Brain Research, 90 (1975) 195-204

195

~) Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands

PROJECTION OF T O O T H PULP A F F E R E N T S TO T H E CAT T R I G E M I N A L N U C L E U S CAUDAL1S

S A M U E L G. N O R D AND RONALD F. YOUNG

Department of Neurology (S.G.N.) and Department of Neurosurgery (R.F.Y.), Upstate Medical Center, Syracuse, N. )1. 13210 (U.S.A.) (Accepted January 20th, 1975)

SUMMARY

Unit activity was recorded extracellularly from cat medullary neurons following electrical stimulation of the canine tooth pulp. Response characteristics of the neurons quickly stabilized at specific suprathreshold stimulus intensities but such properties as spike latency, interspike interval and spike density varied systematically as intensity was raised to maximally effective values. Receptive fields were principally unilateral. The majority included both canines and extended into other oro-facial areas. Suppression of a pulpal response could be effected by preceding tooth stimulation with a conditioning stimulus applied to some other point in the receptive field of the responding cell at an appropriate interstimulus interval. In contrast, a pulpal response could be enhanced by presenting two stimuli successively to the same canine at such intervals. Similar enhancing effects followed simultaneous stimulation of spatially segregated loci in a field. The pulp-responsive neurons were localized histologically in, or in the immediate vicinity of, the nucleus caudalis of the spinal trigeminal complex where the possibility of their existence has been questioned previously. Most of the cells were situated along the ventromedial border of the nucleus, a region reported to contain other pain-related neurons with trigeminal fields.

INTRODUCTION

The nucleus caudalis (NVCaud) of the bulbar trigeminal complex is known, on clinical groundsl0, is, to play an essential role in the mediation of facial pain. Moreover the nucleus has been shown electrophysiologically to contain pain-related neurons 12-t4 which often have peripheral fields limited in area to portions of the gums, oral mucosa, tongue and lips 13,14. Yet, electrical stimulation of the tooth pulp has been used with inconsistent effectiveness to evoke neural activity in the vicinity of

196 NVCaud even though the pulp is largely, if not exclusively, involved with pain sensibility1, l~. For example, some investigators'hi9 failed to obtain any clear-cut pulpa/ responses in NVCaud in experiments in which both unit activity and field potentials were readily recorded at more rostral levels of the complex. Other investigators did record pulpal unit activity in the nucleus caudalis, but relatively few cells were sampled and these were restricted in location to the rostral pole of the nucleus ~,6. The only pulpal responses monitored in larger portions of NVCaud have been field potentials~ and, significantly, it has been shown that pulpal field potentials areconsiderabty smaller in this nucleus than those evoked elsewhere in the complex 6. The results of these investigations lead to the paradoxical conclusion that very few afferent fibers project from the tooth pulp, a known trigeminal pain sensor, to the neurons of NVCaud, the principal trigeminal pain relay of the medulla. The present experiments were undertaken to examine this apparent paradox by using electrophysiological and stereotaxic techniques to explore the nucleus caudalis for individual neurons responsive to electrical stimulation of the tooth pulp. METHODS

Experimental data were collected from 12 adult cats. Initially, each animal was anesthetized with pentobarbital sodium (30 mg/kg). Supplementary doses of 25 mg were administered irregularly during the experiment in order to keep the animal at a light anesthetic level. In addition, the animal was immobilized with gallamine triethiodide and artificially ventilated through a tracheal cannula. Drugs, as well as a continuous drip of 5 To dextrose in saline, were administered through a femoral vein catheter. Expired Pcoz and arterial blood pressure were monitored continuously. Body temperature was maintained by means of a thermostatically controlled hot water pad. Exposed brain tissue was covered with warmed saline which was changed irregularly. In order to facilitate placement of stimulating electrodes and to permit exploration of the buccal cavity, the jaws were fixed in an open position by cementing a 25 mm length of rigid plastic tubing (diameter, 3 ram) between the contralateral upper and lower molars with dental acrylic. Using a dissection microscope, the incisal tips of the ispilateral maxillary and mandibular canine teeth were removed with a dental burr in order to permit access to the pulpal cavities. A concentric, bipolar electrode which had been insulated previously to within 1 mm of the tip was then fitted into each cavity opening until the exposed conductive surface was completely embedded. Using this procedure, the electrode tip just made contact with the distal extreme of the pulp. Each electrode consisted of a 28-gauge (diameter, 0.35 mm) cannula and an insulated inner wire core (diameter, 0.075 mm). After insertion, all electrodes were cemented in place and were completely covered with dental acrylic. The animal was then fixed in a modified stereotaxic position and the medulla exposed as in previous experiments l'~. Extracellular unit activity, recorded through tungsten microelectrodes (tip diameter I #m), was amplified and monitored by conventional methods. Criteria for distinguishing neuronal from fiber responses were those described in detail by Kerr eta/. 7. Appropriate data were stored on magnetic tape for further study and analysis. In each experiment, the recording electrode was positioned stereotaxically at the

197 bulbar surface in order to conform with the location of NVCaud. Stimulation of the tooth pulp was begun when the electrode had penetrated the surface of the medulla and continued while the electrode was advanced by means of a remotely driven micromanipulator. During this phase of the experiment, square-wave stimulus pulses of 10-50 V amplitude and 0.5-10 msec duration were delivered simultaneously through stimulus isolators to the two canines, typically at a frequency of 0.3 pulses/sec. Relatively wide ranges of intensity and duration were necessary since resistance between stimulating electrode poles varied considerably (range, 0.2-1.4 M~)). Current spread to structures outside the tooth pulp was minimized by the use of insulated bipolar stimulating electrodes extending through the outer layers of the tooth into the pulpal canal z0. In our experiments the spread of current was further limited by using concentric electrodes which result in a very restricted region of stimulation iv, particularly when the inner core of the electrode is made the cathode pole. As the electrode was lowered and the pulp stimulated, the face and the buccal cavity were tested manually with simple mechanical probes t4. This procedure was used to activate NVCaud cells whose physiological characteristics and spatial organization in the cat have been described in great detailS, s,9, thereby to ensure that the electrode was positioned in the nucleus. When spike potential responses to pulpal stimulation were encountered, the electrode was lowered a few more micrometers until maximum spike amplitude was obtained. Stereotaxic coordinates of the electrode tip were recorded. Then, receptor field and response properties of the unit were determined. In each case, an attempt was made to measure the latency and spike density of the response at stimulus intensities which consistently evoked discharges in the cell under study. These characteristics were measured when each implanted canine was stimulated alone and during simultaneous stimulation of the teeth. In several instances stimulus voltage or duration was systematically varied from below threshold to supramaximal values so that stimulus-response relationships might be determined. The face and mouth were then explored with the mechanical probes to determine the distribution and relative magnitude of the peripheral field of the active cell. Conditioning experiments were performed with 25 cells to test for evidence of afferent interaction. In these experiments an electrical conditioning stimulus (CS) applied to some part of the receptive field of the cell preceded a test stimulus (TS) applied to the same or to a different part of the field by a predetermined interstimulus interval (ISI). For non-dental stimulation a concentric, bipolar electrode, sharpened, but otherwise identical to the dental pulp electrodes was inserted into the receptive field which had been delineated previously by mechanical stimulation. Such areas included the facial skin, tongue, lip, nose or gum tissue. The pulp stimuli used in conditioning experiments had parameters which were similar to those noted above. Electrical pulses applied to non-dental fields varied from 9,5 to 40 V in amplitude and from 0.3 to 10 msec in duration. In early experiments, blocks of control trials (i.e., TS presented alone) were interspersed with blocks of conditioning trials (paired CS-TS presentations). Each block consisted of 10 or more trials. An intertrial interval of 3 sec was used and the ISI remained constant within each block. When this procedure was followed, however, several units were lost and others gave evidence of fatigue

198

Fig. 1. Photomicrograph of a cross-seciion of the medulla 0.5 mm caudal to the obex. Microlesi(m ;~i arrow was made at the site at which pulp-responsive units were encountered. The horizontal bracket indicates 0.5 ram; NVCaud, Irigcmina! nucleus caudalis. before an adequate n u m b e r o f ISis could be studied. As a result, a within-blocks randomization paradigm was developed. In this paradigm, one control trial and several conditioning trials tone for each ISI to be tested) were presented in a randomized fashion in each block, such that each block had a different arrangement of ISls. Typically, a full conditioning experiment consisted o f 10 or more blocks. Within a single block, trials were separated by pauses of 5 sec. Stimulus pulses and evoked unit activity were recorded on separate bands o f magnetic tape. Postexperimentally, the mean number of spikes obtained for each ISI were c o m p a r e d to the mean number of spikes obtained in control trials. Periodically, gallamine triethiodide was administered while a pulp unit was being studied in order to ensure that it was not responding secondarily to reflex activation o f jaw muscles. A microlesion was made at the ventralmost point in each penetration by passing current through the recording electrode 14. In addition, identical lesions were frequently made at, or near, the loci at which pulpal unit activity was recorded. The animal was sacrificed at the termination o f each experiment. The brain stem was transected rostrally and caudally to the region explored during the experiment and the resulting block of tissue was excised. The histological procedures used in treating the tissue and in preparing microscopic slides, as well as techniques for deter-

199 +1 I

,,,~,~,

0

-I

I

-2

Obex 4

I

-5

I

-4

I

-5 I

-6mm I

( Midline ) Rostral

Caudal

I"

3

7 " : " ." . • LjJ " .-" 5-

E I E

Fig. 2. Diagrammatic representation of the medulla in the longitudinal plane illustrating all electrode placements. Each placement was determined by stereotaxic measurement and inspection of appropriate brain stem sections.

mining the locations of cells activated during experiments, have been described previously 14. A photomicrograph with a microlesion made at the site at which pulp responsive units were encountered is presented in Fig. I. RESULTS

Pulpal stimulation was effective in evoking unit activity in 83 medullary neurons. The neurons were isolated in 33 electrode penetrations which ranged 6.1 mm caudally from a point 0.6 mm rostral to the obex (Fig. 2). Typically, the neurons were spatially intermingled with cells which responded to mechanical manipulation of facial or other fields, and most (87 ~ ) were situated between 1.0 and 2.5 mm below the bulbar surface. The distribution of latencies of cells responding consistently to stimulation of a single tooth pulp was rather broad, but positively skewed, with 86 ~ of the values ranging from 6 to 30 msec. The latencies generally varied inversely with stimulus intensity (1), but became stable when an I value which evoked a relatively constant spike configuration was reached. Pulpal neurons typically responded with brief bursts of very few spikes when a tooth was stimulated. For example, of 48 cells in which this variable was studied systematically, 20 responded with 1 or 2 spikes/stimulation and only 4 responded with more than 5. In some cases, the number of spikes in a burst increased slightly as I was increased beyond the level at which a consistent response was evoked. Once a maximum number of spikes was elicited, further increments in I reduced the interspike interval within the burst in addition to shortening the latency, as noted above. Response bursts in apparently spontaneously active neurons were typically followed by silent periods of at least 200 msec duration. Simultaneous stimulation of two sites in the peripheral field of a cell, such as the two canines, ordinarily produced a greater number of spikes than stimulation of either site alone. However, the resulting response always contained fewer spikes than the arithmetic sum of the spikes in the two individual responses.

200 ~oo! a

°°i

,S--" ~--'= 10 r n s e c

q 20

0

I00 200 Inlerstirnulus

300 400 Interval (msec}

500

~

I

Fig. 3. Response of a single pulpal neuron in NVCaud during conditioning. The TS was applied to the maxillary canine pulp and the CS to the ipsilateral gum. A : response supression at various ISls. Conditioned response levels (mean number of spikes following CS-TS pairings) are expressed as percentages of control trial response levels (mean number of spikes following TS stimulation only). B: three successive control trial responses of the neuron. C: three successive conditioning trial responses (ISI : 200 msec).

Sixty-nine neurons were tested for responsiveness to stimulation of both implanted canines. Of these, 14 were activated by lower tooth stimulation alone, 18 by upper tooth stimulation and the remainder (54 3~ of the total) by stimulation of either tooth. The size and distribution of peripheral fields were determined for 46 of the 'pulpal' neurons. In 38 cases, the fields extended beyond the canines. The majority' of these (61%) were limited to restricted ipsilateral oral or perioral zones. Another sizeable group (26 %) extended beyond the perioral region to include larger areas of the ipsilateral face. Five neurons had widespread fields which included areas of the body other than the face and 3 of these extended into the contralateral skin. Although each of the neurons with non-dental fields responded to peripheral mechanical stimulation, o/ 45/o could only be activated when the stimulus reached noxious levels, as previously described 14. Pulpal response suppression was obtained in 13 conditioning experiments in which the CS and the TS were applied to separate locations in the field of the neuron under study. The degree of suppression varied from total (i.e., no TS spikes) to an appreciable reduction of the mean number of TS spikes below control trial levels (Fig. 3). F o r example, 7 neurons which were studied in full conditioning paradigms using dental pulp stimulation as the TS responded with between 38 and 100 %, fewer spikes at maximally effective ISis than they did in control trials. When the suppression was less than total, the first (or only) spike in the remaining TS response usually fired at the same latency as the first spike of control trial responses (Fig. 3B and C). In such cases, suppression appeared to be specific to the later spikes in the response burst. Generally, suppression was most pronounced with ISis between 60 and 100 msec in duration. Intervals of between 200 and 500 msec had comparatively less effect, and 1Sls which were longer than 500 msec did not alter the TS response in any demonstrable fashion. Brief ISis (20-60 msec) usually produced suppression but in 3 experiments they

201 A

2ooJ t60

-~ 120 _~

3

~

I0 msec

8o-

40-

TS- Upper Canine CS- Upper Canine

0 0

I00 200 300 400 Interstimutus Interval (msec)

Fig. 4. Response of a single pulpal neuron in NVCaud during conditioning. Both CS and TS were applied to the ipsilateral maxillary canine pulp. A: response enhancement at various lSls. Conditioned response levels are expressed as in Fig. 3. B: three consecutive control trial responses. C: three consecutive conditioning trial responses (IS1 80 msec).

brought about a marked increase in the number of test response spikes over control trial levels. Conditioning stimulus sites which produced measurable suppression included the two implanted canines, the gums, lips, nose and the periorbital region. One of the canines was usually the site of TS application, although successful results were also obtained in two experiments in which the lower gum received the TS and one in which the tongue was stimulated. Conditioning experiments in which the CS and the TS were delivered sequentially to the same canine pulp resulted in TS response enhancement (i.e., an increase in the mean number of spikes/response over control levels). This effect was accompanied by a decrease in the latency of the first spike of response groups and a shortening of the interspike interval in those groups, as suggested by Fig. 4. Enhancement was produced over the same range of ISI values which resulted in response suppression when spatially segregated conditioning and test stimuli were used. As noted above similar increments in spike production with concomitant decreases in latency and interspike interval were obtained under two other circumstances: (1) following simultaneous stimulation of two sites in the peripheral field of a cell, particularly the two implanted canines and (2) following brief ISis in some of the conditioning experiments in which the CS was applied to non-dental trigeminal fields and the TS to the canine pulp. Only neurons whose locations in the brain stem could be determined from histologically prepared sections are reported in the present study. The majority of these (84 %) were situated either along the ventromedial border of NVCaud or in immediately adjacent portions of the lateral reticular formation (LRF). Of the remaining ceils, 4 were localized well within the nucleus proprius of NVCaud, 2 near the dorsomedial extreme of NVCaud, 4 in more dorsal regions where numerous trigeminal fibers are intermingled with cell bodies and 3 in deeper lateral reticular zones. No systematic

202 differences were observed to exist between those neurons lying in the LRF and ti~,~sc which were situated either in the ventromedial NVCaud or in the other regions ntqcd above. DISCUSSION

The present experiments demonstrate that an appreciable number of neurons which respond to electrical stimulation of the canine tooth pulp are situated in NVCaud or in contiguous medullary structures. The neurons tend to be distributed along the ventromedial aspect of the nucleus where obvious borders between N VCaud and the adjacent LRF do not existVL Pulpal units have been encountered in this area previously 4,~ as have neurons which are responsive to coarse or noxious stimulation of other trigeminal fields ~,1a,14. However, some investigators have not been able to evoke significant pulpal activity at this level of the medulla and, as a result, they have proposed that pulpal afferents do not project to the cells of NVCaud 2,1~. Such an interpretation of their data can be questioned on methodological grounds. Specifically, they attempted to locate the more deeply situated responding neurons by exploring the surface of the lower medulla lbr field potentials during periods of tooth pulp stimulation. As Greenwood 6 has suggested regarding this point, comparatively few A-delta canine pulp afferents synapse with these neurons and, consequently, their activation by pulpal stimulation should give rise to very low amplitude remote field potentials. Such potentials might easily escape detection by conventional surface recording techniques. Thus, failure to identify evoked surface activity under these circumstances might be more indicative of technical limitations than of an absence of appropriately responding neurons. Response characteristics of the neurons studied in the present experiments compare favorably with those of the pulpal neurons encountered by Dunker et al. ~ in a portion of the spinal trigeminal complex extending from the rostral pole of NVCaud to the caudal extreme of the nucleus oralis. For example, the responses recorded by Dunker et al. 4 also consisted of a single spike, of 2 or 3 spikes with short interspike intervals or of somewhat larger groups with variable intervals. Furthermore, the distributions of initial spike latencies obtained in the two investigations were skewed in the same manner, with more than 90 );; of the values being 5 msec or longer in each. In both investigations, too, the latencies generally became shorter and the number of spikes in response bursts increased as stimulation intensities were raised above threshold. As in our experiments, some of the neurons recorded by Dunker et al. ~ responded to stimulation of one canine (i.e., maxillary or mandibular) whereas others responded to stimulation of either. In the latter case, when brief intervals separated the application of stimuli to the two teeth, the response to the second stimulus was frequently suppressed. The suppression, which is analogous to the effect obtained in our conditioning experiments, was found to be maximal when intervals of 90 msec separated the stimuli. As noted above, the suppression observed in the present investigations was most pronounced when ISis of 60-100 msec were used. The many similarities in response characteristics displayed by the two groups of pulpal neurons, as well as their overlapping spatial distributions, lead to the obvious conclusion that they were

203 sampled from a common population of cells situated ventromedially in the caudal two nuclei of the spinal trigeminal complex. Thus, the data of the two investigations considered together present convincing evidence that neurons receiving pulpal input are dispersed longitudinally throughout a section of the complex in which their existence has been questioned in the past 2,19. We have demonstrated that presumably nociceptive, 'pulpal' neurons situated principally along the ventromedial border of NVCaud are influenced by converging, interacting, impulse barrages originating in dental and non-dental trigeminal fields. The most notable effect of these converging barrages is the response suppression brought about by sequential excitation of oro-facial afferents projecting to the neurons from different points in their receptive fields. As noted above, similar inhibitory influences upon pulpal responses in the spinal trigeminal complex have been reported in another study4. The second major effect of interacting afferent barrages converging upon the pulpal neurons is response enhancement. This was observed in several circumstances which did not typically produce suppression. Other pain-related trigeminal neurons have been localized in the same regions of NVCaud la,14 as our pulpal cells and these are also differentially affected by apparently excitatory and inhibitory impulse trains initiated at different peripheral loci. The results of the present investigation considered in the light of these findings are compatible with two theoretical positions regarding bulbar mechanisms of trigeminal pain. (1)The most deeply situated cells of NVCaud are intimately involved in the mediation of facial pain. These cells lie both within the nucleus proprius of NVCaud and in the immediately adjacent lateral reticular formation la,t4. (2) Normal facial pain sensibility requires the transmission of spatially summating inhibitory and facilitatory effects from broad facial fields to a polyneuronal network via the descending tract of the trigeminal nerve a. Recently, Young and King21 proposed that information concerning noxious facial stimuli projected both rostrally into the main sensory-oralis region of the sensory trigeminal complex and caudally into NVCaud. Interaction between these nuclear zones was emphasized as essential for the differentiation of a noxious from a nonnoxious facial stimulus. They suggested, in particular, that neuronal activity in NVCaud might modulate the firing patterns of the more rostral cells. Sessle and Greenwood la have confirmed that modulating influences upon synaptic transmission of the rostral trigeminal nuclei are exerted via the nucleus caudalis and have thereby provided considerable support for this internuclear interaction proposal. However, previous failures~-,19 to identify substantial numbers of pulp-responsive neurons in NVCaud seriously limited the generality of this position by seemingly restricting its application to non-dental trigeminal pain. This restriction has been removed, at least for the present, by our demonstration that a sizeable population of pulp-responsive neurons does exist in NVCaud. Whether the cells possess all of the functional and anatomical characteristics required of the proposed modulating cells cannot, of course, be determined at this time. ACKNOWLEDGEMENTS

The authors express gratitude to Marguarite Agi, Cindy Bessmer, Rick Matte-

204 son a n d D a v i d R o l i n c e for technical assistance and to K a r e n J u n e for t y p i n g the manuscript. ]-he r e s e a r c h was s u p p o r t e d in p a r t by a g r a n t f r o m the Dr. Merrill H. R o s s M e d i c a l F u n d Inc. a n d by G r a n t s 5 S01 R R 0 5 4 0 2 - 1 3 , 5 T01 N S 0 5 6 0 5 , and NSII248-01AI

f r o m the N a t i o n a l

I R/)I

I n s t i t u t e o f N e u r o l o g i c a l Diseases and Stroke.

REFERENCES I ANDERSON, D. J., HANNUM, A. G., AND MATTHEW& B., Sensory mechanisms in mammalian teeth and their supporting structures, PhysioL Rev., 50 (1970) 171-195. 2 DAVIES, W. I. R., SCOTT, D., VESTERSTRON,K., AND VVKLICKY, L., Depolarization of the tooth pulp afferent terminals in the brain stern of the cat, J. Physiol. (Lond.), 218 (1971) 515--532. 3 DENNY-BROWN,D., AND YANAGISAWA,N., The function of the descending root of the fifth nerve, Brab,, 96 (1973) 783-814. 4 DUNKER, E., GRUBEL, G., UND V. REHREN, D., Dynamische eigenschaften afferenter trigeminusneuren in der medullar oblongata bei electrischer zahnreizung, Pftiigers. Arch. ges. Physiol., 296 (1967) 289-307. 5 GORDON, G., LAND6REN, S., AND SEED, W. A., The functional characteristics of single cells in the caudal part of the spinal nucleus of the trigeminal nerve of the cat, J. Physiol. (Lond.), 158 ( 1961 ) 544-559. 6 GREENWOOD, F., An electrophysiological study of the central connections of primary afferent nerve fibers from the dental pulp in the cat, Arch. oral. Biol., 18 (1973) 771-785. 7 KERR, F. W. L., KRW~ER, L., SCHWASSMANN,H. O., AND STERN, R., Somatotopic organization ot" mechanoreceptor units in the trigeminal nuclear complex of the macaque, J. comp. Neurol., 134 (1968) 127-143. 8 KRUC;ER,L., AND MICHEL, F., A morphological and somatotopic analysis of single unit activity in the trigeminal sensory complex of the cat, Exp. Neurol., 5 (1962) 13%156. 9 KRUGER, L., SIMINOFE,R., ANt) WITKOVSKV,P., Single neuron analysis of dorsal column nuclei and spinal nucleus of trigemina[ in cat, J. Neuro_physioL, 24 (1961) 333-349. l0 KUNC, Z., Significance of fresh anatomic data on spinal trigeminal tract for possibility of selective tractotomies. In R. S, KNIGHTONAND P. R. DUNKE(Eds.), Pain, Henry Ford Hosp. Intern. Syrup., Little, Brown, Boston, Mass., 1966, pp. 351-364. 1 l Mtzo(;ucm, K., The sites of action of morphine and the antagonistic action of levallorphan on the central nervous system of the dog, Folia pharmacol. Jap., 60 (1964) 326-346. 12 Mosso, J. A., AND KRUGER, L., Receptor categories represented in spinal trigeminal nucleus caudalis, J~ Neurophysiol., 36 (1973) 472-488. 13 NORD, S. G., AND KVLER, H. J., A single unit analysis of trigeminal projections to bulbar reticular nuclei of the rat, J. comp. Neurol., 134 (1968) 485-494. 14 NORD, S. G., AND ROSS, G. S., Responses of trigeminal units in the monkey bulhar lateral reticular formation to noxious and non-noxious stimulation of the face: experimental and theoretical considerations, Brain Research, 58 (1973) 385-399. 15 SESSLE, B. J., AND GREENWOOD, F., Influence of trigeminal nucleus caudalis on the responses of cat trigeminal brain stem neurons with orofacial and mechanoreceptive fields, Brain Research, 67 (1974) 330-333. 16 SICHER, H., (Ed.), Orhan's Oral Histolog.v and Embryology, 6th ed., Mosby, St. Louis, Mo.. 1972, p. 131. 17 SILVER, I. A., Other electrodes. In P. E. K. DONALDSON(Ed.), Electronic Apparatusfi)r Biological Research, Butterworth's Scientific Publ., London, 1958, pp. 568-581. 18 SJOQWST, O., Studies on pain conduction in the trigeminal nerve; a contribution to the surgical treatment of facial pain, Aeta payehiat, stand., Suppl. 17 (1938) 1-38. 19 TAMAROVA,Z. A., SHAPORALOV,A. E., AND VYKLICKY, L., Projections of tooth pulp afferents in the brain stem of rhesus monkey, Brain Research, 64 (1973) 442-445. 20 WAGERS,P, W., AND SMITH, C. M., Responses in dental nerves of dogs to tooth stimulation and the effects of systemically administered procaine, lidocaine and morphine, J~ PharmacoL exp. Ther., 130 (1960) 89-105. 21 YOUNG, R. F., AND KIN~, R. B., Excitability changes in trigeminal primary afferent fibers in response to noxious and non-noxious stimuli, J. Neurophysiol., 47 (1972) 87-95.