NERVE REGENERATION: BASIC AND APPLIED ASPECTS

R. Bruce Donolf

Dean, Professor of Oral and Maxillofacial Surgery, Harvard School of Dental Medicine, 188 Longwood Avenue, Boston, Massachusetts 02115

ABSTRACT: Increased knowledge is shedding new light on our understanding of central and peripheral nerve anatomy andmolecular biology and function. New tools and methods provide important methods for the study of the behavior of cells,axons, and receptors. This review discusses the current state of that knowledge, with particular regard to the efficacy of theSeddon classification of nerve injury. The correlation of that new information to damage and repair of the peripheral sensorynerve, especially the inferior alveolar and lingual nerves, serves to highlight the progress and problems that exist.

Keywords. Inferior alveolar nerve, lingual nerve, injury, repair.

Introduction

T he clinically sound and logical management of nerveinjuries is based on the classification of the injury.

Several classification schemes exist, the most commonof these being the Seddon and Sunderland classifica-tions. This paper will review current understanding ofnerve regeneration and attempt to correlate this knowl-edge with clinical management of injured peripheral sen-sory nerves. The inferior alveolar nerve (IAN) and lingualnerve will be highlighted. The most controversial issuesfacing the patient and the surgeon are whether an injuryrequires treatment and, if so, when. This review willexamine these issues through animal models, long-termclinical data for untreated injuries, and the current stateof electrophysiologic measurements. It is well to remem-ber that these classification schemes were devised forother peripheral nerves, and so application to a sensorynerve like the inferior alveolar or lingual may involvesome inaccuracies. More general discussions regardinginjury to the trigeminal nerve offer a better overview ofthe topic (Donoff and Colin, 1990).

Seddon described three types of nerve injury—neu-rapraxia, axonotmesis, and neurotmesis—based uponthe severity of tissue injury, prognosis for recovery, andtime for recovery (Seddon, 1943). Neurapraxia is a con-

duction block resulting from a mild insult to the nervetrunk. There is no axonal degeneration, and sensoryrecovery is complete and occurs in a matter of hours toseveral days. The sensory deficit is usually mild and char-acterized by a paresthesia, with some stimulus detectionbut poor discrimination and disturbed stimulus interpre-tation. Axonotmesis is a more severe injury. Afferentfibers undergo degeneration, but the nerve trunk isgrossly intact with variable degrees of tissue injury.Sensory recovery is good but incomplete. The timecourse for sensory recovery depends on the rate of axon-al regeneration and usually takes several months. Thesensory deficit is characterized by a severe paresthesia.Neurotmesis is a complete disruption of the nerve, themost severe injury in the Seddon classification. Sensoryrecovery is not expected except when the nerve coursesthrough a canal like the mandibular canal. The sensorydeficit is characterized by anesthesia.

The Sunderland classification is based on the degreeof tissue injury (Sunderland, 1978). There is similaritybetween the two systems, with Sunderland's offeringgreater detail of description. For the purposes of thisreview, the reader is referred to good descriptions ofSunderland's classification. The purpose of this review isto assess critically our knowledge of peripheral trigemi-nal nerve injury and regeneration in light of Seddon's

18 Crit Rev Oral Biol Ued 6(1):18-24(1995)

classification. It will focus on answers to questions suchas, how does a crush injury differ from a compete dis-ruption in clinical behavior, clinical examination, neu-roanatomy, and electrophysiology? This paper seeks tocorrelate basic science and clinical knowledge.

Nerve Injury and Mechanisms forParesthesia, Anesthesia, and Dysesthesia

The literature is replete with descriptions of nerveinjuries, and there is a growing body of papers dealingwith the results of treatment. Unfortunately, variabilityexists in clinical descriptions. This means that a standardmethod for describing clinical findings such as paresthe-sia, dysesthesia, and anesthesia is often not found.Terms such as "hypoalgesia" and "hypesthesia" serveonly to confuse the issue. Both neurapraxia andaxonotmesis may be characterized by paresthesia, withthe sensory loss and detection of discrimination usuallyless in the more severe injury. Both may exhibit dyses-thetic symptoms in patients. The definitions of each areoften muddled. For this discussion, paresthesia meanspartial loss of sensation, but still some sense of touch,while anesthesia is the complete absence of touch.Dysesthesia is a partial loss of sensation with a painfulor uncomfortable component.

Animal models of nerve injury have used compres-sion, stretching, or sectioning of nerves under controlledexperimental conditions to examine the anatomic, phys-iologic, and neurochemical consequences of peripheralnerve injury. Results show a spectrum of injury that fur-ther complicates direct application to humans.Compression or stretching of a nerve may damage thenerve so that it degenerates, either completely or par-tially. It may occur acutely or may evolve over time due toprogressive irritation or compression. Nerves either par-tially or completely severed may give rise to pathologyfrom disordered central zones, from hyperactive ganglioncells, or from the formation of a neuroma (Devor andWall, 1990). Nerves injured but continuous maintain con-duction, but pathology may arise from disordered func-tion in the zone of injury. Compression or stretching ofnerves may produce a complex of acutely severed axons,axons that are injured but functional, or axons that haveformed a neuroma-in-continuity.

Even mild injuries may cause segmental demyelina-tion, progressive demyelination at the site of damagewhich can by itself lead to pathology. These sites may actas ectopic impulse generation areas (Nordin et al, 1984).At the other end of the spectrum, neuromas resultingfrom complete nerve transection have elevated mechan-ical and electrical sensitivity (Wall and Gutnick, 1974).This spectrum of consequences resulting from differentlevels of injury does not make the task of recognition ofthe type of injury straightforward. What do we know thatpermits more accurate categorization of injuries into

those that will heal without intervention and those thatmight benefit from treatment?

Normal Response to Nerve InjuryInjury to the peripheral trigeminal nerve results indegeneration the degree of which depends in part uponthe magnitude of the injury, the age of the patient, andthe location of the injury (Lieberman, 1974). Thus, for agiven age range and site of injury, the type of injury isoften the dependent variable that predicts outcome.Transection of a peripheral nerve branch (axotomy)results in a greater degeneration than with compressionor crush injuries. The latter are often reversible. TheSeddon and Sunderland classifications are based on his-tologic findings and do not represent functional out-come analysis. We have already discussed the variabilityof pathology even with identical injuries. Can we add adegree of conditional outcome analysis to our currentstate of knowledge?

The normal anatomy deserves some review for fur-ther appreciation of the complexity of the subject. Thenormal nerve trunk is made up of organized collectionsof axons that are the peripheral extensions of the cellbodies located in the trigeminal ganglion. Schwann cellsenvelop axons in a predetermined fashion and producevarious degrees of myelin. In the peripheral nervous sys-tem, a single Schwann cell envelops one axon with amyelin sheath. Unmyelinated axons differ from myelinat-ed axons in that several are ensheathed by one Schwanncell. The endoneurium surrounds either type of axon andconsists of organized collagen fibers. The outermostlayer of these fibers is the basal lamina of the Schwanncell, called the band of Bungner, and is a basal laminatube running the entire length of the axon. An outerregion of fine collagen fibers is further organized into thesheaths referred to as the endoneurium. Many axonswith their endoneurial sheaths are surrounded by a sec-ond organization of collagen fibers called the perineuri-um, which form the fascicle. Fascicular patterns are quitevariable (Svane et al, 1980; Girod et al, 1989). The nervetrunk is completed by the internal epineurium, externalepineurium, and the mesoneurium, which contains theblood supply.

There is an enormous amount of literature on degen-eration and regeneration and the role of the micro-envi-ronment (Seckel, 1990). Direct comparison of axotomyand crush injuries will serve to highlight differences.Axon degeneration in both directions from the injury sitecharacterizes complete transection of a nerve. Originallytermed "Wallerian degeneration", marked morphological,biochemical, and physiologic changes in the nerve cellbody and its processes occur (Gordon, 1983; Barron,1989; Seil, 1989). If death of the nerve cell does not occur,regenerative activity in the form of nerve sprouts comingfrom the proximal stump may begin as early as 24 hours

6(1): 18-24 (1995) Crit Rev Oral Biol Med 19

after injury.Unlike cutaneous wound healing, the healing of

nerve tissues is unique, because the process is greatlydependent upon cellular rather than tissue repair.Successful nerve regeneration requires neuronal growth.The key histologic structure during axonal regenerationappears to be the Schwann cell and its basal lamina {i.e.,bands of Bungner). Schwann cells multiply in the distalnerve segment and when contacted by an axon sprout,undergo a cascade of changes that trigger the productionof myelin (Pelligrino and Spencer, 1985). Some signalcode likely determines if an invading axon will be myeli-nated or unmyelinated. The early regeneration ofunmyelinated axons may explain the early appearance ofpain and temperature sensation in patients' initial recov-ery (Lundborg, 1988b). Despite this strong basic knowl-edge, the usual growth rate in humans (Seddon et al,1943; Buchthal and Kuhl, 1979) of 1 to 2 mm/day makeslittle sense when applied to the human inferior alveolarnerve.

Using sequential double-fluorescence labeling tech-niques, Zuniga and O'Conner (1987) demonstrated that,following mental nerve axotomy and immediate repair inadult rats, mental sensory cells regenerate from andmaintain an organized somatotopic area within thetrigeminal ganglion; regeneration of axotomized cells isa gradual process that is enhanced by immediate surgi-cal intervention; and, although enhanced, surgical repairdid not result in complete recovery of all transected cells(Zuniga and O'Conner, 1987). Delayed repair showedsimilar results except that the time-dependent dynamicincrease in the number of regenerating cells was lessthan if the nerves were repaired immediately (Zuniga etal, 1989). Finally, morphologic quantification of the out-comes of mental axotomy showed that axotomy resultedin 47% loss of cells, which was unaltered by the immedi-ate or delayed repair of the nerves; repair did not affectthe size range of surviving cells-, and diminished volumeof the ganglion and its associated nerve trunk wasrestored following repair, presumably the result of axon-al branching and Schwann cell proliferation (Zuniga et al,1990).

The well-known limitations of horseradish peroxi-dase in nerve studies—such as diffusion out of cell bod-ies and tissue necrosis at the site of injection—led to theuse of fluorescent latex microspheres as a retrogradetracer in the sensory peripheral nervous system. Whenmicrospheres were used in a crushed or intact rabbitinferior alveolar nerve model, results showed that theywere taken up only by damaged axons, they remained inthe trigeminal cell bodies for up to three months withoutdegradation or diffusion to extracellular structure, andcells containing them were capable of regeneratingaxons, as evidenced by the return of evoked sensoryaction potentials and the retrograde axonal transport of

True blue (Colin et al, 1989). This work used a crush injuryas a model and provided some of the only existing elec-trophysiologic data for that type of damage (Colin et al,1986). It is interesting, at this point, to look at currentknowledge of crush injuries and then compare and con-trast that information with that for axotomy. The rate ofaxonal outgrowth is always faster and more complete fol-lowing crush injuries (Lundberg, 1988a). Minimalchanges in cell size and cell death compared with axoto-my are described and axons appear to regenerate to orig-inal receptor sites because Schwann cell columns aremostly undisturbed and serve as guides (Horch, 1979).This propensity for better regeneration is shown by thefinding that even taste sensations are completely recov-ered in animals whose chorda tympani nerve wascrushed and compared with a normal nerve (Robinson,1988). It is important to appreciate that a relatively sim-ple crush injury may be exaggerated by the effects ofcompression. As a separate entity, compression may pro-duce a sensory deficit. The acute response of compres-sion is inflammation and edema. Secondary effectsowing to fibrosis may include localized nerve fiberchanges including segmental demyelination and evenWallerian degeneration if the compression persists. Themechanisms of these injuries include mechanicaldeforming forces and ischemic factors (Dellon, 1980;Lundborg, 1988a).

The early work of Merrill is of interest despite itstechnical limitations (Merrill, 1966). He showed thatcrush injuries in dogs complicated by bony compressionhealed more poorly than crush injuries alone. The meth-ods used were histologic cell counts and a physiologicpain response. In addition, results strongly supportedthe importance of decompression by four weeks forreversal of degeneration to take place. Machida andOmori (1979) studied recovery of nerve action potentialsafter inferior alveolar nerve sectioning in the rabbit.Some animals had the injury compressed by re-applica-tion of the bony lateral plate, and others did not. Therewere no real differences in the recovery of action poten-tials after the first few weeks, leading the authors to con-clude that compression had no lasting effect. It shouldbe observed, however, that compression by replacementof the lateral cortex of bone is quite different and distinctfrom compression due to a bony spicule placed into themandibular canal, as performed by Merrill. The fact thatMerrill's work used axon counts and the Machida andOmori study used action potentials in a volume conduc-tor highlights differences in techniques but does notdetract from the stronger conclusions of the older work.

Electrophysiologic data of crush injury were alsoreported by Colin et al (1986). Longitudinal sensory-evoked potentials of inferior alveolar nerves whose par-ent cell bodies contained microspheres showed no sta-tistically significant decrements of conduction velocities

20 Crit Rev Oral Biol Med 6(1): 18-24 (1995)

or amplitudes by eight weeks after crush injury whencompared with conduction velocities and amplitudesprior to injury and microsphere injection. This return ofaction potential is identical to results obtained by othersafter nerve sectioning and repair (Yamazaki and Noma,1983). The exact relationship of the number of axons toelectrical activity and function is unclear. The inferioralveolar nerve and the trigeminal system have an ultra-structural propensity to compressive injury, because thetrigeminal system has the greatest proportion of myeli-nated axons in the entire somatosensory system.

Clinical Outcome AnalysisResults correlating clinical findings to resolution of neu-rosensory disturbance are important to our understand-ing of the Seddon classification. Recall that neurapraxiaresults in complete recovery, while in axonotmesis,incomplete recovery is described. Both retrospective andprospective studies have been reported. Kipp et al. (1980)examined outcomes following odontectomy and foundthat most injuries resolved by six months. This study waslimited by loss of patients. Osborn et al. (1985), in aprospective study, described similar results but groupedlingual and inferior alveolar nerve injuries together. Inboth of these studies, attention was drawn to the factthat most deficits improved or disappeared before thesix-month mark. This time period is a traditional bench-mark for sensory return. These results suggest that neu-rapraxias are most common. In a retrospective survey-based paper, Ailing (1986) described persistent deficitsin 13% of lingual nerve injuries which had an incidence of0.06%. Others have shown 11% incidence and 0.05% per-sistence rates (Blackburn and Bramley, 1989). Other data(Carmichael and McGowan, 1992) showed a lingual nerveinjury incidence of 15% at 6 to 24 hours, 10.7% at 7 to 10days, and 0.6% after 1 year. A recent paper (Schultze-Mosgau and Reich, 1993) showed no persistent neu-rosensory deficits in patients found to be anesthetic atfirst examination. Other data for the IAN showed an inci-dence of 5.5% at 6 to 24 hours, 3.9% at 7 to 10 days, and0.9% after 1 year (Carmichael and McGowan, 1992), and0.4% incidence and 3.5% persistence at one year (Ailing,1986). In general, recovery of IAN injuries is better thanthat of lingual nerve injuries, probably owing to the guid-ing provided by the mandibular canal.

The results suggest that a certain number of IAN andlingual nerve injuries do not resolve by themselves. Thetask then is how to ascertain which early lesions theserepresent. Clinical exam with a finding of anesthesiaalone or in concert with a positive Tinel's-like sign is onepotential guide. This sign is elicited more commonlyupon palpation over the lingual alveolus. A tingling orshooting sensation to the tip of the tongue is experi-enced. For the IAN, a directed abnormal sensation to thelip may be seen in a few situations upon palpation over

the third molar socket area. The sensation need not bepainful as with a trigger point. A neuroma usuallyexplains these findings. Other methods of testing wouldbe of great usefulness.

Electrophysiologic TestingElectrophysiologic testing can determine the inducedaction potentials of sensory nerves and conductionvelocity (Yamazaki and Noma, 1983). Some of these elec-trophysiologic methods are applicable in humans, butmost are not as yet useful for the peripheral trigeminalsystem. Nerve conduction velocity testing can probe theperipheral trigeminal complex, and perhaps trigeminalsomatosensory evoked potentials (TSEP) may cast lighton the more central effects of injury and repair.

The trigeminal system has been clinically evaluatedby a variety of electrophysiological techniques; however,each method has inherent shortcomings and limitations.Electrical threshold testing is a technique whereby anelectrical current is passed in increasing increments untila sensation is barely elicited. Using this technique,researchers have found electrical sensory thresholds tobe elevated when placed on a paresthetic lip, but todiminish with time as the paresthesia resolves (Lavant,1967). This test is effectively an electrical version of themany sensory reflex examinations that are available. It islimited by the patient's level of cooperation and pain tol-erance, and certainly does not localize a neural lesion.

A variety of reflex studies have been developed, likethe blink reflex, that test the trigeminal sensory pathwaysand the motor branches of the facial nerve (Kimura,1984a,b), but they involve one or more synapses. Thistest fails to localize the neural lesion in the reflex arc andis a relatively insensitive test unlikely to discern subtleafferent injuries (Leandri and Favale, 1991). The jawreflex and evoked masseteric silent period, which involveboth afferent and efferent loops of the trigeminal nervethroughout the mesencephalic trigeminal nucleus in thebrainstem, have the same shortcomings as described.The TSEP has been advocated as an electrophysiologicmethod for the evaluation of peripheral and centraltrigeminal pathways in health and disease (Prevac, 1970;Stohr and Petruch, 1979; Bennett and lannetta, 1980;Findler and Feinsod, 1982; Barker etal, 1987; Godfrey andMitchell, 1987; Larson and Pogrel, 1992). Electrical stim-ulation of the trigeminal branches will evoke volleys ofperipheral, ganglionic, spinal, and cortical potentials.There is much disagreement as to the normal latencies ofthese signals because of variability (Pogrel, 1992). Ameaningful correlation of the waveforms and the under-lying neural generators is still lacking (Leandri andCampbell, 1986). Therefore, the TSEP has not had rou-tine clinical use.

Some progress has been made in the study of animaland human nerve injuries electrophysiologically, as dis-

6(l):18-24 (1995) Crit Rev Oral Biol Ued 21

cussed previously. A recent study in human volunteers(Colin, 1993) found a conduction velocity of 67 m/s at theonset of the signal and a conduction velocity of 50 m/s atthe peak of the signal. Others (Matsuda, 1980b) found amean conduction velocity of 55.8 ± 2.95 m/s for thehuman IAN by stimulating the mental nerve and record-ing from the mandibular foramen. In this study, theinvestigator averaged the onset latency to be halfwaybetween the first positive and the first negative peak, sothis did not represent either the conduction velocity atthe onset of the signal or the conduction velocity at thepeak of the signal, but rather an average of the two. Also,only one side was studied in each of the 20 male andfemale volunteers, so it is unclear if there is any side-to-side variation or gender difference. Sasaki et al. (1986)found a normal IAN maximal conduction velocity of 61.4± 7.8 m/s, ranging from 51 to 77 m/s, but only studied theright IAN in nine volunteers of both genders. Recently, areport claimed successful IAN recording (Jones andThrash, 1992) but obtained good signals in only six of 10subjects and estimated the conduction distance.

IAN conduction studies have appeared in the litera-ture. Control rabbit IAN signals have had a similar rangeof conduction velocities of 54 m/s in our studies (Colinand Donoff, 1990) and in others, 58 m/s (Matsuda,1980a), 40 m/s (Edinger and Luhr, 1986), and 22.7 m/s(Eppley et al, 1989). In these investigations, the conduc-tion distance was estimated from surface measurements,resulting in erroneous calculations of the conductionvelocity.

Before electrophysiologic tests like conductionvelocity become useful in human patient care, the inac-curacy caused by measurement of the conduction dis-tance at the skin surface must be overcome.Determination of this small distance by a skin measure-ment is especially inaccurate. Radiographs have beenused to try to minimize the inaccuracy of distance deter-mination (Matsuda, 1980b; Colin, 1993). Other aspects ofnerve conduction velocity determination have inherentproblems. The stimulating circuit, recording system, andinadvertent stimulation of nerves and muscles are justsome of the concerns.

The long-term goal of such electrophysiologic stud-ies would be to predict the outcome of injury and providesubstantiation of current clinical methods of testing asguides to decision-making. At this time, the evaluationof these sensory disturbances is still best done by histo-ry and physical examination.

Merging Science and ApplicationMost clinicians agree that early treatment of severe nerveinjuries is indicated, but proof is lacking. Animal studieslike that of Zuniga (Zuniga and O'Conner, 1987) supportthis view, but few hard data exist in humans to supportthis. Most importantly, while an enormous body of

knowledge exists on the neurophysiology of the trigemi-nal system, little of this is pertinent to clinical consider-ations in trigeminal nerve injuries. Problems arise incomparing results of outcome studies because of theinconsistent use of terms like paresthesia, dysesthesia,and anesthesia. Exclusion and inclusion criteria do notexist, and there certainly have not been any randomized,double-blinded trials of surgical intervention.

The scientific core of knowledge on the subject doesstrongly suggest that early intervention will capture the"regenerative" power of the nerve cell. Axonal recovery,cellular recovery, and even receptor recovery appear tobe time-dependent. Some clinical results support thiscontention (Donoff and Colin, 1990), but others have notdifferentiated among patients with regard to demo-graphics of age, gender, race, and mechanism or type ofinjury. A multicenter retrospective study groupedpatients into four categories: patients with lingual andinferior alveolar nerve injuries, and patients who dis-played either hypoesthesia or hyperesthesia stimulustesting, with or without pain (Labanc and Gregg, 1992).The major finding of this study of 521 patients was thatthere was a high predictability of results for the hypoes-thetic patients. Thus, clinical findings of anesthesia,Seddon's neurotmesis, particularly when combined witha Tinel's-like sign upon palpation in the lingual area forlingual nerve deficits, permit surgical treatment to becarried out with confidence. It is unlikely that this type ofinjury falls into any of the categories of nerve damagewhich may show spontaneous resolution of the sensorydeficit.

Summary and ConclusionThe healing of a cutaneous wound is characterized by tis-sue repair with the purpose of regenerating dermis andepithelium. Osseous healing seeks to restore osteocyteand an osteoid matrix capable of calcification. Both ofthese responses take place within reasonable bound-aries of space. Surely there are influences other thanlocal factors, but for the most part, the machinery ofrepair, the cells involved, the molecular biologic eventsof DNA synthesis, cell division, etc., take place close tothe area of injury.

The healing of a damaged nerve is unique. Cellularevents occur in a site, the ganglion, located "miles" fromthe site of injury. Cells around the body of nerve cellextensions—called axons, Schwann cells—have a crucialrole in axonal recovery. A body of knowledge regardingso-called neurotrophic factors exists which hasn't evenbeen mentioned (Seckel, 1990). That nerve tissue is pre-cious goes without saying. The management of nerveinjuries of the IAN and lingual nerves in particular is stillembryonic. Much has been learned, but much moreneeds to be learned about the normal and abnormalbehavior of injured sensory nerves.

22 Crit Rev Oral Biol Ued 6(l):18-24 (1995)

If this review has emphasized anything, it is that theoriginal Seddon classification, if followed literally,remains an excellent guide to the management of nerveinjuries. Consistency of description in clinical examina-tion with careful and close follow-up offers the patientand the clinician the best chance of reaching a success-ful outcome.

REFERENCES

Ailing CC (1986). Dyesthesia of the lingual and inferioralveolar nerves following third molar surgery. I OralMaxillofac Surg 44:454-457.

Barker GR, Bennett AJ, Wastell DG (1987). Applications oftrigeminal somatosensory evoked potential in oraland maxillofacial surgery. Br I Oral Maxillofac Surg25:308-313.

Barron KD (1989). Neuronal responses to axotomy: con-sequences and possibilities for rescue from perma-nent atrophy or cell death. In: Frontiers of clinicalneuroscience. Neural regeneration and transplanta-tion. Vol. 6. Seil FJ, editor. New York: Alan R. Liss, pp.79-99.

Bennett M, lannetta PJ (1980). Trigeminal evoked poten-tials in humans. Electroenceph Clin Neurophysiol 48:517-526.

Blackburn CW, Bramley PA (1989). Lingual nerve damageassociated with the removal of lower third molars. BrDent] 167.103-107.

Buchthal F, Kuhl V (1979). Nerve conductor, tactile sensi-bility, and electromyogram after suture or compres-sion of peripheral nerve. A longitudinal study in man.1 Neurol Neurosurg Psychiatry 42:436-451.

Carmichael FA, McGowan DA (1992). Incidence of nervedamage following third molar removal. Br J OralMaxillofac Surg 30:78-82.

Colin W (1993). Conduction velocity of the human inferi-or alveolar nerve: A preliminary report. J Oral MaxillofacSurg 51:1018-1023.

Colin WB, Donoff RB (1990). Electrodiagnostic evalua-tion of the uninjured rabbit inferior alveolar nerve. IOral Maxillofac Surg 48:170-173.

Colin W, Edwards DD, Sandrew BB, Foote WE, Donoff RB(1986). Trigeminal ganglion cell labelling after crushinjury of rabbit inferior alveolar nerve (abstract). I DentR£s65(SpecIss):245.

Colin W, Donoff RB, Foote WE (1989). Fluorescent latexmicrospheres as a retrograde tracer in the peripheralnervous system. Brain Res 486:334-339.

Dellon AC (1980). Clinical use of vibratory stimuli to eval-uate peripheral nerve injury and compressor neu-ropathy. Plast Reconstr Surg 65:466-476.

Devor M, Wall PP (1990). Cross excitation in dorsal root

ganglia of nerve injured and intact rats. ) Neurophysiol64:1733-1746.

Donoff RB, Colin W (1990). Neurologic complications oforal and maxillofacial surgery. In: Oral and maxillofa-cial clinics of North America. Vol. 2. Assael L, editor.Philadelphia, PA: W.B. Saunders Co., pp. 453-462.

Edinger D, Luhr H (1986). Free autologous nerve graft-ing—comparison of suture methods. ) Maxillofac Surg14:227-230.

Eppley B, Doucet M, Winkelman T, Delfino |( (1989).Effect of different surgical repair modalities on regen-erating of the rabbit mandibular nerve. I Oral MaxillofacSurg 47:257-274.

Findler G, Feinsod M (1982). Sensory evoked response toelectrical stimulation of the trigeminal nerve in man.J Neurosurg 56:545-549.

Girod SC, Nenkam DW, Girod B, Reumann K, Semran H(1989). The fascicular structure of the lingual nerveand the chorda tympani: An anatomical study. I OralMaxillofac Surg 47:607-609.

Godfrey RM, Mitchell KW (1987). Somatosensory evokedpotentials to electrical stimulation of the mentalnerve. Br J Oral Maxillofac Surg 25:300-307.

Gordon T (1983). Dependence of peripheral nerve ontheir target organs. In: Somatic and autonomic nerve-muscle interactions. Burnstock G, O'Brien RA,Urbov'a G, editors. Amsterdam: Elsevier, pp. 289-325.

Horch K (1979). Guidance of regrowing sensory axonsafter cutaneous nerve lesion in the cat. ) Neurophysiol42:1437-1439.

Jones D, Thrash W (1992). Electrophysiologic assessmentof human inferior alveolar nerve function. J OralMaxillofac Surg 50:581-585.

Kimura I (1984a). The blink reflex. In: Electrodiagnosis indiseases of nerve and muscle: Principles and practice.Philadelphia, PA: Davis, pp. 323-351.

Kimura I (1984b). The H-reflex and other late responses.The jaw reflex. In: Electrodiagnosis in diseases ofnerve and muscle-. Principles and practice.Philadelphia, PA: Davis, pp. 385-386.

Kipp DP, Goldstein BH, Weiss WW (1980). Dysesthesiaafter mandibular third molar surgery. I Am Dent Assoc100:185-192.

LaBanc IP, Gregg JM, editors (1992). Trigeminal nerveinjuries. In: Oral and maxillofacial surgery clinics ofNorth America. Vol. 4. Philadelphia, PA: W.B.Saunders, pp. 277-283.

Lavant BA (1967). Mental anesthesia and its prognosis.Br I Oral Surg 4:206-208.

Larson LE, Prevac TS (1970). Somatosensory response tomechanical stimulation as recorded in human EEG.Electroencephalogr Clin Neurophysiol 28:162-166.

Leandri M, Campbell JA (1986). Origin of early wavesevoked by infraorbital nerve stimulation in man.Electroenceph Clin Neurophysiol 65:13-19.

6(l):18-24 (1995) Crit Rev Oral Biol Med 23

Leandri M, Favale E (1991). Diagnostic relevance oftrigeminal evoked potentials following infraorbitalnerve stimulation. J Neurosurg 75:244-250.

Lieberman AR (1974). Some factors affecting retrogradeneuronal responses to axonal lesions. In.- Essays onthe nervous system. Bellars R, Gray EG, editors.Oxford, England: Clarendon Press, p. 71.

Lundborg G, editor (1988a). Compression and stretching.In.- Nerve injury and repair. London, England:Churchill Livingstone, p. 64.

Lundborg G, editor (1988b). Nerve regeneration. In:Nerve injury and repair. London, England: ChurchillLivingstone, p. 149.

Machida K, Omori K (1979). Recovery of nerve actionpotentials and vascularization after inferior alveolarnerve sectioning. Bull Tokyo Dent Coll 20:101 -114.

Matsuda Y (1980a). Effect of nerve injuries on actionpotential of inferior alveolar nerve. 1. Experimentalstudy on rabbit. Shikwa Gakuko 80:1267-1285.

Matsuda Y (1980b). Effect of nerve injuries on actionpotential of inferior alveolar nerve. 2. Clinical Study.Shikwa Gakuko 80:1591-1519.

Merrill RG (1966). Further studies in decompression forinferior alveolar nerve injury. I Oral Surg 24:233-238.

Nordin M, Nystrom B, Wallin U, Hagbarth KE (1984).Ectopic sensory discharges and paresthesia inpatients with disorders of peripheral nerves dorsalroots and dorsal columns. Pain 20:231-245.

Osborn TP, Frederickson G, Small IA, Torgerson TS (1985).A prospective study of complications related tomandibular third molar surgery. I Oral Maxillofac Surg43:767-769.

Pelligrino RG, Spencer PS (1985). Schwann cell mitosis inresponse to regenerating peripheral axons in vivo.Brain Res 341:16-25.

Pogrel MA (1992). Trigeminal evoked potentials and elec-trophysical assessment of the trigeminal nerve. In-.Trigeminal nerve injury: Diagnosis and management.Vol. 4. Gregg J, LaBanc J, editors. Oral and MaxillofacialSurgery Clinics of North America. Philadelphia, PA:W.B. Saunders, pp. 535-541.

Robinson PP (1988). The reinnervation of the tongue andsalivary glands after lingual nerve injuries in cats.Brain Res 483:259-271.

Sasaki K, Masaki H, Miyake S (1986). Restoration of infe-rior alveolar nerve injury after orthognathic surgery ofthe mandible. )pn J Oral Maxillofac Surg 32:1139-1142.

Schultze-Mosgau S, Reich RH (1993). Assessment of infe-rior alveolar and lingual nerve disturbances after den-toalveolar surgery, and of recovery of sensitivity. \nt JOral Maxillofac Surg 22:214-217.

Seckel BR (1990). Enhancement of peripheral nerveregeneration. Muscle and Nerve 13:785-800.

Seddon HI (1943). Three types of nerve injury. Brain66:237-288.

Seddon HJ, Medewar PB, Smiths H (1943). Rate of regen-eration of peripheral nerves in man. J Physiol (London)102.191-215.

Seil FJ, editor (1989). Axonal sprouting in response toinjury. In: Frontiers in clinical neuroscience. Vol. 6.Neural regeneration and transplantation. New York:Alan R. Liss, pp. 123-138.

Stohr M, Petruch F (1979). Somatosensory evoked poten-tials following stimulation of the trigeminal nerve inman. J Neurol 220:95-98.

Sunderland S (1978). Nerves and nerve injuries.Edinburgh: Churchill Livingstone, pp. 31-60.

Svane TJ, Wolford CM, Milam SB, Bass RK (1980).Fascicular characteristics of the human inferior alveo-lar nerve. I Oral Maxillofac Surg 44:431-434.

Wall PD, Gutnick M (1974). Ongoing activity in peripher-al nerves: the physiology and pharmacology ofimpulses originating from a neuroma. Exp Neurol43:580-593.

Yamazaki Y, Noma H (1983). Comparison of suture meth-ods and materials in experimental inferior alveolarnerve grafting. J Oral Maxillofac Surg 41:34-46.

Zuniga JR, O'Conner B (1987). Primary and secondarymicroneuroanastomotic repair of the mental nerve inthe rat. Int J Oral Maxillofac Surg 16:465-472.

Zuniga JR, Pate JL, Hegtvedt AK (1989). Trigeminal gan-glion response to immediate and delayed peripheralnerve injury and repair (abstract). I Oral Maxillofac Surg47(Suppl):102.

Zuniga JR, Pate JL, Hegtvedt AK (1990). Regenerativeorganization of the trigeminal ganglion followingmental nerve section and repair in the adult rat. JComp Neurol 295:548-555.

24 Crit Rev Oral Biol Med 6(I):18-24 (1995)