a substrate for the neuronal norepinephrine transporter, was similar among the groups. Thus the increased neurally evoked response after spinal transection appeared to be due to a reduction in neuronal uptake of released norepinephrine, a mechanism that did not explain the enhanced response of tail arteries after spinal transection that we previously reported. The findings provide further support for potentiated neurovascular responses contributing to the genesis of autonomic dysreflexia.

Adaptations of peripheral vasoconstrictor pathways after spinal cord injury.

The consequences of spinal cord injury on the function of sympathetic pathways in the periphery have generally been ignored. We discuss two types of plasticity that follow disruption of sympathetic pathways in rats . The first relates to the partial denervation of sympathetic ganglia that would follow the loss of some preganglionic neurones. Sprouting of residual connections rapidly reinnervates many postganglionic neurones, restoring functional transmission within a few weeks, but other neurones may be permanently decentralized. Some of the new functional connections may generate inappropriate pathways leading to abnormal reflexes . The second type of plasticity concerns the markedly enhanced and prolonged contractile responses to nerve activity in arterial vessels to which ongoing sympathetic activity has been reduced or silenced following spinal cord transection or ganglion decentralization. In a cutaneous artery (the rat tail artery), the mechanisms underlying this arterial hyperreactivity differ from those in the splanchnic arteries (the rat mesenteric artery). In the former, hyperreactivity is mainly postjunctional but independent of changes in alpha1-adrenoceptor sensitivity, whereas the increased responsiveness in the latter vessels can be attributed to a greater responsiveness to alpha1-adrenoceptor activation. There are enough data from humans to suggest that both of these novel findings in experimental animals are likely to apply after spinal cord injury and contribute to autonomic dysreflexia .

Hyperinnervation of mesenteric arteries in spontaneously hypertensive rats by sympathetic but not primary afferent axons.

Hypertrophy of the perivascular plexus is thought to play a role in the development of hypertension in spontaneously hypertensive rats (SHR). However, it is not known whether the sympathetic varicosities are more numerous or larger, or form more neurovascular junctions. Further, a parallel hypertrophy of primary afferent terminals around the vessels might modulate any effects of hypertrophied sympathetic terminals. We have investigated the perivascular plexus around second-order mesenteric arteries of SHR and Wistar-Kyoto (WKY) rats by electron microscopy. Noradrenergic terminals were identified by the presence of small granular vesicles after chromaffin fixation, and substance P (SP+) afferent axons were identified by immunohistochemistry. The numbers of noradrenergic axon and varicosity profiles were higher (48 and 25%, respectively) in SHR than in WKY rats, and the majority lay closer to the medio-adventitial border. In contrast, there was no difference in the numbers of SP+ axons. Sympathetic and SP+ varicosities were indistinguishable in size, shape, vesicle content and mitochondrion content between each other and between the strains. However, both the number of neuromuscular junctions and the proportion of varicosities that formed them in SHR arteries were more than double those in WKY vessels. The data clearly show that hyperinnervation in SHR is specific for noradrenergic axons.

Chronic decentralization potentiates neurovascular transmission in the isolated rat tail artery, mimicking the effects of spinal transection.

Spinal cord transection produces a marked increase in the response of the isolated rat tail artery to sympathetic nerve stimulation, possibly as a result of a decrease in ongoing sympathetic activity. We have tested the effects of removing ongoing nerve activity on neurovascular transmission by cutting the preganglionic input to postganglionic neurones supplying the tail artery (decentralization). Isometric contractions to nerve stimulation were compared between decentralized arteries and those from age-matched and sham-operated controls. Nerve-evoked responses of decentralized arteries were much larger than those of control arteries at 2 and 7 weeks post operatively. The extent of blockade of nerve-evoked contraction by alpha-adrenoceptor antagonists prazosin (10 nM) or idazoxan (0.1 microM) was reduced. Decentralized arteries were transiently supersensitive to the alpha1-adrenoceptor agonist phenylephrine and the alpha2-adrenoceptor agonist clonidine; the unchanged sensitivity to methoxamine and phenylephrine after 2 weeks indicated no effect on the neuronal noradrenaline uptake transporter. Decentralized arteries were hypersensitive to alpha,beta methylene-ATP, but the P2-purinoceptor antagonist suramin (0.1 mM) did not reduce nerve-evoked contractions. Enlarged responses to 60 mM K+ after both 2 and 7 weeks were correlated with the response of the arteries to nerve stimulation, suggesting that increased postjunctional reactivity contributes to the enhanced contraction. Comparison between data from decentralized arteries and our previous data from spinalized animals showed that the two lesions similarly potentiate nerve-evoked contractions and have similar but not identical postjunctional effects. The enhanced vascular responses following a reduction in tonic nerve activity may contribute to the hypertensive episodes of autonomic dysreflexia in spinally injured patients.

Inflammation in sympathetic ganglia proximal to sciatic nerve transection in rats.

Comparison was made between recruitment of T-lymphocytes and macrophages into lumbar sympathetic ganglia (SGs) and dorsal root ganglia (DRGs) following sciatic nerve transection in rats. In both control and lesioned SGs, resident (ED2+) macrophages expressed less major histocompatibility complex class II (MHC II), but MHC II+ macrophage density was higher, than in equivalent DRGs. The influx of T-cells was larger and the influx and activation of macrophages were more sustained in SGs than in DRGs. Only two of the five subtypes of macrophage that invade lesioned DRGs were recruited to SGs. While some MHC II+ cells phagocytosed dead sympathetic neurones, most phagocytes in SGs lacked a macrophage marker. The different patterns of response between ganglia may provide clues about macrophage involvement in neuronal death and hyperexcitability after peripheral nerve lesions.