Brain Research, 98 (1975) 73-92 © Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands

73

F U N C T I O N A L S I G N I F I C A N C E OF P R O J E C T I O N FROM T H E C E R E B E L L A R N U C L E I TO T H E M O T O R CORTEX IN T H E CAT

H1ROSHI A S A N U M A AND ROBERT W. H U N S P E R G E R *

The Rockefeller University, New York, N.Y. 10021 (U.S.A.) (Accepted April 28th, 1975)

SUMMARY

The functional organization of projections from the cerebellar subcortical nuclei to the motor cortex through the nucleus ventralis lateralis (VL) of the thalamus was studied using tungsten microelectrodes for stimulation and recordings in acute and chronic cats. The following results were obtained. (1) Microstimulation of a small area of the ventral thalamus produced contraction of a single limb muscle as well as movements of whiskers. (2) The stimulus parameters for producing low threshold contraction of limb muscles were different from those for face muscles. The decrease of the frequency gradually increased the threshold values for face muscles whereas the decrease abruptly increased the threshold for limb muscles. The optimum duration of the train for the lowest thresholds was longer for face muscles. (3) Stimulation of cerebellar nuclei (interpositus and lateralis) produced contraction of limb muscles. The stimulus parameters for the minimum threshold were similar to those for producing contraction of limb muscles from the ventral thalamus. (4) The peripheral receptive fields of neurons located around the low threshold sites in the thalamus were diffuse, i.e. they were driven insecurely by twisting the joints or pressure to the deep structures, but could not be driven by touch or light pressure on a circumscribed area of the body. (5) Chronic ablation of the motor cortex did not abolish the muscle contractions produced by thalamic stimulation, excluding the possibility that the effects were produced by stimulation of the branches of the pyramidal tract fibers reaching the ventral thalamus. (6) Chronic section of the brachium conjunctivum abolished or changed the characteristics of the contractions produced by thalamic stimulation indicating that the previous effects were produced by stimulation of cerebellar efferent fibers reaching the thalamus. * Present address: Physiologisches Institut der UniversitAt ZiJrich, Switzerland.

74 (7) From these results it was concluded that the efferent impulses originating from the cerebellar nuclei can produce contraction of a particular muscle through activation of the red nucleus. These impulses are, at the same time, transmitted to a small group of neurons in the VL and then forwarded to the neurons in the motor cortex. (8) The functional significance of the VL projection system has been discussed in relation to the efferent zones within the motor cortex.

INTRODUCTION

Recently, Asanuma et al. 5 reported the existence of special fibers which project from the ventrolateral nucleus (VL) of the thalamus to wide areas of the motor cortex. These wide projection fibers have a special terminal bush which spreads as widely as 1.0 sq.mm within the motor cortex and the size corresponds to that of the cortical efferent zones 4. These fibers do not carry specific information arising from the periphery, but carry diffuse inputs to the cortex, such as those produced by twisting the joints or by pressure to deep structures 5, in accordance with the observation that VL neurons receive only diffuse peripheral information17, is. A question then arises: what would be the functional significance of these inputs to the motor cortex during natural movements? It has already been shown that VL stimulation can produce mono- and polysynaptic activation of pyramidal tract (PT) cells 2 and also contraction of muscles z4. Furthermore, it has been shown that there is a topographical organization of the projection from VL to the motor cortex 24. These results suggested the existence of a finely grained coupling between VL and the motor cortex. We were interested in this problem and planned the experiments to further elucidate the organization of the projection from VL to the motor cortex by using microstimulation techniques. The results obtained were different from what we expected and will be described following the sequence of the experiments. METHODS

Results were obtained from 26 adult cats weighing between 2.5 and 3.0 kg of which 4 were used for chronic experiments. Since the procedures used for the experiments were not uniform, details of the methods will be described at each section and only the general procedures for acute experiments will be described here. Operations were carried out under inhalation anesthesia composed of a mixture of oxygen ( 4 0 ~ ) and nitrous oxide (60 ~o) supplemented with 1.0 ~ halothane. The cat was mounted on a stereotaxic instrument and the skull was opened over the thalamus or the cerebellum. A special closed chamber, which was used as the substitute for the usual stereotaxic instrument, was then attached on the craniotomy opening in the following manner. The closed chamber, on the lid of which was mounted a micromanipulator (Fig. 1), was attached to the stereotaxic frame by an electrode carrier. The coordinates of the stereotaxic instrument were adjusted in such a way

75

V Fig. 1. Schematic diagram of the experimental set up. The scales grooved on the rim of the chamber serve as stereotaxic coordinates. The head is fixed with the screws drilled into the skull which in turn are anchored to the arm attached to an upright (not shown). Further details are in the text. that the center of the micromanipulator, i.e. the microelectrode, was directed to the target nuclei. The closed chamber was then brought down to a position where the base of the chamber was close to, but still not touching, the skull. While holding the chamber in that position, the base of the chamber was fastened to the skull with melted dental impression wax which after cooling fixed the chamber firmly to the skull. This procedure did not distort the orientation of the manipulator and the scales grooved on the rim of the chamber (Fig. 1) were used to substitute for stereotaxic coordinates during the experiments. After the installation of the chamber a few screws were drilled into the skull and they were covered with dental cement. Additional screws were embedded into this cement mass and these screws were used to anchor the headholder as shown in Fig. 1. The cisterna magna was opened for drainage to prevent swelling of the brain during the experiments. All the wound areas were infiltrated with a long lasting local anesthetic (Zyljectin, Abbott). At the end of the operation, a tranquilizing dose of sodium pentobarbital (Nembutal, 10 mg/kg) was injected intraperitoneally, the ear bars and the snout clamp were removed and inhalation anesthesia was discontinued. This amount of Nembutal effectively eliminated the initial struggling of the cat due to the fixation of the head and the period of tranquilization, which lasted 1-2 h, was long enough for the cat to become accustomed to the head-fixed posture so that additional injections of Nembutal were unneccessary. The experiments were started 1-2 h after the operation. By giving milk or food, be-

76 fore and during the experiments, the animals were usually cooperative to the experiments. Stimulation and recordings

Glass insulated tungsten microelectrodes of the type described by Stoney et al. 23 were used for stimulation as well as for recording unitary spikes. Instead of

using a small tip electrode (exposed tip: 10 #m × 10/~m), a larger electrode with an exposed tip of 20/zm in diameter and 40/zm in length was used. Electrodes of this size were small enough to record unitary activities of neurons in VL, probably because of the sparser density of the neuronal population in this nucleus than in the cerebral cortex where finer electrodes were necessary to isolate unit spikes 4. Electrodes of larger size were necessary in this series of experiments for the following reasons. During the experiments, trains of high frequency (100-500 Hz) negative pulses of 0.2 msec duration which sometimes lasted more than 100 msec were passed through the electrode and the intensity of the pulses frequently was as great as 50 #A. This amount of current inevitably produces bubbles at the tip of the electrode and damages the neighboring tissues 6. To avoid the bubbling from the tip, positive pulses (0.1 msec duration) of roughly the same intensity were passed immediately after the negative pulses. This procedure has been shown to cancel the polarization at the tip without changing the efficiency of the stimulating pulses 5. This cancellation current, however, etches the electrode tip rapidly so that it becomes incapable of passing the current. The size of the electrode used in the present investigation was large enough to enable us to complete a series of necessary examinations without changing the electrode, but still small enough to pick up unitary spikes. The intensity of the current was continuously monitored as the voltage difference across a resistance placed in series with the electrode (Fig. 1). For recording unitary activities, the electrode was switched to a second amplifier. The effect of microstimulation was monitored by visual observation of the contraction and palpation of the muscles, but when quantitative analysis was necessary an electromyogram was recorded through bipolar electrodes from the target muscles. The location of the E M G electrodes was determined by autopsy after the experiments. The threshold intensity was determined by 50 ~ appearance of the effects. The variation in the thresholds of a given effect was usually less than 10~o; but when it was more than 2 0 ~ , the experiment was abandoned. Each train of stimuli was delivered at an interval of 6-10 sec. Histological examinations

At or near to the termination of the experiments, several lesions were made by passing negative currents of 20 #A for 20 sec. The animals were given food or milk as a reward for their cooperation in the experiments and then sacrificed by injecting an overdose of Nembutal into the abdominal cavity. The brain was perfused by saline followed by 10 ~ formalin and removed. On the following day, the appropriate parts of the brain were serially sectioned in the frontal plane (50 #m) by the frozen method and stained by Kliiver and Barrera's method la. All the electrode tracks were reconstructed in reference to the sites of the lesions made during the experiments.

77 RESULTS

(I) Motor effects produced by VL stimulation A train of negative-positive current pulses (0.2 and 0.1 msec duration) of 300 Hz and total duration of 60 msec was used throughout the experiments unless specified. The reasons for choosing these parameters will be stated later. The electrode was first brought down to around the level of horizontal + 3 or + 4 while recording unitary spikes. Usually clearly isolated unitar2y spikes could be recorded at this level and none of these units could be driven from circumscribed areas of the body by natural stimulation. Experiences derived from the early part of the experiments showed that microstimulation of the area dorsal to this level never produced contraction of muscles with currents of 50 #A or less, hence, in the later experiments, stimulation was started at this level. Fig. 2 shows examples of the motor effects produced by microstimulation along the penetrations at the frontal plane of A-10. Stimulation was delivered at 0.5 mm steps and when the effect was observed, the threshold was determined. During the lateral penetration shown in the left of Fig. 2, the effect first appeared as a movement of whiskers and then changed to contraction of forelimb muscles as the electrode was advanced. At the depth of + 1 . 0 mm, contraction of wrist flexor was observed with 50/zA stimulation and at +0.5 mm, simultaneous contraction of elbow extensor and wrist flexor was produced. When the stimulus intensity was reduced to near threshold, contraction of wrist flexor disappeared and the effect appeared to be focused in an elbow extensor. Although it was not difficult to describe the movement in a rough manner such as flexion of elbow

Threshold '~

1.0

•

20

-(15 1"

O"

40 •

q

\ -,.o-l-

~

(jua /

"'"o ~0

-15 -2.0 Depth

(men)

Fig. 2. Left: electrode traverses and sites of stimulation in the thalamus. Throughout the penetrations 50/~A was used. O, no effect; ~ , wrist flexion; ~ , elbow extension; V, shoulder movement; I-, whiskers movement. Larger symbols indicate sites where the effect appeared with thresholds of 25/~A or less. Abbreviations: VPL, ventral-posterolateral nucleus; VL, ventrolateral nucleus; VM, ventromedial nucleus. Right: threshold changes along the penetration shown in the left. Filled circle: m. ext. dig. lat. (wrist flexor); open circles: m. tri. brachii (elbow extensor). Further details are in the text.

78 or wrist, it was very difficult to identify the target muscle without implanting E M G electrodes especially when the stimulus was near threshold. Several electrodes were implanted into the muscles in this trial and two of them picked up E M G activities which followed the threshold thalamic stimulation. At threshold stimulation, the effect appeared only in one muscle, i.e., either in m. extensor digitorum lateralis (wrist flexor) at the depth of +0.5 m m or m. triceps brachii (elbow extensor) at --0.5 mm as shown in the graph in Fig. 2. This is a typical example of the results obtained during the course of the experiments and these results and others revealed the following. The threshold currents for producing contraction of muscles were frequently very low, i.e., about 10/~A or less and the low threshold points of 25 #A or less were confined in a narrow focus of less than 1.0 m m to 2.0 m m along the electrode track. Threshold stimulation at the focus of the effective area usually produced contraction of one muscle, although occasionally it produced contraction of two or more muscles. When multiple muscles were activated simultaneously, the combination of muscles was random; at one time, they were synergists and at another time, antagonists. To obtain an idea about the neuronal mechanisms subserving the muscle contractions produced by VL stimulation, the following studies were made. It is known that transmission of impulses through synapses is subject to temporal summation. It has been known that single pulse stimulation of the surface of the motor cortex produces contraction of muscles with threshold currents of the order of milliamperes 16. Although single pulse microstimulation within the motor cortex does not produce contraction of muscles, repetitive stimulation is able to produce contraction with the currents of the order of microamperes 6. At a frequency of 300~00 Hz, the total duration of the train stimulation necessary to reach the minimum threshold was 25-30 msec, i.e. a total of around 10 pulses was sufficient to saturate the temporal summation. To compare the temporal characteristics of the synaptic connections from VL to the motoneurons with those from the motor cortex, several threshold-

80

oEDL ~TRB • Whisk

:t

3 \

'~

+0

!i \

40 ,

'\

"'-o.+..~'_'.'-_B--__---_-..ea__-:-_-..-:.-_~ ~ - - / ' / " - - - ,

20

,

40

el0

;0

'

100

/.'

•-----A ,

150

,

200

Train Duration (mse¢)

Fig. 3. Relationship between threshold and total duration of train stimulation for producing muscle contractions. Frequency of train stimulation was fixed to 300 Hz. Note that longer train duration was necessary for contraction of face muscles. EDL, m. dig. long.; TRB, m. tri. brach.; Whisk, whiskers movement.

79 duration curves for producing muscle contractions were constructed. Fig. 3 illustrates examples of the results. In all the cases when VL stimulation elicited muscle contractions in the contralateral limbs, the train duration necessary for the saturation of the effects was rather short, i.e., 40-60 msec as shown by open squares and circles in the figure. On the other hand, when the contractions were elicited in face muscles, the saturation necessitated a much longer duration; it was always more than 100 msec. These results indicate that the characteristics of temporal summation involved in the activation of limb muscles from VL somewhat resemble those of the corticospinal tract. On the other hand, the pathway to the face muscle is more complicated and seems different from that to limb muscles. A stimulus frequency of 300 Hz was used throughout the trials because this was the lowest frequency which produced the maximum effect, as will be described below. Another factor that has to be considered for an understanding of synaptic transmission is spatial summation. I f the neurons responsible for producing a given effect are scattered diffusely within the area around the stimulating electrode, then an increase of stimulus strength should increase the effect gradually. On the other hand, if the responsible neurons are located only within a small area near the electrode, an increase of the strength beyond a certain level should not increase the effect. Single shock stimulation of VL within a reasonable range of intensity (less than 100/tA), however, did not produce contraction of muscles, hence examination of spatial summation had to be carried out utilizing repetitive stimulation. Further problems concerning this method will be taken up in the Discussion. The examination of spatial summation has been done in the following way. First, the threshold distribution for a given muscle along the penetrations was mapped out. Then, at the lowest threshold point, the duration of the train stimulation was fixed to 60-100 msec and the frequency within the train was changed to determine the threshold value for each frequency. Fig. 4 illustrates typical results obtained by this method. When the effect of VL stimulation appeared in a face muscle, as shown by

80

,

o TIA

0

!~ o',

60

\

\,~

4o

o EDL ,,

• Whisk.

xo .\ \ \

, \

20 •

,

"\

-'----

I

I

I

,oo

200

300

Frequency

0 .......

I

,oo

0

I

soo

(c/sec)

Fig. 4. Relationship between threshold and frequency of train stimulation for producing muscle contractions. Train duration of 60 msec was used for limb muscles and 100 msec for face muscle. TIA, m. tib. ant.; EDL, m. ext. dig. lat.; Whisk., whiskers movement.

80 filled triangles in Fig. 4, a decrease in frequency gradually increased the threshold. On the other hand, when the effect appeared in limb muscles, the threshold increased rapidly when the frequency was reduced below 300 Hz. Altogether, 8 curves were constructed from different experiments, 5 for limb muscles and 3 for face muscles. In all cases, the results obtained were similar to the results shown in Fig. 4, although on one occasion the curve for limb muscle was less steep than the one illustrated here. It should be noted, however, that the less steep curve was still steeper than the curve for whiskers movements obtained in the same experiment. In summary, it became clear that the spatial characteristics of neuronal pathways producing contraction of face muscles are different from those of limb muscles. To obtain a general idea about the distribution of the effective areas in the thalamus, all the results were pooled in one illustration which is shown in Fig. 5. Since not all the effects were examined by electromyography, the effects were classified by the gross movements produced. All the movements produced by stimulation of 50/zA or less were listed and larger signs indicate those produced by 25/~A or less. The minimum threshold obtained for producing whiskers movement was 8.0 /~A and for limb muscle was 9.0/~A. As shown in the diagram, the effects were produced by stimulation of the ventral portion of VL, the lateral VPM, the medial portion of VPL and also the lamina medullaris externa, as delineated by the atlas of Jasper and Ajmone Marsan 14. Stimulation of dorsal VL did not produce contraction of limb muscles but occasionally produced whiskers movements. There was a rough topographical organization within the effective area. Stimulation of the rostral portion of the area yielded contraction of contralateral hindlimb muscles and the caudal portion produced that of forelimbs. The distribution of the effective sites roughly corresponds to the results obtained by Strick 24, although he did not describe the effect produced by stimulation of VPM and VPL. In addition to the effects produced by stimulation of thalamic nuclei, low threshold effects were elicited from the internal capsule (F-I 1.0 in Fig. 5) and also from Forel's fields (F-9.5). Since we were primarily interested in the thalamocortical projection system, further analysis was not carried out for these effects. In the early stage of the experiments, the receptive fields of neurons were examined whenever they were found at low threshold areas (less than 25/zA) and in the later parts, they were examined occasionally, but in none of the cases did these neurons have clearly delineated peripheral receptive fields. Some neurons were driven insecurely by twisting a limb or pressure applied to deep structures and others could not be driven at all. These characteristics were similar to those of neurons which could be activated antidromically by stimulation of the motor cortex and located in VL ~. (II) Acute ablation o f the motor cortex The results described in the preceding section show that there is a finely grained localization of m o t o r function within the thalamus especially for the movements of the limbs. It is known that VL projects primarily to the m o t o r cortex, although the

81

SITES OF EFFECTIVE MICROSTIMULATION

• \

GONTRALATKRAL

MOVEMENTS

shoulder

elbow

MOVEMIWNTS o f FACE ~- c o n f r . l a f e r a l "4

~t~ ~]~

wrist

[]

hip

[] []

knee

t[~ (])

toot

flexor exfenlor flexor exlensor

flexor exfenlor

Ipellateral flexor exlerllor

.o .,feel ( ~ o . A )

Fig. 5. Summary diagram of the effective points obtained with microstimulation of less than 50/~A. The effects were produced mainly from the ventral thalamus, but also from the internal capsula as well as Ford's fields. Larger symbols denote the sites where the threshold was 25 pA or less. Further details are in the text.

projection f r o m V P M and V P L to the m o t o r cortex has not yet been established 24. A natural question then is whether or n o t these effects are mediated through the m o t o r cortex. T o answer the question, the m o t o r cortex was cooled or ablated in 2 cats. In these experiments, the m o t o r cortex was exposed in addition to the routine procedure described in the Methods. The cortex was then covered with vaseline to prevent the cooling o f the brain. A microelectrode was inserted into the thalamus and after finding a site which produced l o w threshold contraction o f a limb muscle, the vaseline was r e m o v e d and a cooling probe, m a d e o f stainless steel and having a

82 surface area of 5.0 mm × 10.0 m m was placed on the surtace of the motor cortex. The m o t o r cortex was then cooled by pumping saline of 0 °C through the probe while stimulating the thalamus periodically. Several trials were made in this way and in all cases, the muscle contraction produced by thalamic stimulation did not disappear nor did the threshold change. The cooling probe used was large enough to cover the whole pericruciate cortex and the effect of cooling with this probe was powerful enough to suppress all the neuronal activities underneath the probe as has been described in detail in a previous communication 2~. To further confirm these observations, the entire m o t o r cortex was suctioned at the end of the experiments while temporarily anesthetizing the cats with halothane. In both cases, threshold for eliciting muscle contraction was unchanged after the removal of the cortex. The results indicated that the effects of thalamic stimulation were not mediated by excitation of cortical neurons which can also produce contraction of individual limb muscles. (11I) Chronic ablation o f the motor cortex

A question still remained whether the thalamic stimulation activated collaterals of the pyramidal tract fibers terminating in VL and subsequently activated the motoneuron pools through axon reflexes. It is known that the cat's m o t o r cortex sends an abundance of efferent fibers to VL as well as to the ventrobasal complex 2e and a recent physiological study has shown that at least some of them are branches of fibers in the pyramidal tract 12. To examine whether this pathway is responsible for the effect produced by thalamic stimulation, incisions were made to undercut the motor cortex 3 weeks prior to the terminal experiments. Two cats were used for this purpose and in both cats the placing reaction disappeared for the period of observation (3 weeks) and later histological examination showed that the incisions had completely isolated the motor cortex. At the terminal experiments, openings were made in the skull on both sides over the thalamus and a large closed chamber which permitted access to the thalamus on both sides was mounted on the skull. The intact side of the thalamus was used for the control trials. Stimulation of the thalamus on either side produced contraction of limb muscles similar to those observed in the intact cats. A threshold frequency curve was constructed in one cat stimulating the thalamus on the side of the chronic lesion and is shown in Fig. 6. The characteristics of the curve are similar to those shown in Fig. 4. The results hitherto obtained exclude the possibility that the m o t o r cortex plays a determining role in the thalamic activation of limb muscles. So far, descending pathways originating from the thalamic nuclei have not been reported. Still, there was a finely grained localization of m o t o r function within the thalamus when examined by microstimulations. One possibility which might explain the m o t o r effects elicited by thalamic stimulation is that they were mediated through the red nucleus. It has been shown by Tsukahara et al. 26 that stimulation of VL produces monosynaptic EPSP's in red nucleus cells. They attributed these effects to axon reflexes of neurons in the inter-

83

O i

\ ~,

0 BIB o

0

b •x

I 100

I 200

b .....

O ....

O

I 300

I 400

I 500

(Frequency/sec)

Fig. 6. Threshold-frequency relationships for producing contraction of m. biceps brachii (BIB) in a cat whose motor cortex was ablated chronically. The ordinate is expressed by the multiple of the minimum threshold which was 14/~A. Note that the threshold increased rapidly at frequencies less than 300 Hz.

positus nucleus of the cerebellum which project to both the VL and the red nucleus. To examine whether this pathway was responsible for the m o t o r effect, the rubrospinal tract was sectioned in a cat whose m o t o r cortex was isolated chronically. F o r this purpose a larger opening was made around the cisterna magna during the surgical operation (see Methods) so that the rubrospinal tract could be sectioned at the termination of the experiment. After confirming that the chronic ablation of the m o t o r cortex did not abolish the effect produced by thalamic stimulation, a light anesthetic dose of Nembutal (20 mg/kg) was injected into the peritoneal cavity. In the meantime, the electrode was kept in the same position in the control side to allow periodic examination of the effect of stimulation. This amount of Nembutal increased the threshold of thalamic stimulation by about 20 ~o but did not prevent the effect from appearing in the same target muscle. After confirming that the cat was anesthetized, the lateral quadrant of the cervical cord was sectioned with fine scissors, step by step until the effect of thalamic stimulation disappeared. Later histological examination revealed that the area sectioned included the lateral pyramidal tract and the rubrospinal tract as shown in Fig. 7, left side. The electrode was then moved to the side where the m o t o r cortex had been severed chronically. The electrode was advanced and then fixed to a position where stimulation produced contraction of a wrist flexor (m. ext. digitorum communis) with a threshold of 25/~A. Later histological examination revealed that the site stimulated was located at the border line between VL and VPM. The incision into the spinal cord was made in two steps. The first small section was made at position 2 shown in the figurine in Fig. 7. This section did not alter the threshold of the thalamic stimulation. The second incision was made at position 3 which is 1 m m caudal to position 2. The incision was made step by step

84

,

m

C1 Fig. 7. Effects of spinal section of the thalamically induced muscle contractions. Section No. 1 abolished effect produced by stimulation of the contralateral thalamus. Section No. 2 did not, but No. 3 did abolish the thalamic effect. The pyramidal tract in the right side was dysfunctioned by chronic ablation of the left motor cortex. Further details are in the text. until the effect in the wrist flexor suddenly disappeared. The increase of the stimulus intensity at the same site after the incision did not produce motor effect at all until the intensity reached 100 #A. At that time, the effect appeared in the m. biceps brachii. The possibility that the effect disappeared by shock induced by the incision is small because the cat was anesthetized by Nembutal and the first incision on the intact side did not abolish the low threshold effect in the opposite side. Furthermore, the effect appeared again, although in a different muscle, when stronger stimulation was used. The extent of the section confirmed by the later histological examination is shown in Fig. 7, right (positions 2 and 3). From these results, it became clear that the responsible pathway was located in the hatched area shown in Fig. 7, right. Since the participation of the corticospinal tract had been excluded by chronic ablation of the motor cortex in this experiment, the most likely pathway responsible for mediation of the thalamic effect appeared to be the rubrospinal tract.

(IV) Stimulation of cerebellar nuclei The preceding experiments strongly suggested that the motor effect produced by thalamic stimulation was elicited by backfiring the cerebellar efferent fibers which in turn activated neurons in the red nucleus. If this is the case, stimulation of the cerebellar subcortical nuclei should produce low threshold contraction of ipsilateral limb muscles and this has been the case in 3 cats. The experimental procedures were the same as described in the Methods except that the chamber was mounted on the occipital skull so that the electrode could reach the cerebellar nuclei. When the electrode was inserted into the cerebellum, it first passed through an area where an abundance of unit discharges could be recorded (cerebellar cortex) and then went into a silent area. With further advancement, the electrode reached

85

an area where unitary spikes could be recorded and where the stereotaxic coordinates indicated to be the subcortical nuclei. Stimulation was delivered to this area which elicited contraction of the ipsilateral limb and trunk muscles as has been reported elsewhere 21. Once the effect appeared in the course of a penetration, the threshold decreased sharply. The minimum threshold value obtained for activating distal limb muscles in this series of limited trials was 4.0/~A which was considerably less than the minimum value obtained in the thalamus (9.0 #A). The effects produced by stimulation of cerebellar nuclei were complicated and the details of the effects from further study will be reported elsewhere (Shinoda, Zarzecki and Asanuma, in preparation). In this paper, only a relevant part o f the results obtained from limited trials will be presented. Fig. 8 shows an example of the results. Stimulation was delivered at 0.5-ram steps along the penetration and later histological examination showed that this penetration was in the plane of P --8.5 shown in the stereotaxic atlas of Berman 7. In the inset figurine, the points where stimulating currents of 25/~A or less produced effects are mapped and at each point, the effect produced by the lowest stimulation is listed. After completing the penetration, the electrode was pulled back to the position marked with an arrow where contraction of elbow flexor could be produced with the threshold current of 9.0/~A. E M G electrodes were inserted into the target muscle (m. biceps brachii) and the temporal and spatial characteristics of the neuronal pathway mediating the cerebellar effects were examined. As clearly shown in the graph in Fig. 8, the characteristics are similar to those for limb muscles obtained by stimulation of the thalamic nuclei. The threshold reached a minimum at the frequency

5

o

I

3

\

A~

\

o\ \

,.\,, o

I

\ • - -W - = - - -

I

I

i

(Frewency o )

1~

~'~

3oo

(Dwatie . )

20

40

6o

o-- ___-,

I

4~ so

1

5~

~00

Fig. 8. Threshold-frequency and threshold-duration curves for contraction of m. biceps brachii obtained by stimulation of nucleus interpositus. Ordinate is expressed in multiples of the minimum

threshold (9.0 #A). Inset diagram shows electrode track and the site of stimulation (arrow).

86 of 300 Hz with a duration of 50 msec. Altogether 3 sets of curves were constructed, one from m. extensor digitorum communis (wrist flexor) and 2 from m. biceps brachii and all the curves showed similar characteristics. The results described in this section thus support the interpretation outlined in section II1 that stimulation of the ventrobasal complex of the thalamus backfired the fibers originating from the cerebellar nuclei to produce the m o t o r effects.

(V) Chronic section of the brachium conjunctivum The next step taken was to examine whether elimination of fibers originating from the cerebellar nuclei would abolish the thalamic effects. This has been done in 5 cats by sectioning the brachium conjunctivum 3 weeks before the terminal experiments. Under Nembutal anesthesia, the foramen magnum was exposed and enlarged. The cerebellum was then lifted with a spatula inserted through the foramen and the brachium conjunctivum of one side was sectioned under direct visual observation. Of 5 cats operated, 2 survived for 3 weeks and the experiments were carried out on these 2 cats. Later histological examination revealed that the section of the brachium was complete in both cats. In one cat, the m o t o r cortex was exposed at the terminal experiments as described in section III. In the other cat, the m o t o r cortex was not exposed for reasons which will be described later. In the first experiment, penetrations were made into the control side as well as into the chronically treated side in the same frontal plane. The distribution of the effective points in the control side was similar with those in the previous cats indicating that the unoperated side was in normal condition. Stimulation of the chronically lesioned side, however, did not produce any movement at all even with stronger stimulation of up to 100 # A except for the face movements produced by stimulation of the internal capsule. The results demonstrated that the muscle contractions similar to those produced by cerebellar stimulation were abolished by elimination of the brachial fibers. It was not clear, however, why stimulation of VL did not produce at least some movements since it is well established that VL stimulation can produce mono- and polysynaptic activation of PT cells 2. A possibility which might account for the lack of muscle contraction is that although we prevented cooling of the exposed motor cortex by carefully covering it with vaseline, some unknown factors might have changed the excitability of the cortex which resulted in a decrease in the efficiency of synaptic transmission in the motor cortex. To exclude this possibility, the motor cortex was not exposed in the second experiment. Instead a small hole was made in the skull between the motor cortex and the thalamic chamber. This hole was, later on, used to undercut the motor cortex. The acute experiment was carried out in the same way as in the first cat. The thalamic insertions were made first into the control side and the effective areas were mapped. Then the electrode was moved to the operated side and was inserted into the area which corresponded to the effective area in the control side. Two penetrations were made and both passed through an area which produced low threshold contraction of distal forelimb muscles as show in Fig. 9. The minimum threshold

87

I

~.o ,

, i~, ._

VA -, ,,,._....

5

,

VL

,"""-..

4

\

2

•

cu

o '-o_.

1

o ....

o ....

o" - - - ~ / / - - -

•

I

(

Frequency

/Duration

o )

loo

200

300

400

500

• )

20

40

80

eo

~oo

200

Fig. 9. Characteristics of contractions of m. ext. dig. communis produced by thalamic stimulation after chronic section of contralateral brachium conjunctivum. The ordinate is expressed by the multiples of the minimum threshold (16/*A). Note that the shapes of the curves are different from those shown in Fig. 8. Inset figurine indicates the sites where the effects were elicited with stimulation of less than 50/~A. Arrow indicates the site where the curves were obtained. Further details are in the text.

obtained was 16 /zA (train of 300 Hz, 100 msec duration). The character of the contractions, however, seemed different from that of control contractions produced by stimulation of the control side. Although it was difficult to objectively distinguish these two contractions by simple visual observations or palpation of the muscles, the impression obtained was that these contractions were smoother than the control contractions which were twitches. EMG electrodes were then implanted into the m. ext. dig. communis and threshold-duration and threshold-frequency curves were constructed at the point indicated by an arrow in Fig. 9. As clearly shown in the graphs, the characteristics of temporal and spatial summation were different from those obtained in the normal animals. The decrease of the frequency to less than 300 Hz gradually increased the threshold and the increase of the duration to more than 60 msec gradually decreased the threshold. The white substance underneath the motor cortex was then sectioned. This time, the contraction disappeared completely and subsequent stimulations around the effective area with 100/~A did not produce any movement. The results thus clearly indicate that stimulation of VL can produce contraction of limb muscles by activating the motor cortex and that the characteristics of the contractions produced through this pathway are different from those produced by stimulation of the cerebellar subcortical nuclei. The functional significance of these two different systems will be discussed in the following section.

88 DISCUSSION

(1) The problem o f apatial summation The threshold-frequency curves were used to assess the spatial distribution o f neuronal elements responsible for producing muscle contractions (section I). Since it is obvious that impulses originating from VL travel through polysynaptic pathways to activate motoneurons, it is likely that spatial and temporal summation o f facilitation as well as inhibition occur at each synaptic station. One factor that has to be considered, however, is the finding that the contraction o f a given muscle could be produced in a stable manner by weak stimulating currents. This would imply the existence o f a functionally stable pathway for the contraction o f each muscle. If this is the case, then it may be justified to assume each multisynaptic pathway as a solid fixed pathway and the characteristics o f the pathways can be examined by changing the input parameters. As briefly discussed in the results, the parameters we used were the threshold and the frequency, i.e. the total a m o u n t o f inputs to this system per unit period. It has been shown by Stoney et aL 23 that the effective spread (r) o f stimulating current (i) is proportional to the square root o f the stimulating current (r -----kx/i). F r o m this it is calculated that the total number of neurons (n) excited by a given current is proportional to the cube o f the square root of the stimulating current (n = ( k v q ) z) because neurons are distributed in three-dimensional volume. From this, the frequency-intensity relationship which keeps constant the total a m o u n t o f nerve impulses produced per unit period o f time can be calculated as shown by the curve 'a' in Fig. 10. One end o f the curve is drawn linearly because we know that the increase o f the frequency above 300 Hz does not increase the efficiency o f the stimulating current (Fig. 4). a 5 A 4

3 Diffuse inputs

2

\,

I

Specific inputs

1 I

O

I

I

I

i

200

300

400

500

cy/sec Fig. 10. Theoretical relationships between afferent inputs and efferent outflow in diffuse and specific systems. A: diffuse inputs to a theoretical neuronal pathway. B: specific inputs to the same pathway. Note that in the case of B, increase of effective area does not increase the total amount of inputs. C: theoretical frequency-intensity relationships which keep the total amount of inputs constant in the diffuse system. Curves are based on the equation that the effective spread (r) of the current is proportional to the square root (Ci) of the intensity (r-- kxJq; by Stoney et al.23), a: for neurons. b : for fibers. Further details in the text.

89 The results obtained during this series of experiments made it clear that the effects elicited by stimulation of VL in the intact animals were produced by excitation of the brachium fibers which passed through the tissue surrounding the electrode. This implies that the total number of fibers excited by a given current is proportional to the square of the stimulating current because the spread o f the current along the longitudinal direction of the fibers does not increase the number of the fibers excited, i.e., the effective spread of current is two-dimensional. The frequency-intensity curve for the fibers thus calculated is shown by curve 'b' in Fig. 10. The frequency-intensity curve shown in Fig. 9 for a wrist extensor and the curve for the whiskers movement in Fig. 4 are similar to the theoretical curves for neurons and fibers respectively (Fig. 10C a and b) suggesting that the neuronal elements responsible for the contraction of the muscles were scattered diffusely around the electrode (Fig. 10A). On the other hand, the curves for the limb muscles shown in Figs. 4, 6 and 8 may indicate that the fibers responsible for these movements were concentrated around the electrode, the situation shown in Fig. 10B. The above arguments apply only when the stimulating electrode was in the center o f the effective area. Since we know that the lowest threshold value for producing the contraction of the limb muscle was 9.0 ktA, all of the examinations were made when the threshold was around 15 #A or less except for the one whose threshold was 25 #A. On that occasion, the curve was less steep as described in the results (section 1), although it was still steeper than the theoretical curve shown in Fig. 10Cb. When the initial threshold was higher than 25/~A, it was difficult to complete the curve because the microelectrodes we used have a limitation in passing currents beyond a certain level (usually, 100 /~A). When stronger currents were passed, the amount of current fluctuated during the course of a train stimulation making the measurements of the current unreliable. So far, we did not use constant current stimulus isolators because of the difficulty in providing the compensation currents immediately following t h e stimulating currents (see Methods).

(2) Organization of projection from the subcerebellar nuclei to VL It has been shown that microstimulation of VL in normal animals produced contraction of individual muscles. The results in section IV revealed that temporal and spatial characteristics of neuronal pathways from the thalamus and the cerebellar nucleus to limb muscles are similar. The lesion experiments indicated that the thalamic effects were most likely due to antidromic activation of the branches of fibers originating in the cerebellar nuclei. The results altogether suggest that the neurons in the cerebellar nuclei which upon stimulation produce contraction of a particular limb muscle send their branches to a small area in the ventral thalamus. Although we do not know the exact site of termination of these fibers, the results obtained indicate that the fibers of a similar function are grouped together in a small bundle when they approach the final destination in the thalamus, strongly suggesting that their terminal sites are also confined to a small area in VL. The similarity of the characters of contractions produced by stimulation of the thalamus and the cerebellar nuclei also support this interpretation.

90 Concerning the anatomical pathways, it is known that neurons in the interpositus and lateral nuclei send axons to the ventral lateral nucleus of the thalamus as well as to the red nucleus s,l°. These fibers are shown to travel through the ventrolateral strip of the VL and the adjoining medial most region of the VPL 3 where most of the thalamic effects were produced in the present experiments. The anatomical studies did not determine whether the fibers terminating in the red nucleus and the VL are the branches of the same brachial fibers, but a physiological study has shown that at least some of them are branches of the same fibers 28. Whether the fibers which innervate both VL and red nucleus originate from interpositus or lateral or both has not been determined 8, but it is known that most of the fibers originating from the interpositus nucleus reach the red nucleus. On the other hand, the majority of the fibers from the lateral nucleus reach the VL TM. In view of the solid, stable contractions produced by stimulation of the ventral thalamus, it is likely that the axon reflex between the thalamus and the red nucleus is a powerful one, at least functionally. Hence it is possible that the majority of the neurons projecting to VL also send branches to the red nucleus.

(3) Functional significance of afferent inputs from VL to the motor cortex It has recently been shown that the neurons in the interpositus nucleus receive peripheral inputs which are somatotopically organized n. The characteristics of the inputs to the interpositus neurons, however, are different from those to the neurons in the somatosensory z° and in the motor 9 cortices in that they are polymodal and arise from a rather wide area in the periphery. Although a precise study on the receptive fields of dentate neurons is not available, it is unlikely that they receive specific inputs from the periphery 1. Neurons in the VL do not receive specific inputs from the periphery TM. They are driven insecurely by twisting the joint and/or pressing the deep structures 5, suggesting that the details of peripheral information forwarded to the interpositus neurons are lost when the information is transferred to the VL neurons. It has been shown in the present experiments that there is a group of neurons in the cerebellar nuclei which sends axons to a small area in the VL, as well as to the red nucleus. Activation of each group of neurons was shown to produce contraction of individual muscles. The functional significance of afferent inputs to VL neurons from the cerebellum, therefore, is likely to be that each VL neuron is informed which muscle or muscles are being commanded to contract from the cerebellum. This would imply that the information transferred to the motor cortex through VL is concerned more with the outgoing impulses causing movements, rather than with afferent inputs elicited by movements. A question then arises as to the manner in which the thalamic neurons transfer this information to the motor cortex. Do they transfer the information to a particular part of the cortex which controls contraction of a particular muscle ? It has been shown that there is a colony of neurons which project to a particular muscle and these neurons are concentrated in a small area of the motor cortex, the cortical efferent zone 4. As shown in Fig. 9, the characteristics of temporal and spatial

91 summation in the cat after chronic brachium section were different from those elicited by cerebellar stimulation (Fig. 8). The gradual rise of the threshold-frequency curve shown in Fig. 9 suggests that the increase of the stimulus strength increased the amount of inputs arriving in the same efferent zone, i.e., the existence of converging inputs to a given efferent zone. This in turn indicates that the projection from a given area of VL diverges to a wide area of the motor cortex. This interpretation is in good agreement with the anatomical observation by Strick ~4 that a small lesion in VL produces sparse degeneration spread over a relatively wide area of the motor cortex whereas a larger lesion simply increases the density of degeneration within the same cortical area. These anatomical findings are supported by the physiological evidence that some of the VL neurons branch and terminate at a rtumber of different sites in the motor cortex 19. Thus, all the results, including the present ones, indicate the existence of a divergent projection of VL neurons to the motor cortex. How does this divergent projection from VL to the motor cortex function in the control of movement in concert with the finely grained localization of motor function in the motor cortex4? It has been shown recently that there are special wide projection fibers which arise from VL and reach multiple sites in the motor cortex 5. Each terminal of these fibers branches extensively in the gray matter and the diameter of the terminal bush is as wide as 1.0 mm corresponding to the size of the cortical efferent zone. This may suggest that although the projection from VL to the cortex as a whole is diffuse, each VL neuron innervates a specific group of cortical efferent zones. In view of the fact that a larger lesion in the thalamus increases the density of degeneration in the cortex 24, it is likely that each cortical efferent zone receives multiple innervation from VL neurons. The mode of convergence from various VL neurons may determine the hierarchical order in activating the cortical efferent zones, the details of which are not yet known. In addition, the motor cortex receives afferent inputs from the periphery and also from various areas of the cortex through association and commissural fibers. Impulses from the VL, therefore, may excite a group of cortical efferent zones depending on the level of excitability preset by the other inputs in combination with the hierarchical order intrinsic to the VL pojection. The remaining question of how the red nucleus controls contraction of limb muscles has been studied by Ghez 13. ACKNOWLEDGEMENTS

This research was supported by Grant NS-10705 from the National Institute of Health. Dr. R. W. Hunsperger was a visiting professor on his sabbatical leave from the Physiologisches Institut der Universit/it Ziirich, Switzerland. The authors would like to express their gratitude to Miss K. Alexieva for her technical assistance. REFERENCES l ALLEN, G. I., AND TSUKAHARA,N., Cerebrocerebellar communication system, Physiol. Rev., 54 (1974) 957-1006.

92 2 AMASSIAN,V. E., AND WEINER, H., Monosynaptic and polysynaptic activation of pyramidal tract neurons by thalamic stimulation. In D. P. PURPURA AND M. D. YAHR (Eds.), The Thalamus, Columbia University Press, New York, 1966, pp. 255-282. 3 ANGAUT, P., The ascending projections of the nucleus interpositus posterior of the cat cerebellum : an experimental anatomical study using silver impregnation methods, Brain Research, 24 (1970) 377-394. 4 ASANUMA, H., Cerebral cortical control of movement, Physiologist, 16 (1973) 143-166. 5 ASANOMA, H., FERNANDEZ, J., SCHEmEL, M. E., AND SCHEmEL, A. B., Characteristics of projections from the nucleus ventralis lateralis to the motor cortex in the cats: an anatomical and physiological study, Exp. Brain Res., 20 (1974) 315-330. 6 ASANUMA, H., AND WARD, J. E., Patterns of contraction of distal forelimb muscles produced by intracortical stimulation in cats, Brain Research, 27 (1971) 97-109. 7 BERMAN,A., The Brain Stem of the Cat. A Cytoarchitectonic Atlas with Stereotaxic Coordinates, The University of Wisconsin Press, Madison, Wisc., 1968. 8 BRODAL, A., Neurological Anatomy in Relation to Clinical Medicine, Oxford University Press, London, 1969, 283 pp. 9 BROOKS, V. B., RUDOMIN, P., AND SLAYMAN,C. L., Peripheral receptive fields of neurons in the cat's cerebral cortex, J. Neurophysiol., 24 (1961 ) 302-325. l0 COHEN, D., CHAMBERS,W. W., AND SPRAGUE, J. M., Experimental study of the efferent projection from the cerebellar nuclei to the brain stern of the cat, J. comp. Neurol., 109 (1958) 233-259. 11 ECCLES, J. C., RANTUCCI, T., ROSEN, 1., SCHEXD, P., AND TABOmKOVA, H., Somatotopic studies on cerebellar interpositus neurons, J. Neurophysiol., 37 (1974) 1449-1559. 12 ENDO, K., ARAKI, T., AND YAGI, N., The distribution and pattern of axon branching of pyramidal tract cells, Brain Research, 57 (1973) 484-491. 13 GHEZ, C., Input-output relations of the red nucleus in the cat, Brain Research, 98 (1975) 93 108. 14 JASPER, H. H., AND AJMONE MARSAN, C., A Stereotaxic Atlas of the Diencephalon of the Cat, National Research Council of Canada, Ottawa, 1954, 15 pp. 15 KLCVER, H., AND BARRERA,E., m method for the combined staining of cells and fibers in the nervous system, J. Neuropath. exp. Neurol., 12 (1953) 40(0403. 16 LIDDELL, E. G. T., AND PHILUPS, C. G., Thresholds of cortical representation, Brain, 73 (1950) 125-140. 17 MASSION,J., ET ALBE-FESSARD, P. A., Activit6s 6voqu6es chez le chat darts la r6gion du nucleus ventralis lateralis par diverses stimulations sensorielles. 1. Etude macrophysiologique, Electroenceph, clin. Neurophysiol., 19 (1965) 433-451. 18 MASSION, J., ET ALBE-FESSARD, P. A., Activit6s 6voqu6es chez le chat darts la r6gion du nucleus ventralis lateralis par diverses stimulations sensorielles. 11. Etude microphysiologique, Electroenceph, clin. Neurophysiol., 19 (1965) 452-469. 19 MASSION, J., AND RISPAL-PADEL, L., Differential control of motor cortex and sensory areas on ventrolateral nucleus of the thalamus. In T. L. FR~GYESl, E. RIr~WK, AND M. D. YAHR (Eds.), Corticothalamic Projections and Sensorimotor Activities, Raven Press, New York, 1973, pp. 357 374. 20 MOUN'rCASTLE, V. B., Modality and topographic properties of single neurons of cat's somatic sensory cortex, J. Neurophysiol., 20 (1957) 408-434. 21 POMPEIANO,O., Functional organization of the cerebellar projections to the spinal cord. In C. A. Fox AND R. S. SNIDER (Eds.), The Cerebellum, Progr. Brain Res., Vol. 25, Elsevier, Amsterdam, 1967, pp. 282-321. 22 R~NWK, E., The corticothalamic projection from the pericruciate and coronal gyri in the cat. An experimental study with silver-impregnation methods, Brain Research, l0 0968) 79-119. 23 STONEY, S. D., JR., THOMPSON, W. O., AND ASANUMA,H., Excitation of pyramidal tract cells by intracortical microstimulation: effective extent of stimulating current, J. Neurophysiol., 31 0968) 659-669. 24 STRICK, P. L., Light microscopic analysis of the cortical projection of the thalamic ventrolateral nucleus in the cat, Brain Research, 55 (1973) 1-24. 25 THOMPSON, E. D., STONEY, S. D., JR., AND ASANUMA, H., Characteristics of projections from primary sensory cortex to motorsensory cortex in cats, Brain Research, 22 (1970) 15-27. 26 TSUKAHARA, N., TOYAMA, K., AND KOSAKA, K., Electrical activity of red nucleus neurons investigated with intracellular microelectrodes, Exp. Brain Res., 4 (1967) 18-33.