Raphe-induced suppression of the jaw-opening reflex and single neurons in trigeminal subnucleus oralis, and influence of naloxone and subnucleus caudalis

Pain, 10 (1981) 19--36 © Elsevier/North-Holland Biomedical Press

19

RAPHE-INDUCED SUPPRESSION OF T ~ JAW-OPENING REFLEX AND SINGLE NEURONS IN TRIGEMINAL SUBNUCLEUS ORALIS, AND INFLUENCE OF NALOXONE AND SUBNUCLEUS CAUDALIS

BARRY J. SESSLE and JAMES W. HU Division of Biological Sciences, Faculty of Dentistry, University of Toronto, Toronto, Ont. M5G 1G6 (Canada) (Received 26 February 1980, accepted 2 September 1980)

SUMMARY (1) The effects of stimulation of the nucleus raphe magnus (NRM) and the periaqueductal gray (PAG) were tested on the digastric (jaw-opening) reflex and on the activity of functionally identified single neurons recorded in trigeminal (V) subnucleus oralis in the brain stem. Reflex and neuronal responses evoked by tooth pulp stimulation could be readily suppressed for 250--1000 msec by PAG and NRM conditioning stimuli. The effects were not specific for tooth pulp afferent inputs, however, since suppression was also apparent in jaw-opening reflex responses evoked by low-intensity electrical or tactile stimulation of oral-facial sites, and in the mechanically or electrically evoked responses of oralis neurons with localized low-threshold mechanoreceptive fields. (2) The modulatory effects on the jaw-opening reflex and oralis neuron activity were not altered by reversible cold block of synaptic transmission in V subnucleus caudalis. Thus it appears that the PAG- and NRM-induced effects on the reflex and oralis neurons are not dependent on relays via caudalis. (3) Some of the suppressive influences on responses to oral-facial stimuli could be reversed by the administration of the opiate antagonist naloxone. This suggests that some of the modulatory influences involve endogenous opiate-related mechanisms. (4) Many of the oralis neurons were identified as trigeminothalamic relay neurons on the basis of their antidromic response to ventrobasal thalamic stimulation; PAG and NRM conditioningproduced not only a suppression of their orthodromic responses to oral-facial stimuli but also caused a decrease in the antidromic excitability of the relay neurons. This decrease may be indicative of raphe-induced postsynaptic inhibition of oralis neurons, and/or presynaptic .a~ili~tion of their thalamic endings.

20 INTRODUCTION R~,cent interest in the subject of modulation of sensory transmission by descending central neural influences has centered on centrifugal modulation of nociceptive transmission. In particular, the raphe system and associated structures (nucleus raphe magnus: NRM; periaqueductal gray: PAG) have been studied and implicated in endogenous opiate-related mechanisms of analgesia. For example, electrical stimulation of PAG or NRM, or microinjection of opiate-related substances into these structures, induces analgesia and suppressien of nociceptive spinothalamic neurons [2,4,23]. The effects can be partly overcome by the administration of the opiate antagonist naloxone, and it is generally thought that these intrinsic descending influences are opiate-related and specific or at least preferential for nociception [2,4,23]. In conformity with these findings, we [17,36] and others [1,22,45] have recently noted that PAG or NRM stimulation causes profound suppression of responses to noxious facial and tooth pulp stimuli in nociceptive neurons within trigeminal (V) subnucleus caudalis; this subnucleus is the caudalmost component of the V spinal tract nucleus and is considered the essential brain stem relay of orofacial pain [see refs. 7, 10, 20]. Some of these effects appear opiate-related since they can be reversed by the opiate antagonist naloxone [36]. In addition, we have, however, found that the response of the majority of low-threshold mechanoreceptive neurons in caudalis to non-noxious orofacial stimuli can also be suppressed by raphe conditioning [ 17,36], and suppression of responses to tactile stimuli has recently been reported for spinothalamic neurons [ 11,24,44]. A major aim of the present study was to determine if the responses of low-threshold mechanoreceptive neurons at more rostral levels of the V brain stem complex are also susceptible to descending raphe influences, and if the effects are opiate-related. The V brain stem complex can be rostrocaudally divided into the main (or principal) sensory nucleus, and the oralis, interpolaris and caudalis subdivisions of the V spinal tract nucleus [7,8,10,20]. In the present study, neurons in the subnucleus oraiis were specifically examined. Its neurons show many features in common with those of neurons in the adjacent V main sensory nucleus which is considered the lemniscal homolog of the V system [7,8,10,20]. For example, oralis neurons typically have "lemniscal-type" properties capable of precisely coding the spatiotemporal characteristics of low-threshold mechanical stimuli applied to their loc~ized ipsilateral mechanoreceptive fields [7,10]; a significant proportion of these neurons also project directly to the contralateral ventrobasal thalamus, while others connect to local brain stem reflex centers where, for example, they may be involved as excitatory interneurons for the digastric jaw-opening reflex [see ref. 10]. There is little evidence suggesting a cutaneous nociceptive afferent input to oralis, although some oralis neurons do respond to tooth pulp stimulation [19,25,35]. Thus, in view of these various characteristics, we could compare the effects of PAG and NRM conditioning stimulation on the responses of oralis neurons to

21 pulp or low-intensity orofacial stimuli, and on jaw-opening reflex activity evoked b y these stimuli, and also compare the effects with those we previously reported on caudalis nociceptive and non-nociceptive neurons. The second major aim of this study relates to the possible involvement of subnucleus caudalis in any observed modulatory effect of raphe stimulation on oralis neurons and on associated jaw-opening reflex activities. In contrast to subnucleus oralis which appears to lack cutaneous nociceptive neurons and a laminated morphology, subnucleus caudalis is a layered structure with nociceptive neurons concentrated in its laminae I and V. Its functional and morphological characteristics indicate that it is the V homolog of the spinal dorsal horn [7,8,10,12,20]. Moreover, it receives a direct raphe projection [see refs. 2, 22], is one of the major sites of concentration of enkephalin and opiate receptor sites [ 15,28,37], and has a direct ascending modulatory projection to oralis [12,13,17,20,42,46]. Thus it might be expected that any observed modulatory changes induced by PAG or NRM stimulation in oralis would be dependent on relays via caudalis. This possibility was tested by studying the modulatory effects before, during, and after reversible cold block of synaptic transmission in subnucleus caudalis. Some of the findings on PAG conditioning effects on oralis neurons have been briefly reported elsewhere [ 33,34]. METHODS

The methods employed have been described in detail previously [ 17,35] and will only be briefly outlined here. The study was carried out in 14 adult cats (2.5--4.0 kg) of which 12 were anesthetized with chloralose (60 mg/kg), and two decerebrated electrolytically at the mid.collicular level; blood pressure, % expired CO2 and rectal temperature were continuously monitored. The head was placed in a stereotaxic frame and the right brain stem exposed to allow the introduction of a tungsten microelectrode for recording the extracellular activity of single units in histologically verified sites in subnucleus oralis; for such recording:~ the animal was paralyzed and artificially ventilated. In some of the cats, single neuron activity was also recorded in V subnucleus caudalis, and the results of the caudalis studies are reported elsewhere [36]. Prior to brain stem recording, bipolar EMG electrodes were inserted into the ipsilateral anterior digastric muscle to record the effect of raphe conditioning (see below) on the digastric (jaw-opening) reflex evoked by tooth pulp or infraorbital nerve (IO) stimulation. Bipolar stimuli (0.01-5.0 mA, 0.1--0.2 msec) were delivered to the ipsilateral IO, facial skin, oral mucosa and maxillary or mandibular tooth pulp [35]; the facial skin was also stimulated by innocuous mechanical stimuli (air puffs, blunt probes) or by noxious mechanical stimuli delivered by serrated forceps. In addition, an electronically controlled mechanical stimulator [43] capable of delivering accurate and reproducible tactile stimuli was often employed to help delineate the mechanoreceptive field properties of the neurons or to evoke the digastric reflex.

22

Bipolar stimulating electrodes were stereotaxically placed [ 38] in the contralateral ventrobasal thalamus of most of the anesthetized animals at sites from which m a x i m u m field potential and neuronal activity could be evoked by facial tactile and IO stimuli. These electrodes were subsequently used for an~idrolnic activation of oralis neurons, utilizing criteria previously employed [32,35] to identify trigeminothalamic relay neurons. Bipolar concentric electrodes were also placed in the vicinity of the ipsilateral PAG in the 12 anesthetized cats and in NRM in 8 of the 14 cats (6 anesthetized, 2 decerebrate) to permit us to test for effects of conditioning stimulation of each structure on peripherally evoked responses. The electrodes were sited, under stereotaxic guidance (PAG: anterior: 0 to 1, lateral: 1 to 2, vertical: 0 to --2; NRM: posterior: 6.5 to 9, lateral: 0 to 1, vertical: --7 to --9), at the histologically verified loci in PAG and NRM that had the lowest threshold for producing suppression of the ipsilateral tooth pulpevoked digastric reflex [ 17,33]. Up to 3 penetrations were made in 0.5 m m lateral steps at both the PAG and NRM levels, and loci were tested at every 1--2 mm of vertical depth. A change in the electrode position by just 1--2 mm often led to a marked change in the threshold stimulus required for suppressing the digastric reflex, as we previously described [ 36]. The PAG conditioning stimulus was a single 200 msec train of 100 Hz constant-current pulses each of C.2 msec duration; NRM stimulus parameters were a 20 msec train of 400 Hz 0.2 msec pulses. We routinely tested the conditioning effect on near-threshold responses of the digastric reflex and single neurons to stimulation of IO, skin or tooth pulp. The threshold PAG and NRM conditioning stimulus intensities required for suppression (viz. complete inhibition) of peripherally evoked neuronal activity was also routinely noted. In testing for a conditioning effect on the digastric reflex response or on an oral~s neuron's response to a peripheral stimulus, a suprathreshold conditioning stimulus was applied at various time intervals preceding the peripheral {test) stimulus, as previously described [17,32,35]. For some trigeminothalamic relay neurons, conditioning effects were similarly tested on the antidromic response to the thalamic test stimulus. The time course of a conditioning effect was determined by comparing, at various conditioningtest intervals, the incidence of reflex or neuron excitation evoked by 10-20 test stimuli (delivered at 0.5--1 Hz) with the incidence elicited by a similar number of stimuli in the presence of the central conditioning stimulus. The influence of V subnucleus caudalis on any observed conditioning effect on digastric reflex or oralis neuron activity was assessed by noting the effect of block of caudalis. Caudalis block was achieved and studied, as previously described [13,27], b y the technique of reversible cold block of synaptic transmission in caudalis. This technique involves the circulation of warm ( 3 7 ° C ) o r cold (2 --5 ° C)al cohol through a silver thermode block positioned on the dorsal surface of the caudal medulla and upper cervical spinal cord overlying subnucleus cauda!is. The 3 mm wide thermode extended from a point 0--1 mm rostral to the obex to 11--12 mm caudal to the obex, and

23 ~----~ I I I

I I I

8O ,

o 60 r,,,,,

!

:

I

!

/ l

.:. I

2{

/

:

;/

:

'k',

0

_ /

/,_..._ /

200

/

/

jfl~

~

{~.----PAGCOND.TRAIN A. . . . PAGCOND.SI NGLE PULSE r

,

,

,

400

6 IX}

8{D

CONDIT iON ING -T E ST INTERVAL *

1O00

I~'

CONTROL

oco.o.. 1 1200

msecb

Fig. 1. Time courses of suppression of the tooth pulp-evoked digastric reflex by PAG c o n ditioning stimulation. The time course of suppression produced by a single PAG conditioning stimulus is shown, as well as that induced by a 200 msec train of 100 Hz pulses delivered to PAG. The time courses were determined as outlined in Methods. Note that the PAG conditioning train produced a longer-lasting inhibitory effect that had a duration of approximately 1000 msec. Moreover, the threshold for reflex suppression by the PAG train (60 /~A) was 10 times less than that required by the sing!e conditioning stimulus. The traces illustrate an example of the reflex suppression produced by the conditioning train; the top trace is 5 superimposed control responses to 5 successive tooth pulp stimuli delivered at 0.5 Hz, and the bottom trace shows the inhibitory effect on the reflex of a preceding PAG conditioning stimulus (COND.). Voltage calibration: 0.1 mV; time calibration: 10 msec.

had a thermistor incorporated in its contact surface with the medulla and spinal cord so that the thermode temperature could be continuously monitored. Previous experiments [ 13,27] with the thermode had established that the spread of cold block rostral to the thermode was less than 2 mm and that it was very effective in reversibly interrupting synaptic transmission in subnucleus caudalis. We confirmed this effectiveness by recording the field potential evoked in caudalis by IO stmmlation and noting the reversible blockade of postsynaptic activity in the field potential during cold block of caudalis. RESULTS

Digastric reflex In agreement with earlier studies [see ref. 10], tooth pulp or IO stimulation could evoke a consistent digastric reflex at a latency of 8--10 msec. The reflex was, however, susceptible to descending influences from both PAG and NRM," at histologically verified sites in these two regions that were cornparable to those we had earlier described as being effective in supp~'essing

24 the digastric reflex and caudalis neuron activity [36]. In both anesthetized and decerebrate preparations, a single train of PAG or NRM conditioning stimuli could suppress the digastric reflex elicited by t o o t h pulp stimulation (e.g. Fig. 1). The time course of reflex suppression had an onset of 10--20 msec, peaked at 40--50 msec, and usually lasted at least for 250 msec and sometimes as long as 1000 msec or more (Fig. 1); as Fig. 1 shows, single conditioning pulses to PAG or NRM also produced reflex suppression, but with a shorter duration. The digastric reflex could also be evoked by tactile stimulation of the canine tooth or low-intensity IO stimulation, in conformity with other studies [ 10,26,30,33,34]. We found that the reflex elicited by these stimuli could also be suppressed by conditioning stimulation at the same sites in PAG and NRM effective for suppressing the pulp-evoked reflex. The time course of suppression was similar to that for the pulp-elicited reflex. The threshold for NRM-induced suppression was also the same (mean 60/~A) for both the IO and pulp-evoked reflex; for the PAG-induced suppression, the threshold intensity required for the IO-evoked reflex was sometimes nearly two times higher than that required (mean 280/~A) for the reflex evoked by t o o t h pulp stimulation. .

Oralis neurons General characteristics.

A total of 92 oralis neurons was studied '.and they showed typical properties of oralis cells [7,10,35]. They had localized ipsilateral orofacial mechanoreceptive fields, and were arranged in an "inverted-head" somatotopic fashion with neurons having mandibular mechanoreceptive fields in the dorsal part of the subnucleus, ophthalmic field cells in the ventral part, and maxillary fields represented in between. The neurons also had short-latency excitatory inputs from the facial skin or IO, e.g., tt;ose with a localized infraorbital mechanoreceptive field could be activated by IO stimulation with a latency of 3.99 +- 1.73 msec (mean -+ S.D., n = 27). Fiftysix of the oralis cells responded to tooth pulp stimulation with a latency of 6.28 -+ 4.13 msec (e.g. Fig. 2); this pulp representation in oralis is in ag~:eement with earlier findings [19,25,35], although the proportion of pulpdriven neurons is higher and biassed towards these cells since we often used the pulp s~imuli as searching stimuli in the present study. None of ~;he neurons responded to noxious mechanical stimulation of the face. On the basis of antidromic activation, 22 of the oralis neurons were ider~tifled as trigeminothalamic relay neurens. Their mean antidromic latency was 1.35 +- 0.62 msec. An example of one of these neurons is shown in Fig. 3. We did not systematically test every neuron to determine if it could be orthodromically or antidromically activated from PAG or NRM since we were primarily interested in this study in the suppressive effects of raphe conditiening stimulation~ Nonetheless, it was apparent that such stimulation could induce transsynaptic and/or antidromic activity in a number ..gf oralis neurons. For example, 10 neurons could be orthodromically activated by PAG or NRM stimulation at latencies ranging between 3 and 8 msec; two °~

25 49 ~

.......... o

30 ,d t~

it

l

i

W

0

i

,, 20 ...I

.¥

I-,.-

0

,--- I0

~. /

P,

i/

,

~

0 ........ CONTROL

il

CONTROL

, . . . . PAGCOND. %7 . . . . "NRM COND.

//• ~-,,Ifl

I

" . ~ . / / w "'~z

I

I

I

I//4

0.5 1.0 1.5 2.0" STIMULUS INTENSITY ( mA

,

,

~ . , PAG CONI).

I .0

5.0

Fig. 2. Effects of FAG and NRM conditioning stimuli on the stimulus-response relationship of an oralis neuron. This neuron, with the localized infraorbital mechanoreceptive field shown, could be activated by maxillary t o o t h pulp stimulation as well as by light tactile stimulation of its mechanoreceptive field; some of its other features are illustrated in Fig. 5B. The traces show 5 superimposed control responses to 5 successive pulp test stimuli delivered at 1 Hz (top) and the inhibitory effect of a preceding PAG conditioning (COND.) stimulus ( b o t t o m ) . The graph illustrates the neuron's stimulus-response relationship (o) to graded intensities of tooth pulp stimulation, and the inhibitory effects on this relationship of a PAG conditioning stimulus set at 130 pA (o) and a NRM conditioning stimulus set at 60 # A (v); the conditioning-test intervals were kept at 50 msec. Each point on the graph represents, for each intensity of the pulp stimulus, the total n u m b e r of spikes elicited bF 10 successive pulp stimuli. Note the shift of the stimulus-response relationship down and to the right, indicating depression of both the threshold and suprathreshold responsiveness of the neuron. Voltage calibration: 0.1 mV; time calibration: 4 mgec.

could be antidromically activated. Most of the 10 neurons had orofacial afferent inputs, but an orofaci~d input could not be found in three. The orthodromic responses indicate a direct raphe projection to oralis, while antidromic activation suggests an oralis projection to FAC or NRM. It is also noteworthy that in addition to orthodromic activation, five of the 10 neurons showed the long-lasting PAG- or NRM-induced suppression of orofacial afferent inputs that was a feature of this study (see below}. Conditioning effec~s. The responses to electrical or mechanical stimulation of oral-facial sites could be suppressed by a single train of PAG or NRM conditioning stimuli at threshold intensities comparable to those effective in suppressing the digastric reflex. The thresholds remained stable for many hours, and neurons with or without a pulp afferent input were susceptible to PAG and NRM modulation. Conditioning stimulation of PAG produced suppression of pulp-evoked responses in 39 of 40 neuro_-~s tested (e.g. Fig. 2). Responses to low-intensity IO or skin electrical stimuIation were suppressed in 29 of 41 neurons tested; 17 of these 41 neurons had a pulp afferent input,

26

÷=

!

/

i

,

....

TACTILECONTROL

|

_

L |

--

_~.

THAL. CONTROL

THAL. lOOHz

| •

PAG COND.

PA6 COND. Fig. 3. Effects of PAG conditioning stimulation on a low-threshold mechanoreceptive trigeminothalamic neuron. The neuron's antidromic response e v o k e d by a 100 Hz threshold stimulus to the thalamus (THAL.) is shown on the e x t r e m e right. The neuron could be orthodromically activated only by light tactile stimulation or electrical stimulation of its localized infraorbital mechanoreceptive field illustrated in the face figurine. Five superimposed responses to 5 successive tactile stimuli (450 pm, 40 msec) are shown along with the electronic analogue of the mechanical stimulus, as well as the suppressive effect of a preceding PAG conditioning (COND.) stimulus. Also shown are 5 superimposed thalamic control responses and the suppressive effect of PAG conditioning on this antidromic activity. Increasing by less than 100 pA the thalamic stimulus intensity could overcome this PAG-induced suppressive effect (not illustrated). Naloxone (0.4 mg/kg) was administered while the activity of this neuron was recorded, and failed to reverse the PAGinduced suppression of either the tactile or antidromic evoked activity. Voltage calibration: 0.4 mV; time calibration: 10 msec (tactile), 4 msec (THAL.).

and 16 of the 17 showed inhibition of their IO or skin-evoked response. Likewise, NRM conditioning inhibited IO or skin-evoked responses in all 10 neurons tested {only one of these 10 had a pulp afferent input} and pulpelicited responses in all 7 neurons tested. These inhibitory effects had a time course comparable to that outlined above for reflex suppression. Moreover, as we had previously noted [36] for caudalis neurons, progressively increasing the intensity of the pulp or facial test stimulus could gradually reduce the suppressive effect of the conditioning stimulus. Fig. 2 shows an example of these effects. Not only were responses to electrical stimulation of IO, skin or tooth pulp suppressed, but the responses of oralis neurons to electronically controlled tactile stimuli delivered to their orofacial mechanoreceptive field could be suppressed, as Fig. 3 illustrates. Suppression of responses evoked by mechanical test stimuli was seen in 24 of 30 neurons examined. Seven of these neurons had a pulp afferent input, and inhibition of their responses to tactile stimuli occurred in all seven. As we had noted for reflex and oralis responses to pulp or IO stimuli, the inhibition lasted for 200 msec or more, and could be reduced by an increase in test stimulus intensity. The antidromically evoked activity of oralis neurons, as well as their orthodromically evoked activity, was also susceptible to descending raphe influences. A decrease in antidromic excitability could be produced in 13 of 17 trigeminothalamic relay neurons tested; 2 of these 13 had a pulp afferent input in addition to a low-threshold mechanoreceptive input. Although we rou~inely tested at a co~ditioning-test interval of 50 msec, in six of these

27

neurons the time course of this effect was determined. In five the decreased excitability became apparent at 20--40 msec and lasted for 200--300 msec; in the other neuron, which could be excited by pulp as well as by facial tactile stim~dation, the effect lasted only 100 msec. An example of a decrease in antidromic excitability produced by PAG conditioning is seen in the mechanosensitive trigeminothalamic neuron illustrated in Fig. 3. Six of the relay neurons showing a decreased antidromic responsiveness as a result of raphe conditioning also were tested to determine if the control level of antidromic excitability could be restored during conditioning by a small (0.05--0.2 mA) increase in the intensity of the thalamic stimulus that antidromically excited the neuron. In five of the neurons, the antidromic excitability could be restored, but in the remai:dng neuron tested, the antidromic excitability could n o t be restored by this maneuver. This particular neuron was the same pulp-driven neuron in which the decreased antidromic excitability lasted for only 100 msec (see above}. These findings suggest that a postsynaptic inhibitory mechanism induced by raphe stimulation may have been operating on this neuron, as opposed to possible presynaptic influences acting on the other 5 neurons (see Discussion). Effect of naloxone. In view of the stability of the conditioning influences, the effect of the administration of the opiate antagonist naloxone (0.4

4

8 12 TIME ~ minutes) NALOXONE ( 0. 4mg/kg t

16

20

24

Fig. 4. Effect of the intravenous administration of naloxone on the PAG-induced suppression of an oralis neuron's response to maxillary tooth pulp stimulation. This neuron had a mechanoreceptive field localized to the lateral aspect of the vibrissa pad on the upper lip. The traces show 5 superimposed control responses of the neuron to 5 successive pulp test stimuli delivered at 1 Hz (top) and the inhibitory effect on these responses of a preceding PAG conditioning (COND.) stimulus (bottom). The effect of naloxone (0.4 mg/kg) on this suppression was tested by setting the test and conditioning stimuli at levels just sufficient for inducing a clear and consistent inhibitory effect, and then noting if the inhibitory effect was altered after naloxone. Periodically over more than 20 rain, 5--10 control sequences and a similar number of conditioning sequences were examined, and the spike discharge of the neuron was expressed as a percentage of the control level of responsiveness. Note that reversal of the suppressive effect of PAG conditioning occurred, first becoming apparent between 3 and 6 min and lasting for approximately 15 rain. Voltage calibration: 0.1 mV, time calibration: 19 msec.

28

mg/kg, i.v.) could readily be tested on the conditioning influences over a period of 20--40 min for each neuron. Twelve oralis neurons showing PAGinduced suppression were tested with naloxone. Eight responded to pulp stimulation as well as to low-intensity tactile stimulation of their localized orofacial mechanoreceptive field, and the remaining four responded to tactile stimuli only; 6 of the 12 neurons were trigeminothalamic relay neurons. Naloxone produced a partial reversal of the PAG-induced suppression of pulp-evoked activity in 7 of 8 of the pulp-driven neurons tested. Fig. 4 shows an example. The reversal commenced 3--5 rain after naloxone administration and lasted for 15--25 min. In 10 of the 12 neurons, naloxone was tested on the suppression of their responses to low-intensity IO or tactile stimulation of their mechanoreceptive field. Only two of these neurons showed evidence of reversal with naloxone; both responded also to pulp stimulation, and naloxone produced a partial reversal of the PAG-induced suppression of their pulp-driven activity as well as their facial-evoked responses. The decreased antidromic excitability that was a feature of conditioning stimulation in trigeminothalamic relay neurons (see above) could not be reversed by naloxone in 3 neurons tested. Effects of cold block of caudalis. To determine if the PAG- or NRMinduced suppression of the digastric reflex and oralis neurons might be dependent on relays via subnucleus caudalis, reversible cold block of synaptic transmission in caudalis was employed and tested on reflex and oralis cell suppression in both anesthetized and decerebrate cats. In accordance with earlier f i n d ~ [ 30], we noted that reversible cold block of synaptic transmission could result in a depression of the IO-evoked reflex and sometimes the pulp-evoked reflex. This ~upports the concept [see refs. 13, 17, 46] that caudalis is exerting an ascending tonic facilitatory influence on neurons in V subnucleus oralis since oralis is thought to be the major site of the excitatory interneurons for the digastric reflex [ 10,40]. We then tested the effect of cold block of caudalis on the PAG- and NRMinduced depression of the reflex. Before, during, and after cold block of caudalis, the PAG and NRM stimulus intensities were kept constant at their threshold level for inhibiting the IO or pulp-evoked reflex. The IO or tooth pulp intensities were kept just suprathreshold for evoking a consistent digastric reflex. Even though it was often necessary during cold block to increase the intensity of the test IO or pulp stimulus evoking the reflex (because caudalis cold block removed the ascending tonic influence on oralis neurons), PAG- and NRM-induced depression of the reflex still remained during caudalis block (Fig. 5A). The retention of PAG- and NRM-induced effects was also confirmed after a V tractotomy at the level of the obex. Our findings that digastric reflex suppression is unaffected by cold block of caudalis were complemented by similar findings for oralis neurons. Fifteen neurons showing PAG- and/or NRM-induced suppression were examined. The effect of cold block was tested on the response of eight of them to IO stimulation, and on the response of eight to tooth pulp stimulation. In con-

29 A

38oc

CONTROL

J1 - "

N R M COND .

.

PAG C O N D .

5oc

.

.

.

,--~_

.

-

38~

_A ~

,

~

,

.

.

-_

m

,

,

.

.

_

.Tj .. ~_

m

-

.

.

.

,

,

. .....

.

J B CONTROL

~

i

~

NRM COND.

PAG COND.

"r -~ " -~ "

]r---

- ~

-

,

t

,

~

Fig. 5. Effect of reversible cold block of subnucleus caudalis on the PAG- and NRMinduced suppression of digastric reflex (A) and oralis neuron (B) activity evoked by maxillary t o o t h pulp stimulation. Prior to cooling (38°C), the pulp, PAG and NRM stimuli were set at intensity levels just sufficient to induce clear and consistent suppression as shown; the conditioning-test interval was kept at 50 msec. Note that during cold block (5°C), the control reflex responses to the pulp test stimuli illustrated in A showed little change, and no change was also apparent in the suppressive effects of PAG or NRM conditioning (COND.) stimuli. Likewise, for the neuronal responses to pulp stimulation (B), pulp stimulation evoked 10--15 spikes before (38°C), duri,lg (5°C) and following (38°C) cold block, and no change was apparent in the PAG- ana NRM-induced suppressive effects during cold block; these data are from the same neuron shown in Fig. 2. Note that during PAG conditioning, another neuron with a smaller amplitude spike was sometimes excited as a result of the preceding PAG stimulation. Each trace in A and B illustrates 5 superimposed responses to 5 successive pulp stimuli delivered at 1 Hz. Voltage calibration: 0.2 mV (A), 0.1 mV (B); time calibration: 10 msec (A), 4 msec (B).

formity with earlier findings in our laboratory [13], we noted that the responses of some of these neuron~ to IO or pulp stimulation could be depressed during cold block of caudalis. This again is a reflection of the removal of the tonic ascending influence that caudalis e x e c s on oralis neurons. However, we found no alteration in the threshold suppressive effects of PAG or NRM conditioning in all 15 oralis neurons tested. Fig. 5B shows examples in one of these neurons. DISCUSSION

Our studies have revealed that digastric reflex activity and the activity of neurons in V subnucleus oralis elicited by tooth pulp or low-intensity orofacial s~imuli are susceptible to considerable modulation by descending influences from PAG and NRM. The observation of raphe-induced suppression of the digastric (jaw-opening) reflex concurs with simil~ findings in this [14,

30 17,33] and other [1,22,26] laboratories. With respect to oralis neurons, we had previously briefly noted that oralis neurons can be suppressed by PAG conditioning stimulation [33], and NRM-induced inhibition of oralis neurons has also been reported by others [22]. In the present study we carefully characterized the properties of these neurons and have shown that the oralis neurons involved have the "lemniscal-type" properties typical of subnucleus oralis [7,8,10]. Many of the neurons also responded to tooth pulp stimulation, but this was not an unexpected finding in view of the welldocumented pulp afferent input to a large proportion of oralis neurons [ 16, 19,25,35]. The mechanoreceptive field properties of the oralis neurons were consistent with their role in the precise spatiotemporal coding of tactile stimuli delivered to their localized orofacial mechanoreceptive fields [7,8~ 10], and many were involved in relaying this information directly to the contralateral thalamus, in agreement with earlier findings of antidromic activation of oralis neurons [7,32,35]. The majority of oralis neurons do not, however, project to the thalamus, but many project to adjacent brain stem regions where for example they may be involved as excitatory interneurons in craniofacial mo t o r activities such as the digastric reflex [10,40]. Thus it seems likely that the suppressive effects we noted on the digastric reflex could have reflected the PAG- and NRM-induced suppression occurring on some of our sample of oralis neurons. In each cat, the PAG and NRM stimulating electrodes were located at the optimal sites for suppressing the tooth pulp-evoked digastric reflex, and the subsequent suppression of oralis neurons reflected the effects of stimulation at these sites. Since a change in the position of the PAG or NRM concentric electrode b:' only 1--2 mm could often produce a marked decrease in the effectivenes, of the stimulus, in accordance with our earlier observations [36], it is unlikely that the reflex or neuronal suppression would have resulted from stimulus spread from our histologically verified sites in PAG and NRM. This is supported by other reports, including our own with digastric reflex and caudalis neuron suppression [36], of the localized effects of PAG and NRM stimulation with bipolar constant-current pulses delivered by concentric electrodes [ 6,24,26,44]. We do not wish to infer from this that these sites constitute the only loci in the vicinity of PAG and NRM that can modulate the digastric reflex and V brain stem neuron activity. Other areas such as the nuclei reticularis magnocellularis and gigantocellularis, and the lateral reticular formation, also constitute important potential modulatory influences [ 2,4,23,24,36 ]. Some of our current studies ~xe presently addressing this possibility and tentatively point to the involvement of a number of brain stem sites in the modulation of digastric reflex and V neuron activity [41]. Subnucleus caudalis might reasonably be expected to play an important role in the observed modulatory influences in oralis and associated digastric reflex activity, for the following reasons: caudalis is the essential brain stem link necessary for the appreciation of orofacial pain and has structural and functional properties suggesting a homology with the spirial dorsal horn

31 [7,8,10,12,20]; it receives a direct projection from NRM [2,22] which, as indicated earlier, is implicated in endogenous mechanisms of analgesia; enkephalin and opiate receptor sites are highly concentrated in caudalis [15,28,37]; and caudalis sends a direct ascending tonic modulatory projection to oralis [ 12,13,17,20,42,46]. Our studies with reversible cold block of caudalis confirmed the tonic modulatory influence from caudalis, but failed to find any evidence that caudalis influences the PAG- and NRM-induced effects on both oralis neurons or the digastric reflex. These findings suggest that the modulatory effects exerted by PAG and NRM on oralis neurons and on associated digastric reflex activities are borne over relatively direct pathways from PAG and NRM to oralis, and are not dependent on a relay in subnucleus caudalis. We cannot rule out a contribution from a relay via caudalis, but its influeace would be minor. Our findings receive support from recent electrophysiological and anatomical observations [2,22] of a direct projection from NRM to subnucleus oralis, and from findings [ 41] that cold block of caudalis causes no alteration in the presynaptic excitability changes induced in the oralis endings of tooth pulp afferents by PAG and NRM stimuli. Thus it is possible that at least the NRM-induced suppression results directly from the action of raphe-oralis inhibitory fibers; the PAG-induced effects are probably relayed to oralis via the descending path linking PAG with NRM. Despite the evidence for direct projections to oralis from NRM, we cannot, '.mwever, dismiss the other possibility that the inhibitory neurons are not located in NRM but that they occur in or adjacent to oralis. The oralis neurons that can be transsynaptically excited by PAG and NRM stimuli should be examined to see if they serve as inhibitory interneurons that mediate the r:tphe-induced modulation on adjacent oralis neurons. The suppressive effects of PAG and NRM conditioning stimulation on oralis neurons have similar features to the suppression that they exert on the activity of neurons in V subnucleus caudalis [17,33,36]. There we found, under comparable experimental conditions to those used in the present study, that PAG and NRM also exert powerful and prolonged suppressive influences on the responses of neurons to tooth pulp and tactile orofac~al stimuli (as well as on responses to noxious facial stimulation). These effects in caudalis, as in oralis, could be partially or totally overcome by an increase in the pulp or orofacial test stimulus exciting the neurons. The stimulusresponse relationship of the oralis and caudalis neurons is shifted down and to the right during PAG or NRM conditioning (e.g. Fig. 2), indicating that both the threshold as well as the suprathreshold responsiveness of the neurons are being influenced by PAG and NRM stimuli. This feature of rapheinduced inhibition in the V brain stem subnuclei contrasts with that recently reported to occur on spinal dorsal horn neurons [ 6]. The threshold responses of these dorsal horn neurons to noxious cutaneous heat was unaffected; only the suprathreshold responses could be suppressed by stimulation of PAG and adjacent "structures. Whether these variations reflect a real functional difference between the spinal dorsal horn and the V brain stem subnuclei, or can

32

be explained by differences in experimental preparation, procedures, test stimuli utilized, etc., remains to be determined. The similarities between the PAG and NRM effects in oralis and caudalis also warrant further c o m m e n t in view of our findings in both subnuclei that PAG and NRM stimulation can suppress not only neuron responses to noxious facial or t o o t h pulp stimuli, but also the activity of the majority of non-nociceptive (low-threshold mechanoreceptive) neurons evoked by tactile or low-intensity electrical stimulation of their localized mechanoreceptive fields. Lovick and Wolstencroft [ 22] have recently reported that NRM is less effective in suppressing the responses of an unspecified number of oralis or caudalis neurons to low-intensity IO or facial hair stimuli than responses to tooth pulp stimulation, but they did n o t fully characterize the functional properties of their neurons or indicate the relative incidence of suppression. We have noted that the incidence of suppression of responses to low-intensity electrical or tactile facial stimuli is higher in oralis neurons with a pulp afferent input, and in caudalis neurons with a pulp or nociceptive facial input [33,36]. Nonetheless, the majority of neurons showing only a lowthreshold rnechanoreceptive input can be suppressed by raphe conditioning. Our findings ~ e compatible with recent observations in the spinal cord [11, 24,44] and dorsal column nuclei [9] where it has been shown that neuronal responses to mechanical stimuli can be suppressed by PAG or NRM stimuli. Thus, despite the emphasis on the involvement of PAG and NRM in endogenous mechanisms of suppression of nociception, it is now clear that the descending raphe influences are not specific for nociceptive transmission. Transmission of tactile as well as nociceptive information can be modulated. It is also noteworthy that visceral afferent transmission can in addition be affected since PAG and NRM conditioning suppresses evoked or spontaneous neuronal activity within the solitary tract nucleus as well as associated respiratory activity and upper respiratory tract reflexes [ 31]. The mechanisms by which these raphe-induced suppressive influences are produced are still unclear. The release of endogenous opiate-related substances (e.g. enkephalin) that presynaptically inhibit nociceptive afferent inputs to V or spinal dorsal horn neurons has been recently implicated [2, 18,39]. Our results provide some support for an opiate-related mechanism, since the opiate antagonist naloxone could reverse the suppressive effects of PAG conditioning of most pulp-evoked responses and some facially elicited responses. However, nalox~ne was often ineffective, especially on suppression of IO or tactile evoked responses. These and comparable findings in subnucleus cmldalis [36] and the spinal dorsal horn [5] point to the presence of multiple descending influences from the raphe and associated structures on sensory transmission; some of these influences may be opiaterelated, while otbers may operate by other mechanisms. It was not the purpose of the present experiments with naloxone to resolve the question of whether or not the raphe acts by opiate-related presynaptic mechanisms. However, the view for a presynaptic action has not received support from other recent findings by us in the V brain stem sub-

33

nuclei [ 16]. Primary afferent depolarization, which is believed to reflect presynaptic inhibition [3,29], can be produced by PAG or NRM conditioning in the oralis or caudalis endings of tooth pulp afferents, but the effect is not reversed by the opiate antagonist naloxone [16]; this finding has been confirmed by Lo,Sck et al. [21]. Nonetheless, our observations of a rapheinduced decrease in the antidromic as well as the orthodromic responsiveness of oralis neurons are of relevance since the decrease in antidromic excitability could be taken as evidence in favor of at least a postsynaptic mechanism of action of the raphe influences. This view is supported by our observation that in a small proportion of these neurons tested the decrease could not be overcome by increasing the thalamic stimulus intensity. We did note, however, in most of our sample of trigeminothalamic relay neurons that a small increase in the intensity of the thalamic stimulus evoking the antidromic response could indeed restore the excitability of the neuron during raphe conditioning. Such an effect has previously been noted [32] with facial or cortical conditioning of the antidromic excitability of trigeminothalamic relay neurons in the V main sensory nucleus. As Sessle and Dubner [32] have argued, this probably is indicative of presynaptic hyperpolarization of the thalamic endings of the trigeminothalamic relay neurons rather than postsynaptic inhibitory mechanisms blocking antidromic invasion of the neurons. Since presynaptic hyperpolarization might be a reflection of presynaptic facilitation [3,29,32], these observations suggest the possibility of an ascending presynaptic facilitatory effect from the raphe primarily on the mechanoreceptive afferent input to the ventrobasal thalamus, in addition to its descending inhibitory effects on V and spinal dorsal horn neurons. ACKNOWLEDGEMENTS

We acknowledge with thanks the technical assistance of R. Egizii and K. MacLeod, the art and photographic services of R. Bauer, S. Burany and M. Heam, and the secretarial services of D. Tsang. We are also grateful to Dr. J.O. Dostrovsky for his comments on an earlier draft of this paper. This research was supported by Grant I R01 DE04786 awarded by the U.S. National Institute of Dental Research, DHEW. Naloxone was a gift from Endo Laboratories Inc., Garden City, N.Y. 11530, U.S.A. REFERENCES 1 Andersen, R.K., Lund, J.P. and Puil, E., Excitation and inhibition of neurons in the trigeminal nucleus caudalis following periaqueductal gray stimulation, Canad. J. Physiol. Pharmacol., 56 (1978) 157--161. 2 Basbaum, A.I. and Fields, H.L., Endogenous pain control mechanisms: review and hypothesis, Ann. Neurol., 4 (1978) 451--462. 3 Burke, R.E. and Rudomin, P., Spinal neurons and synapses. In: Handbook of Physiology -- The Nervous System, Sect. 1, Vol. 1, American Physiological Society, Bethesda, Md., 1977, pp. 877--944. 4 Cannon, J.T. and Liebeskind, J.C., Descending control systems. In: R.F. Beers, Jr.

34

5 6 7 8 9 10 11 12

13

14 15 16 17 18 19 20 21

22

23

24

and E.G. Bassett (Eds.), Mechanisms of Pain and Analgesia Compounds, Raven Press, New York, 1979, pp. 171--184. Carstens, E., Klumpp, D. and Zimmermann, M., The opiate antagonist, naloxone, does not affect descending inhibition from midbrain of nociceptive spinal neuronal discharges in the cat, Neurosci. Lett., 11 (1979) 323--327. Carstens, E., Yokota, T. and Zimmermann, M., Inhibition of spinal neuronal responses to noxious skin heating by stimulation of mesencephalic periaqueductal gray in the cat, J. Neurophysiol., 42 (1979) 558--568. Darian-Smith, I., Neural mechanisms of facial sensation, Int. Rev. Neurobiol., 9 (1966) 301--395. Darian-Smith, I., The trigeminal system. In: A. Iggo (Ed.), Handbook of Sensory Physiology, Vol. 2, Somatosensory System, Springer, Berlin, 1973, pp. 271--314. Dostrovsky, J.O., Inhibition of innocuous tactile-evoked responses in cat dorsal column nucleus and trigeminal nucleus by stimulation of brainstem periaqueductal gray and nucleus raphe magnus, Canad. Physiol., 10 (1979) 157. Dubner, R., Sessle, B.J. and Storey, A.T., The Neural Basis of Oral and Facial Function, Plenum, New York, 1978. Duggan, A.W. and Griersmith, B.T., Inhibition of the spinal transmission of nociceptire information by supraspinal stimulation in the cat, Pain, 6 (1979) 149--161. Gobel, S., Falls, W.M. and Hockfield, S., The division of the dorsal and ventral horns of the mammalian caudal medulla into eight layers using anatomical criteria. In: D.J. Anderson and B. Matthews (Eds.), Pain in the Trigeminal Region, Elsevier/NorthHolland, Amsterdam, 1977, pp. 443--453. Greenwood, L.F. and Sessle, B.J., Inputs to trigeminal brain stem neurons from facial, oral, tooth pulp and pharyngolaryngeal tissues. II. Role of trigeminal nucleus caudalis in modulating responses to innocuous and noxious stimuli, Brain Res., ]17 (1976) 227--238. Henry, a.L., Sessle, B.J., Lucier, G.E. and Hu, J.W., Effects of substance P on nociceptive trigeminal brain stem neurons, Pain, 8 (i980) 33--46. HSkfelt, T., Ljungdahl, 4., Terenius, L., Elde, R. and Ni~sson, G., Immunohistochemical analysis of peptide pathways possibly related to pain and analgesia: enkephalin and substance P, Proc. nat. Acad. Sci. (Wash.), 74 (1977 t 3081--3085. Hu, J.W., Dostrovsky, J.O. and Sessle, B.J., Primary afferent depolarisation of tooth pulp afferents is not affected by naloxone, Nature (Lond.), 276 (1978) 283--284. Hu, J.W. and Sessle, B.J., Trigeminal nociceptive and ncn-nociceptive neurons: brainstem intranuclear projections and modulation by orofacial, periaqueductal gray and nucleus raphe magnus stimuli, Brain Res., 170 (1979) 547--552. Iversen, L.L., The chemistry of the brain. Sci. Amer., 24! (1979) 134--149. Khayyat, G.F., Yu, Y.J. and King, R.B., Respo~se patterns to noxious and nonnoxious stimuli in rostral trigeminal relay nuclei, Brain ~,es., 97 (1975) 47--60. Kruger, L., Functional subdivision of the brain stem sensory trigeminal nucleus complex. In: J.J. Bonica, J.C. Liebeskind and D.G. Albe-Fes3ard (Eds.), Advances in Pain Research and Therapy, Vol. 3, Raven, New York, 1979, UP. 197--209. Lovick, T.A., West, D.C. and Wolstencroft, J.H., A pres~ ,laptic action of the raphe on tooth pulp fibre terminals: is this mediated by an opioid peptide? In: J.M. Van Ree and L. Terenius (Eds.), Characteristics and Function of Opioids, Elsevier/North-Holland, Amsterdam, 1978, pp. 175--177. Lovick, T.A. and Wolstencroft, J.H., Inhibitory effects of nucleus raphe magnus on neuronal responses in the spinal trigeminal nucleus to nociceptive compared with nonnociceptive inputs, Pain, 7 (1979) 135--145. Mayer, D.J., Endogenous analgesia systems: neural and behavioral mechanisms. In: J.J. Bonica, J.C. Liebeskind and D.G. Albe-Fessard (Eds.), Advances in Pain Research and Therapy, Vol. 3, Raven, New York~ 1979, pp. 385--410. McCreery, D.B., Bloedel, J.R. and Haines, E.G., Effects of stimulating in raphe nucleus and in reticular formation on response of spinothalamic neurons to mechani-

35 cal stimuli, J. Neurophyslol., 42 (1979) 16(~ 182. 25 Nord, S.G., Responses of neurons in rostral ::~d caudal trigeminal nuclei to tooth pulp stimulation, Brain Res. Bull., 1 (1976) 489-=:92. 26 Oliveras, J.L., Guilbaud, G. and Besson, J~. ~., The use of tooth-pulp stimulation to evaluate the analgesia induced by stimulat~n of raphe nuclei in the cat. In: D.J. Anderson and B. Matthews (Eds.), Pain in th~. Trigeminal Region, Elsevier/North-Holland, Amsterdam, 1977, pp. 319--330. 27 Rowe, M.J. and Sessle, B.J., Somatic afferent input to posterior thalam~c neurones and their axon projection to the cerebral cortex in the cat, J. Physiol. (Lond.), 196 (1968) 1 9 - 3 5 . 28 Sat, M., Stumpf, W.E., Miller, R.J., Chang, K.-J. and Cuatrecasas, P., Immunohistochemical localization of enkephalin in rat brain and spinal cord, J. comp. Neurol., 182 (1978) 17--39. 29 Schmidt, R.F., Presynaptic inhibition in the vertebrate central nervous system, Ergebn. Physiol., 63 (1971) 20--101. 30 Sessle, B.J., Modulation of alpha and gamma trigeminal motoneurons by various peripheral stimuli, Exp. Neurol., 54 (1977) 323--339. 31 Sessle, B.J., Ball, G.J. and Lucier, G.E., Endogenous opiate-related influences from periaqueductal gray and nucleus raphe magnus on solitary tract neurons and respiratory-related functions, Neurosci. Abstr., 5 (1979) 616 (Abstract 2096). 32 Sessle, B.J. and Dubner, R., Presynaptic excitability changes of trigeminothalamic and corticothalamic afferents, Exp. Neurol., 30 (1971) 121--125. 33 Sessle, B.J., Dubner, R., Greenwood, L.F. and Lucier, G.E., Descending influences of periaqueductal gray matter and somatosensory cerebral cortex on neurones in trigeminal brain stem nuclei, Canad. J. Physiol. Pharmacol., 54 (1976) 66--69. 34 Sessle, B.J., Dubner, R., Hu, J.W. and Lucier, G.E., Modulation of trigeminothalamic relay and non-relay neurones by noxious, tactile and periaqueductal gray stimuli: implications in perceptual and reflex aspects of nociception. In: D.J. Anderson and B. Matthews (Eds.), Pain in the Trigeminal Region, Elsevier/North-Holland, Amsterdam, 1977, pp. 285--294. 35 Sessle, B.J. and Greenwood, L.F., Inputs to trigeminal brain stem neurones from facial, oral, tooth pulp and pharyngolaryngeal tissues. I. Fesponses to innocuous and noxious stimuli, Brain Res., 117 (1976) 211--226. 36 Sessle, B.J., Hu, J.W., Dubner, R. and Lucier, G.E., Functional properties of neurons in cat trigeminal subnucleus caudalis medullary dorsal horn. II. Modulation of responses to noxious and non-noxious stimuli by periaqueductal gray, nucleus raphe magnus, cerebral cortex and afferent influences and effect of naloxone, J. Neurophysiol., 45 (1981) 211--225. 37 Simantov, R., Kuhar, M.J., Uhl, G.R. and Snyder, S.H., Opioid peptide enkephal]ns: immunohistochemical mapping in the rat central nervous system, Proc. nat. Acad. Sci. (Wash.), 74 (1977) 2167--2171. 38 Snider, R.S. and Niemer, W.T., A Stereotaxic Atlas of the Cat Brain, University of Chicago Press, Chicago, Ill., 1961. 39 Snyder, S.H., Neurotransmitter activity in the brain: focus on the opiate receptor, Interdisc. Sci. Rev., 3 (1978) 45--54. 40 Sumino, R., Central neural pathways involved in the jaw opening reflex in the cat. In: R. Dubner and Y. Kawamura (Eds.), Oral-Facial Sensory and Motor Mechanisms, Appleton-Century-Crofts, New York, 1971, pp. 315--322. 41 Sumino, R., Hu, J.W., Dostrovsky, J.O. and Sessle, B.J., Trigeminal primary afferent depolarization and reflex inhibition produced by periaqueductal gray and raphe magnus stimuli: influence of nucleus caudalis and effectiveness of other brainstem sites, Neurosci. Abstr., 5 (1979) 616 (Abstract 2098). 42 Tiwari, R.K. and King, R.B., Fiber projections from trigeminal nucleus caudalis in primate (squirrel monkey and baboon), J. comp. Neur01., 158 (1974) 191--206.

36

43 Wermr, G. and Mountcastle, V.B., Neural activity in mechanoreceptive cutaneous afferents: stimulus-response relation, Weber functions, and information transmission, J. Neurophysiol., 28 (1965) 359--397. 44 Willis, W.D., Haber, L.H. and Martin, R.F., Inhibition of spinothalamic tract cells and interneurons by b r ~ ~tem stimulation in the monkey, J. Neurophysiol., 40 (1977) 968--981. 45 Yokota, T. and Hashimoto, S., Periaqueductal gray and tooth pulp afferent interaction on uniL~in caudal medulla oblongata, Brain Res., 117 (1976) 508--512. 46 Young, R.F. and Kin~ R.B., Excitability changes in trigeminal primary afferent fibers in response Co noxious and non-noxious stimuli, J. Neurophysiol., 35 (1972) 87--95.