The ultrastructural morphology and distribution of trigeminohypoglossal connections labeled with horseradish peroxidase

Brain Research, 422 (1987) 235-241 Elsevier

235

BRE 12910

The ultrastructural morphology and distribution of trigemino. hypoglossal connections labeled with horseradish peroxidase Rosemary C. Borke and Martin E. Nau Department of Anatomy, USUHS, F. Herbert School of Medicine, Bethesda, MD 20814-4799 (U..S.A.)

(Accepted 3 March 1987) Key words: Spinal trigeminal nucleus, pars interpolaris; Horseradish peroxidase; Anterograde labeling; Axon terminal in XIt nucleus; Retrograde labeling; Hypog!ossal motoneuron

Axon terminals projecting to the hypoglossal nucleus have been identified and characterized by electron microscopy following injections of horseradish peroxidase (HRP) into pars interpolaris of the spinal trigeminal nucleus (SPVN) in adult rats. Over 70% of the anterogradely labeled terminals contained spherical vesicles (S-terminals) and their synaptic densities were chiefly asymmetrical (Gray Type I). The rest (28%) of the labeled terminals had flattened vesicles (F-terminals) and predominantly established symmetrical (Gray Type II) synaptic contacts. The diameters of labeled terminals were 0.5-2.5 gm. Two-thirds of the S-terminals had diameters <1.25/~m, whereas, F-terminals were distributed equally in the higher (>1.25) and lower (<1.25) diameter ranges. Most axon terminals ended on dendrites of hypoglossal neurons; some, chiefly F-terminals, formed axosomatic endings. Dendrites had diameters of 0.5-5/~m. The majority of S- and F-terminals ended on dendrites with diameters of <2.5/xm. However, more F-terminals (17%) than S-terminals (11% ) were presynaptic to dendrites >2.5 um in diameter. Experiments in which anterograde HRP labeling of trigemino-hypoglossal projections was combined with retrograde WGA-HRP labeling of motoneurons projecting to the tongue, demonstrated that SPVN axons end on dendrites of these motoneurons. Whether some of the tfigeminal fibers also terminate on intrinsic hypoglossal interneurons remains to be determined. INTRODUCTION The tongue is a muscular organ, capable of highly organized and complex movements, associated with oro-facial functions such as mastication, deglutition and phonation. In order to understand the control of tongue movements, the structural, physiological and pharmacological aspects of neural pathways mediating tongue motility must be examined. To date, the somatotopic organization 1s'24 and synaptic morphology5 of the hypoglossal nucleus have been delineated and neurotransmitters utilized by terminals synapsing on neurons in the X I I t h nucleus have been identified 1°. Recent work of the current investigators has focused on specifying the brainstem origins of hypoglossal input 6 and analyzing the synaptic organization and distribution of axon terminals originating from the various sources of hypoglossal afferents 7.

Portions of the trigeminal nuclei, particularly pars interpolaris of the spinal trigeminal nucleus (SPVN), serve as somatic centers, that receive cortical 39'4° and peripheral 16,30 fibers, and project to the XIIth nucleus 1'6'36. the only source of motor innervation for the tongue musculature. Since the sensory trigeminal nuclei act as interneuronal relays in the coordination of tongue movements, the ways in which SPVN, pars interpolaris and the hypoglossal nucleus relate structurally is a basis of interest. T h e object of the current investigation was to analyze the synaptic organization and distribution of trigemino-hypoglossal connections and to interpret their significance. Horseradish peroxidase ( H R P ) and wheat germ agglutin i n - H R P ( W G A - H R P ) were utilized as anterograde and retrograde tracers to designate respectively, the ultrastructural patterns of presynaptic and postsynaptic connectivity.

Correspondence: R.C. Borke, USUHS, F. Hebert School of Medicine, Department of Anatomy, 4301 Jones Bridge Road, Bethesda. MD 20814-4799, U.S.A.

236 MATERIALS AND METHODS Anterograde labeling experiments were carried out in 15 adult male rats of the O s b o r n - M e n d e l strain. In anesthetized rats (7% chloral hydate, 5 ml/kg), single injections (40-100 hi) of 15-30% HRP in 0.1 M Tris buffer were made through glass micropipettes, stereotaxically guided into SPVN, pars interpolaris as detailed previously s. Following stereotaxic surgery, the left hypoglossal nerve was transected in 4 rats and 50/A of 1% W G A - H R P in phosphate-buffered saline (PBS) was injected into the dorsal and lateral aspects of the right side of the tongue to produce ipsilateral retrograde labeling of hypoglossal motoneurons. Control animals were also included in the experimental design. Four rats were used for the control series: two rats for each of 2 survival times (1 and 2 - 3 days) 8. All procedures were identical to those followed for the experimental groups except that a small volume (40 nl) of the tracer vehicle, 0.1 M Tris buffer, was injected into the SPVN, pars interpolaris. After a survival period of 1-3 days, experimental and control animals were reanesthetized and killed by transcardiac perfusion of Ringer's solution, dilute and concentrated aldehydes 2s and a final solution of (I.1 M phosphate buffer. Transverse sections of the medulla were cut at 50 ~m on the Vibratome and reacted histochemically for H R P according to the cobait-glucose oxidase method ( C O - G O D ) 15. Alternate sections were mounted and counterstained with Neutral red to evaluate the position and extent of the injection site. The remaining reacted sections were processed for electron microscopy according to previously detailed procedures 7-9. Ultrathin sections of the hypoglossal nucleus were cut and left unstained for ultrastructural evaluation. Using a Micro Plan II image analysis system interfaced to an IBM-PC, diameters of the labeled presynaptic structures and their postsynaptic part-

ners were measured from electron micrographs (x22,500-47,500) and labeled S- and F-terminals were counted. Labeled S- and F-terminals were subdivided into (1) those axon terminals apposed to neuronal cell parts and (2) those axon terminals unapposed to neuronal cell parts. In addition, the presence or absence of a synaptic specialization was indicated and the type of synaptic specialization was specified for each labeled terminal. RESULTS

Anterograde HRP experiments Light microscopic preparations of HRP-injected tissue revealed numerous labeled cell bodies and processes of neurons in SPVN, pars interpolaris. Most injection sites were confined to the pars interpolaris, and Spinal V tract (Fig. 1A), although in a few cases the injection spread into the rostral portion of SPVN, pars caudalis. Labeled nerve fibers coursed medially from the injection site. Some fibers entered the lateral aspect of the XIIth nucleus; others capped the lateral margin to enter the nucleus along its ventral border. Punctate labeling was distributed throughout the ipsilateral hypoglossal nucleus. Ultrastructural examination disclosed electrondense granules of reaction product in single axon terminals scattered sparsely throughout the hypoglossal nucleus. Electron-dense granules of reaction product could also be identified on occasion in very small myelinated axons and pericytes, but were not seen in somata or dendrites of hypoglossal neurons. H R P granules were observed in axon terminals filled with spherical synaptic vesicles (S-terminals) as well as those containing flattened synaptic vesicles (F-terminals). Unlabeled dense core vesicles were sometimes found in the labeled S-terminals (Fig. 1B). Quantitative evaluation of 576 labeled axon terminals revealed that: (1) 72% were S-terminals and (2) 28% were F-terminals. All of the labeled S-terminals were

Fig. 1. A: bright-field micrograph of HRP injection site in pars interpolaris of SPVN (dotted outline) at level of Xllth nucleus (solid outline) rostral to the obex. B: granules of HRP reaction product (arrows) in a spherical-vesicle terminal that contains a few unlabeled dense core vesicles (arrowheads) and contacts a dendrite (d) in the XIIth nucleus. C: a small spherical-vesicle terminal (s) with membrane-bound HRP label forms an asymmetrical synapse with a small dendrite (d). HRP (arrow) is also seen in a distal axon that ends as flattened-vesicle terminal (f) on the same dendrite. D: HRP label (arrow) in a spherical-vesicle terminal (s) that forms a Taxi body synapse (arrowheads) on a small dendrite. E: axon terminal with flattened vesicles (f) contains HRP label (arrows) and synapses with medium dendrite (d). F: flattened-vesicle terminal (f) with tubular profile of HRP label (arrow) contacts a medium dendrite (d).

---,,1

238 apposed to neuronal cell parts, but synaptic specializations were lacking in 32% of the labeled S-terminals (Fig. 1B). Synaptic specializations of 3 distinct types were associated with the majority (68%) of the labeled S-terminals. Asymmetrical specializations (Gray Type I) 12, characterized by pronounced accumulation of electron-dense material on the cytoplasmic face of the postsynaptic membrane, were common (58%) (Fig. 1C). Taxi bodies 34,35, consisting of subsynaptic electron-dense bodies that form a plaque beneath an asymmetrical synaptic contact, were seen occasionally (9%) (Fig. 1D). Cistern-type specializations z9, distinguished by a flattened subsynaptic cistern that is closely applied to the length of the postsynaptic membrane, were rarely (1%) identified. For the labeled F-terminals, 72% had synaptic specializations (Fig. 1E). The remainder of labeled Fterminals were apposed to neuronal cell parts, but no synaptic specialization was evident (Fig. 1F). Specializations were predominantly symmetrical (Gray Type II) 12, typified by a proportionate amount of electron-dense material associated with the cytoplasmic faces of pre- and postsynaptic membranes. Asymmetrical membrane specializations (3%) and cistern type specializations (1%) were occasionally associated with labeled F-terminals. The diameter of axons terminals of S- and F-types, originating from neurons in SPVN, ranged from 0.5 to 2/zm. Two-

thirds of the S-terminals had diameters <1.25/~m, whereas there was an equal distribution of F-terminals with diameters measuring in the higher (>1.25) and lower (<1.25) diameter ranges. Most of the labeled axon terminals (97%) were presynaptic to dendrites. The diameter of postsynaptic dendrites ranged from 0.5 to 5/~m. The majority of S- and F-terminals ended on dendrites with diameters of <2.5 ~m. However, a higher proportion of Fterminals (17%) were presynaptic to dendrites >2.5 ~m in diameter than S-terminals (11%). Some labeled terminals established axosomatic contacts (3%); those that did were chiefly F-terminals. One labeled terminal apposed the nodal region of a myelinated axon, but an unequivocal synapse could not be distinguished.

Anterograde and retrograde HRP experiments Electron-dense granules of HRP were detected in some apposing pre- and postsynaptic structures when retrograde labeling of somata and dendrites of motoneurons that project to tongue musculature was combined with the anterograde labeling of trigemino-hypoglossal terminals. Quantitative evaluation of 204 labeled synaptic pairs revealed that: (1) 78% had spherical vesicles in the presynaptic structure (Fig. 2A) and (2) 22% contained flattened vesicles in the axon terminals (Fig. 2B). Membrane specializations

Fig. 2. A: spherical-vesicle terminal (s), anterogradely labeled with HRP (arrowheads), in synaptic contact with medium dendrite (d) of a hypoglossal motoneuron retrogradely labeled with HRP (arrows). B: flattened-vesicle terminal (f) from SPVN neuron contains anterograde HRP label (arrowheads) and forms symmetrical synapse with large dendrite (d) of hypoglossal motoneuron retrogradely labeled with HRP (arrows) by injecting W G A - H R P into the tongue.

239 associated with the S-terminals were chiefly asymmetrical, while symmetrical densities were common for the F-terminals. Cistern-type synapses were occasionally labeled.

Control experiments In the current work, block staining with uranyl acetate was not performed during the processing of tissue for electron microscopy and ultrathin sections were not post-stained with uranyl and lead salts. While these precautions decreased the possibility of misidentification of normally occurring membranebound granules with electron-dense contents as containing HRP reaction product, these measures did not eliminate the possibility. The control experiments were included to safeguard against such misidentification. The injection of 0.1 M Tris buffer into SPVN, pars interpolaris, and histochemical treatment of Vibratome sections through the medulla with the COGOD reaction for HRP produced peroxidatic activity only within erythrocytes at the injection site and in those associated with sinusoids of area postrema. Peroxidatic activity was not seen in neuronal perikarya and axons surrounding the injection site. At the ultrastructural level, membrane-bound granules with electron-dense contents (dense core vesicles) were seen in some axon terminals of the hypoglossal nucleus of the control rats. The size and shape of these unlabeled dense core vesicles resembled some of the HRP-labeled vesicles in axon terminals from HRP tracer experiments. However, the electron density of the membrane-bound granules in the HRP-injected material (Fig. 1D) was markedly intensified and vesicles containing HRP reaction product could easily be distinguished from the unlabeled dense core vesicles (Fig. 1B). DISCUSSION Retrograde labeling experiments using HRP have demonstrated bilateral, although chiefly ipsilateral projections from the trigeminal sensory nuclei to the hypoglossal nucleus 6'36. Anterograde HRP and autoradiographic experiments verified these connections 1. Specifically, the dorsal portion of SPVN, pars interpolaris and oralis 6 and the dorsal part of the principal sensory nucleus (PSN) of V 1'6 have been

identified as sources of hypoglossal afferents. The present study, utilizing HRP as an anterograde tracer, provides a detailed accounting of the ultrastructural morphology and distribution of trigemino-hypoglossal projections. Preliminary electron microscopic findings, reported for a small number of axon terminals (n = 8) of the PSN-XII projections. indicated that labeled terminals contained only spherical vesicles and established asymmetrical swlaptic densities exclusively with dendritic processes in the hypoglossal neuropil 1. The current findings of SPVN-XII differed from those of the PSN-XII projections in several ways: (1) S-terminals were not the only type of trigemino-hypoglossal ending, although labeled S-terminals outnumbered labeled F-terminals by as much as 3:1. (2) S- and F-terminals ended chiefly, but not exclusively, on dendrites, as some axosomatic terminals were labeled. Whether these inconsistencies reflect differences in the synaptic organization of hypoglossal afferents from individual trigeminal sensory nuclei and/or modifications obtained by the increased number of labeled terminals sampled in the current work could not be determined. The morphological findings of the current work, however, do bear on certain physiological evidence. The electrical properties of disynaptic pathways from peripheral 14,2°,25and cortical 17"23'26'~2afferents to hypoglossal motoneurons indicate that neurons in the SPVN modulate the control of tongue movements by their excitatory and inhibitory input to hypoglossal motoneurons. If the assumption that synaptic vesicle shape 3A9'37 and type of synaptic density 27 are correlated to function of the synapse is valid, the current finding of two morphological types of labeled terminals, S-asymmetrical and F-symmetrical would provide the presumed morphological substrates to corroborate the excitatory and inhibitory actions ~322 31,33of trigemino-hypoglossal projections. The utilization of combined anterograde and retrograde HRP labeling resulted in the ultrastructural identification of reaction product in synaptic pairs involved in the trigemino-hypoglossal connections. The current findings provide the initial demonstration that SPVN fibers end as S- and F-terminals in synaptic contact with hypoglossal motoneurons. If Sand F-terminals are related to the functional activity of synapses, one could predict that excitatory and in-

240 hibitory inputs from SPVN neurons are directed to hypoglossal motoneurons. The spatial distribution of synapses from our work is compatible with the anatomical a r r a n g e m e n t of spinal m o t o n e u r o n afferents, in which S-terminals end on smaller dendrites of anterior horn cells than F-terminals, but both synaptic types are widely distributed over the neuronal membrane 2'el. It was not possible, however, to specify if a different a r r a n g e m e n t of S- and F-input exists for p r o t r u d e r m o t o n e u r o n s located ventrally and retractor m o t o n e u r o n s situated dorsally in the XIIth nucleus. W h e t h e r some SPVN fibers synapse on a population of intrinsic interneurons, recognized in the X l l t h nucleus of rats 4'1~, also could not be determined. Solving these questions would have significant import on elucidating a paradigm of the modulatory mechanisms associated with trigemino-hypoglossal tongue movements.

not be ascertained, but it is of interest that pars interpolaris of SPVN contains e n k e p h a l i n - i m m u n o r e a c tive neurons 38 and the current findings related to some S-terminals originating from this subdivision of SPVN share features in c o m m o n with those e n k e p h a lin terminals identified by immunocytochemical methods 1°. This raises the possibility that enkephalin could be a putative n e u r o t r a n s m i t t e r for some of the S-terminals of the trigemino-hypoglossal projections. The current observations should prove useful for future attempts to interrelate synaptic m o r p h o l o gy and distribution of particular sources of hypoglossal inputs with functional activity and neurotransmitter content of synapses.

W h e t h e r or not the labeled trigemino-hypoglossal terminals are associated with specific neurotransmitters remains to be d e t e r m i n e d . However, a recent ultrastructural immunocytochemical study 1° describing the m o r p h o l o g y and distribution of axon terminals in the XIIth nucleus of rats may bear on this point. E n k e p h a l i n - i m m u n o r e a c t i v e terminals were 0 . 7 - 0 . 8 u m in diameter, contained spherical vesicles with occasional dense core vesicles and made asymmetrical synaptic contact with large and small dendrites. The source of the enkephalin terminals could

The author thanks Dr. Malcolm B. C a r p e n t e r for his valuable suggestions and comments. The skilled typing of Mary Thomson is gratefully acknowledged. This work was s u p p o r t e d by the D e p a r t m e n t of Defense Grant C07019. The opinions or assertions contained herein are the private ones of the author and are not to be construed as official or reflecting the views of D o D or the U S U H S . The experiments reported herein were conducted according to principles set forth in the ' G u i d e for Care and Use of L a b o r a t o -

REFERENCES

(1985) 127-132. 8 Borke, R.C. and Nau, M.E., Factors affecting the ultrastructural pattern of anterograde labeling in axon terminals with HRP, Brain Res. Bull., 16 (1986) 259-265. 9 Carson, K.A. and Mesulam, M.-M., Electron microscopic demonstration of neural connections using horseradish peroxidase: a comparison of the tetramethylbenzidine procedure with other histochemical methods, J. Histochem. Cytochem., 30 (1982) 425-435. 10 Connaughton, M., Priestley, J.V,, Sofroniew, M.V., Eckstein, F. and Cuello, A.C., Inputs to motoneurones in the hypoglossal nucleus of the rat: light and electron microscopic immunocytochemistry for choline acetyltransferase, substance P and enkephalins using monoclonal antibodies, Neuroscience, 17 (1986) 205-224. 11 Cooper, M.H., Neurons of the hypoglossal nucleus of the rat, Otolarvngol. Head Neck Surg., 89 (1981) 10-15. 12 Gray, E.G., Axo-somatic and axo-dendritic synapses of the cerebral cortex: an electron microscopic study, J. Anat., 93 (1959) 420-433. 13 Green, J.D. and Negishi, K., Membrane potentials in by_ poglossal motoneurons, J. Neurophysiol., 26 (1963) 835-856.

I Aides, L.D. and Boone, T.B., Organization of projections from the principal sensory trigeminal nucleus to the hypoglossal nucleus in the rat: an experimental light and electron microscopic study with axonal tracer techniques, Exp. Brain Res.. 59 (1985) 16-29. 2 Bernstein, J.J. and Bernstein, M.E., Ventral horn synaptology in the rat, J. Neurocytol., 5 (1976) 109-123, 3 Bodian, D., Synaptic types on spinal motoneurons: an electron microscopic study, Bull. Johns Hopkins Hosp., 119 (1966) 16-45. 4 Boone, T.B. and Aides, LD., The ultrastructure of two distinct neuron populations in the hypoglossal nucleus of the rat, Exp. Brain Res., 54 (1984) 321-326. 5 Boone, T.B. and Aides, L.D., Synaptology of the hypoglossal nucleus in the rat, Exp. Brain Res., 57 (1984) 22-32. 6 Borke, R.C., Nau, M.E. and Ringler Jr., R.L., Brainstem afferents of hypoglossal neurons in the rat, Brain Research, 269 (1983) 47-55. 7 Borke, R.C. and Nau, M.E., The ultrastructural identification of reticulo-hypoglossal axon terminals anterogradely labeled with horseradish peroxidase, Brain Research, 337

ACKNOWLEDGEMENTS

ry A n i m a l s ' , D H E W Pub. No. ( N I H ) 78-23.

241 14 Hunter, I.W. and Porter, R., Glossopharyngeal influences on hypoglossal motoneurones in the cat, Brain Research, 74 (1974) 161-166. 15 Itoh, K., Konishi, A., Nomura, S., Mizuno, N., Nakamura, Y. and Sugimoto, T., Application of coupled oxidation reaction to electron microscopic demonstration of horseradish peroxidase: cobalt-glucose oxidase method, Brain Research, 175 (1979) 341-346. 16 Jacquin, M.F., Sembra, K,, Egger, M.D. and Rhoades, R.W., Organization of HRP-labeled trigeminal mandibular primary afferent neurons in the rat, J. Comp. Neurol., 215 (1983) 397-420. 17 Kaku, T., Functional differentiation of hypoglossal motoneurons during the amygdaloid or cortically induced rhythmical jaw and tongue movements in the rat, Brain Res. Bull., 13 (1984) 147-154. 18 Krammer, E.B., Rath, T. and Lischka, M.F., Somatopic organization of the hypoglossal nucleus: an HRP study in the rat, Brain Research, 170 (1979) 533-537. 19 Larramendi, L.M.H., Fickenscher, L. and LemkeyJohnston, N., Synaptic vesicles of inhibitory and excitatory terminals in the cerebellum, Science, 156 (1967) 967-969. 20 Lowe, A.A., The neural regulation of tongue movements, Progr. Neurobiol., 15 (1981) 295-344. 21 McLaughlin, B.J., The fine structure of neurons and synapses in the motor nuclei of the cat spinal cord, J. Comp. Neurol., 144 (1972) 429-460. 22 Morimoto, T. and Kawamura, Y., Inhibitory postsynaptic potentials of hypoglossal motoneurons, Exp. Neurol., 37 (1972) 188-198. 23 Morimoto, T. and Kawamura, Y., Properties of tongue and jaw movements elicited by stimulation of the orbital gyrus in cat, Arch. Oral Biol., 18 (1973) 361-372. 24 Odutola, A.B., Cell grouping and Golgi architecture of the hypoglossal nucleus of the rat, Exp. Neurol., 52 (1976) 356-371. 25 Porter, R., The synaptic basis of a bilateral lingual-hypoglossal reflex in cats, J. Physiol. (London), 190 (1967) 611-627. 26 Porter, R., Cortical actions on hypoglossal motoneurones in cats: a proposed role for a common internuncial cell, J. Physiol., 193 (1967) 295-308. 27 Raviola, G. and Raviola, E., Light and electron microscopic observations on the inner plexiform layer of the retina, Am. J. Anat., 120 (1967) 403-426.

28 Reese, T.S. and Karnovsky, M.J., Fine structural localization of a blood-brain barrier to exogenous peroxidase. J. CelIBiol., 34 (1967) 207-217. 29 Rosenbluth, J., Subsurface cisterns and their relationship to the neuronal plasma membrane, J. Cell BioL, 13 (1962) 405-421. 30 Shiegenaga, Y., Chen, I.C., Suemune, S., Nishimori, T., Nasution, I.D., Yoshida, A., Sato, H., Okamoto, T., Sera, M. and Hosoi, M., Oral and facial representation within the medullary and upper cervical dorsal horns in the cat, .! Comp. Neurol., 243 (1986) 388-408. 31 Takata, M., Lingually induced inhibitory postsynaptic potentials in hypogiossal motoneurons after axotomy, Brain Research, 224 (1981) 165-169. 32 Takata, M. and Nagahama, T., Cortically induced postsynaptic potentials in hypoglossal motoneurons after axotomy, Neuroscience, 13 (1984) 855-862. 33 Takata, M. and Ogata, K., Two components of inhibitory postsynaptic potentials evoked in hypoglossal motoneurons by lingual nerve stimulation, Exp. Neurol., 69 (1980) 299-310. 34 Taxi, J., l~tude de l'ultrastructure des zones synaptiques dans les ganglions sympathiques de la Grenouitte, C;R. Acad. Sci. (Paris), 252 (1961) 174-176 35 Taxi, J., l~tude de certaines synapses interneuronales du syst~me nerveux autonome, Acta Neurove&, (Wien), 26 (1964) 360-370. 36 Travers, J.B. and Norgren, R., Afferent projections to the oral motor nuclei in the rat, J. Comp. Neurol., 220 (1983) 280-298. 37 Uchizono, K., Characteristics of excitatory and inhibitory synapses in the central nervous system of the cat, Nature (London), 207 (1966) 642-643. 38 Uhl, G.R., Goodman, R.R., Kuhar, M.J., Childers, S.R. and Snyder, S.H., Immunohistochemical mapping of enkephalin-containing cell bodies, fibers and nerve: terminals in the brainstem of the rat, Brain Research, 166 (1979) 75:-94. 39 Wold, J.E, and Brodal, A., The projection of cortical sensorimotor regions onto the trigeminat nucleus in the cat. An experimental anatomical study, Neurobiology, 3 (1973) 353-375. 40 Wold, J.E. and Brodal, A., The cortical projection of the orbital and procreate gyri to the sensory trigeminal nuclei in the cat. An experimental anatomical study, Brain Research, 65 (1974) 381-395.