Topographical organization of the projections from the reticular thalamic nucleus to the intralaminar and medial thalamic nuclei in the cat

The topography of the projections from the reticular nucleus of the thalamus (RT) to the intralaminar and medial thalamic nuclei were studied in the cat by the method of retrograde transport of horseradish peroxidase (HRP). Single small injections of the enzyme were made in the different intralaminar nuclei--mediodorsal, ventromedial, midline, and habenular--and in anterior group nuclei. The location and density of the neuronal labeling in the different parts of the RT were studied in each case. Our results show that 1) after injections located in all the nuclei here studied, a consistent number of labeled neurons were found in the RT, except for the injections in the lateral habenula and the anterior thalamic nuclear complex, both of which did not label neurons in the RT. 2) Among the other thalamic nuclei here studied, the most medially situated receive less numerous RT projections than those most laterally located. 3) Injections in all the nuclei studied gave rise to a cellular labeling in the anterior sectors of the RT, except for the anterior nuclear group and the lateral habenula. The projections from the rostral pole of the RT were topographically mediolaterally organized. 4) The anterodorsal part of the pregeniculate sector of the RT projects upon the large-celled part of the lateral central nucleus and to a lesser extent upon the paracentral, centromedian, and ventromedial nuclei, the anterior part of the lateral central nucleus, and the lateral band of the mediodorsal nucleus. The posterodorsal part of the RT pregeniculate sector only projects to the large-celled part of the lateral central nucleus. The dorsal portion of the posteroventral part of the RT pregeniculate sector also projects upon the large-celled part of the lateral central nucleus; its ventral portion projects to the ventromedial nucleus, the posterior part of the paracentral nucleus, the lateral band of the mediodorsal nucleus, and the centromedian nucleus. 5) The infrageniculate sector of the RT projects to the posterior part of the ventromedial nucleus. A weaker projection to the large-celled part of the lateral central nucleus, the centromedian nucleus, and the lateral band of the mediodorsal nucleus was also observed. 6) The ventral lateral geniculate nucleus projects upon the large-celled part of the lateral central nucleus, the lateral band of the mediodorsal nucleus, and the ventromedial nucleus. All these findings suggest an important modulatory action of the RT on the activity of the thalamic nuclei considered here.     

Axonal branching in the periaqueductal gray projections to the thalamus: a fluorescent retrograde double-labeling study in the cat

The double-labeling technique based on the retrograde axonal transport of fluorescent tracers (Evans blue, EB; Fast blue, FB; Nuclear yellow, NY) was used in the cat in order to investigate the occurence of axonal branching in the periaqueductal gray (PAG) projections to some thalamic nuclei (n. ventralis postero-lateralis, VPL; n. ventralis postero-medialis, VPM; n. parafascicularis, Pf). In a first group of cats, FB and EB were injected, respectively, within the right and left VPM. In another two groups of cats, FB injections into Pf were combined with either EB or NY injections within VPL or VPM. Double-labeled neurons were found within the PAG only in the animals of the first group. The present results show that some PAG neurons project bilaterally to VPM by means of axons collaterals.     

Subcortical projections to the centromedian and parafascicular thalamic nuclei in the cat

The primary objective of this study is to identify the totality of input to the centromedian and parafascicular (CM-Pf) thalamic nuclear complex. The subcortical projections upon the CM-Pf complex were studied in the cat with three different retrograde tracers. The tracers used were unconjugated horseradish peroxidase (HRP), horseradish peroxidase conjugated to wheat germ agglutinin (WGA-HRP), and rhodamine-labeled fluorescent latex microspheres (RFM). Numerous subcortical structures or substructures contained labeled neurons with all three tracing techniques. These labeled structures included the central nucleus of the amygdala; the entopeduncular nucleus; the globus pallidus; the reticular and ventral lateral geniculate nuclei of the thalamus; parts of the hypothalamus including the dorsal, lateral, and posterior hypothalamic areas and the ventromedial and parvicellular nuclei; the zona incerta and fields of Forel; parts of the substantia nigra including the pars reticularis and pars lateralis, and the retrorubral area; the pretectum; the intermediate and deep layers of the superior colliculus; the periaqueductal gray; the dorsal nucleus of the raphe; portions of the reticular formation, including the mesencephalic, pontis oralis, pontis caudalis, gigantocellularis, ventralis, and lateralis reticular nuclei; the nucleus cuneiformis; the marginal nucleus of the brachium conjunctivum; the locus coeruleus; portions of the trigeminal complex, including the principal sensory and spinal nuclei; portions of the vestibular complex, including the lateral division of the superior nucleus and the medial nucleus; deep cerebellar nuclei, including the medial and lateral cerebellar nuclei; and lamina VII of the cervical spinal cord. Moreover, the WGA-HRP and rhodamine methods (known to be more sensitive than the HRP method) revealed several afferent sources not shown by HRP: the anterior hypothalamic area, ventral tegmental area, lateral division of the superior vestibular nucleus, nucleus interpositus, and the nucleus praepositus hypoglossi. Also, the rhodamine method revealed labeled neurons in laminae V and VI of the cervical spinal cord.     

Spinal neurons which project to the periaqueductal gray and the medullary reticular formation via axon collaterals: a double-label fluorescence study in the rat

Fluorescent retrograde double-labeling methods were used in which Fast blue and Nuclear yellow or Diamidino yellow dihydrochloride were injected into the midbrain periaqueductal gray (PAG) and medullary reticular formation (MRF). Double-labeled neurons were most frequently observed in the lateral part of lamina V, in laminae VII, VIII and X and in the lateral cervical and lateral spinal nuclei. The data demonstrate that some spinal neurons project to both the PAG and the MRF via axon collaterals.     

Raphe and periaqueductal gray induced suppression of non-nociceptive neuronal responses in the dorsal column nuclei and trigeminal sub-nucleus caudalis

Electrical stimulation of the feline periaqueductal gray matter and nucleus raphe magnus was found to inhibit the firing of trigeminal sub-nucleus caualis nocipeptive neurons, as has been previously reported. However, stimulation at these same sites using similar current intensities was also found to be equally effective in inhibiting the responses of non-nociceptive neurons in both nucleus caudalis and in the dorsal column nuclei.     

Spinal afferents to functionally distinct periaqueductal gray columns in the rat: An anterograde and retrograde tracing study

The segmental and laminar organization of spinal projections to the functionally distinct ventrolateral (vlPAG) and lateral periaqueductal gray (lPAG) columns was examined by using retrograde and anterograde tracing techniques. It was found 1) that spinal input to both vlPAG and lPAG columns arose predominantly from neurons in the upper cervical (C1–4) and sacral spinal cord; 2) that there was a topographical separation of vl-PAG projecting and lPAG-projecting neurons within the upper cervical spinal cord; but 3) that below spinal segment C4, vlPAG-projecting and lPAG-projecting spinal neurons were similarly distributed, predominantly within contralateral lamina I, the nucleus of the dorsolateral fasciculus (the lateral spinal nucleus) and the lateral (reticular) part of lamina V. Consistent with the retrograde results, the greatest density of anterograde label, within both the vlPAG and lPAG, was found after tracer injections made either in the superficial or deep dorsal horn of the upper cervical spinal cord. Tracer injections made within the thoraco-lumbar spinal cord revealed that the vlPAG column received a convergent input from both the superficial and deep dorsal horn. However, thoraco-lumbar input to the lPAG was found to arise uniquely from the superficial dorsal horn; whereas the deep dorsal horn was found to innervate the “juxta-aqueductal” PAG region rather than projecting to the IPAG. These findings suggest that similar to spino-parabrachial projections, spinal projections to the lPAG (and juxta-aqueductal PAG) are topographically organised, with distinct subgroups of spinal neurons projecting to specific lPAG or juxta-aqueductal PAG subregions. In contrast, the vlPAG receives a convergent spinal input which arises from the superficial and deep dorsal horn of cervical, thoracic, lumbar, and sacral spinal segments. J. Comp. Neurol. 385:207–229, 1997. © 1997 Wiley-Liss, Inc.     

Influences of locus ceruleus, raphe dorsalis, and periaqueductal gray matter on somatosensory-recipient thalamic nuclei

Spontaneous and evoked discharge of neurons in the nucleus ventralis posterolateralis (VPL) and spontaneous discharge of neurons in the posterior group and nucleus lateralis posterior (LP) were conditioned by brief trains of stimuli to the locus ceruleus (LC), raphe dorsalis (RD), and periaqueductal gray matter (PAG) in cats anesthetized with pentobarbital or ketamine. Stimulation of LC and RD was without effect on VPL neurons, but induced a long-latency, long-lasting inhibition of LP neurons. Stimulation of the PAG induced marked inhibition of the firing of neurons in all three thalamic nuclei. No differences were found between cats anesthetized with ketamine or pentobarbital.     

Cortical and brain stem afferents to the ventral thalamic nuclei of the cat demonstrated by retrograde axonal transport of horseradish peroxidase

After horseradish peroxidase (HRP) injections into various parts of the ventral thalamic nuclear group and its adjacent areas, the distribution of labeled neurons was compared in the cerebral cortex, basal ganglia, and the brain stem. The major differences in distribution patterns were as follows: Injections of HRP into the lateral or ventrolateral portions of the ventroanterior and ventrolateral nuclear complex of the thalamus (VA-VL) produced retrogradely labeled neurons consistently in area 4 gamma (lateral part of the anterior and posterior sigmoid gyri, lateral sigmoid gyrus and the lateral fundus of the cruciate sulcus), the medial division of posterior thalamic group (POm), suprageniculate nucleus (SG) and anterior pretectal nucleus ipsilaterally, and in the nucleus Z of the vestibular nuclear complex bilaterally. Injections into the medial or dorsomedial portion of the VA-VL resulted in labeled neurons within the areas 6a beta (medial part of the anterior sigmoid gyrus), 6a delta (anterior part of ventral bank of buried cruciate sulcus), 6 if. fu (posterior part of the bank), fundus of the presylvian sulcus (area 6a beta), medial part of the nucleus lateralis posterior of thalamus and nucleus centralis dorsalis ipsilaterally, and in the entopeduncular nucleus (EPN) and medial pretectal nucleus bilaterally. Only a few neurons were present in the contralateral area 6a delta. After HRP injections into the ventral medial nucleus (VM), major labeled neurons were observed in the gyrus proreus, area 6a beta (mainly in the medial bank of the presylvian sulcus), and EPN ipsilaterally, and in the medial pretectal nucleus and substantia nigra bilaterally. Following HRP injections into the centre médian nucleus (CM), major labeled neurons were found in the areas 4 gamma, 6a beta, and the orbital gyrus ipsilaterally, and in the EPN, rostral and rostrolateral parts of the thalamic reticular nucleus, locus ceruleus, nucleus reticularis pontis oralis et caudalis and nucleus prepositus hypoglossi bilaterally. The contralateral intercalatus nucleus also possessed labeled neurons. With HRP injections into the paracentral and centrolateral nuclei, labeled neurons were observed in the gyrus proreus and the cortical areas between the caudal presylvian sulcus and anterior rhinal sulcus ipsilaterally, and in the nuclei interstitialis and Darkschewitsch bilaterally. Minor differences in the distribution pattern were observed in the superior colliculus, periaqueductal gray, mesencephalic and medullary reticular formations, and vestibular nuclei in all cases of injections.     

Damage to the dorsomedial thalamic nucleus,central lateral intralaminar thalamic nucleus,and midline thalamic nuclei on the right-side impair executive function and attention under conditions of high demand but not low demand

This study reports a patient, OG, with a unilateral right-sided thalamic lesion. High resolution 3T magnetic resonance imaging revealed damage to the parvicellular and magnocellular subdivisions of the dorsomedial thalamus (DMT), the central lateral intralaminar nucleus (also known as the paralamellar DMT), the paraventricular and the central medial midline thalamic nuclei. According to the neuropsychological literature, the DMT, the midline and intralaminar thalamic nuclei influence a wide array of cognitive functions by virtue of their modulatory influences on executive function and attention, and this is particularly indicated under conditions of low arousal or high cognitive demand. We explored this prediction in OG, and compared his performance on a range of low and high demand versions of tests that tapped executive function and attention to a group of 6 age- and IQ-matched controls. OG, without exception, significantly underperformed on the high-demand attention and executive function tasks, but performed normally on the low-demand versions. These findings extend and refine current understanding of the effects of thalamic lesion on attention and executive function.     

Identification and somatotopic organization of nuclei projecting via the dorsolateral funiculus in rats: A retrograde tracing study using HRP slow-release gels

Horseradish peroxidase (HRP) in slow-release gel was unilaterally implanted in the transected dorsolateral funiculus (DLF) of either cervical, midthoracic, lumbar or sacral spinal cord levels of adult male rats. Cell mappings were made of all brain areas projecting through the DLF. Following cervical implants, dense labelling was observed within a band along the dorsolateral border of the inferior olive, locus coeruleus, the paralemniscal reticular formation, the mesencephalic central gray, the red nucleus and the paraventricular nucleus of the hypothalamus. The DLF projection from the central gray consisted of a rostrocaudal line of cells extending from the level of the mesen-/diencephalic junction to the rostral red nucleus. The somatotopic organization of labelled nuclei was assessed by comparing the pattern of filled cells following HRP implants at various cord levels. The potential role of these areas in pain modulation was discussed in light of their descending projections through the DLF of the spinal cord.     

Effects of periaqueductal gray and raphe magnus stimulation on the responses of spinocervical and other ascending projection neurons to non-noxious inputs

Experiments were performed in cats anesthetized with α-chloralose to examine the effects of stimulating in the periaqueductal gray (PAG) and nucleus raphe magnus (NRM) on the responses of spinocervical cells and unidentified ascending projection neurons to non-noxious peripheral stimuli. Peripheral stimuli consisted of low amplitude sinusoidal displacements applied to either the glabrous skin or the hairy skin of the neuron's receptive field. Stimulating in either the periaqueductal gray or nucleus raphe magnus reduced the impulse activity of most neurons in both groups. By applying brainstem stimuli at various phases of the sinusoidal peripheral stimulus, it was demonstrated that the effects of stimulating either the PAG or NRM on the responses of both types of neurons was dependent on the timing of the electrical stimuli relative to the peripheral input. The effects of stimulating in the PAG and NRM on the responses of these cells to non-noxious stimuli were reversibly blocked by naloxone. It was concluded that stimulating in the nucleus raphe magnus and in the periaqueductal gray can produce dramatic modifications in the responses of spinocervical cells and unidentified ascending projection neurons to non-noxious peripheral stimuli, suggesting a role for these descending systems in non-noxious information processing.     

Axon collaterals of pontine taste area neurons project to the posterior ventromedial thalamic nucleus and to the gustatory neocortex

Retrograde axonal transport of fluorescent dyes was used to demonstrate collateral projections from neurons of the pontine taste area (PTA) to gustatory-responsive areas of the posterior ventromedial thalamic nucleus (VPM), and to the gustatory neocortex (GN) of the rat. Dual-labeled PTA neurons were reliably observed following application of two different fluorescent dyes to the GN and to VPM thalamus. Dye injections into the GN and into thalamic regions surrounding the VPM nucleus, the bed nucleus of stria terminalis or the infralimbic neocortex, did not result in dual-labeled cells within the PTA. This finding suggests that gustatory information may be relayed simultaneously and specifically to VPM thalamus and to the GN via collateral axons of PTA neurons.     

Relative contributions of the nucleus raphe magnus and adjacent medullary reticular formation to the inhibition by stimulation in the periaqueductal gray of a spinal nociceptive reflex in the pentobarbital-anesthetized rat

Ammonia intoxication decreases the hyperpolarizing action of postsynaptic inhibition. This study examines the metabolic state of the spinal cord during this effect of ammonia intoxication on spinal motoneurons. ATP, ADP, AMP, the adenylate energy charge, glucose, PCr, pyruvate, alpha-ketoglutarate and glutamate were unchanged during the effect of ammonia on the hyperpolarizing action of postsynaptic inhibition. NH4+, glutamine and lactate were increased. Ammonia intoxication affected postsynaptic inhibition without changes of the resting membrane potential, the neuron input resistance, the action potential and EPSPs. The encephalopathy caused by ammonia intoxication is known to occur without an alteration of the tissue energy state. The effect of ammonia intoxication on postsynaptic inhibition can be considered as a cause of the encephalopathy because postsynaptic inhibition is altered without a change of the tissue energy state, the resting membrane potential, the whole neuron resistance, the action potential and EPSPs.     

Cortical modulation of neuronal activity in the cat's lateral geniculate and perigeniculate nuclei

The cortico-thalamic influence on spontaneous and visually evoked activity of single cells in the dorsal lateral geniculate (LGN) and perigeniculate (PGN) nuclei were examined in unanesthetized cats with pretrigeminal brainstem transections by means of reversible cooling of cortical areas 17 and 18. The spatio-temporal characteristic of cells' RFs was tested with light spot randomly presented at different points along the receptive field axis. The cessation of cortical input decreased spontaneous activity of most of the LGN cells (64%; as compared to 36% with increased background firing). Similarly, their visually evoked responses were reduced (70% cells; compared to 24% with increased response) and extent of central excitatory domains diminished. In contrast, the majority of PGN neurons increased their spontaneous activity (62%; compared to 38% with decreased firing rate). Cortical cooling resulted also in a decrease of the ON and OFF central responses of most PGN cells (55%; as compared to 20% with increased responses). The described effects were more pronounced within the population of cells in X than in Y pathway. Although the removal of descending cortical excitation disturbed the balance of activity within the network of thalamic cells the gain of the geniculate relay was preserved. We conclude that the main role exerted by the cortico-thalamic pathway serves facilitation of the ascending retino-cortical flow of visual information at the level of lateral geniculate nucleus.     

Serotonergic projections from the midbrain periaqueductal gray and nucleus raphe dorsalis to the nucleus parafasccicularis of the thalamus

By a double-labeling method combining the retrograde tracing of horseradish peroxidase and the immunocytochemical technique, serotonin-like immunoreactive neurons in the midbrain periaqueductal gray (PAG) and nucleus raphe dorsalis (DR) of the rat were observed to send projection fibers to the nucleus parafascicularis of the thalamus bilaterally with an ipsilateral dominance. These serotonin-containing projecting neurons were observed mainly at the middle-caudal levels of the ventrolateral subdivision of the PAG and less at the middle-rostral levels of the DR.     

The topographical organization of neurons in the rat medial frontal, insular and olfactory cortex projecting to the solitary nucleus, olfactory bulb, periaqueductal gray and superior colliculus

In 19 rats two different retrograde tracers (Fast Blue, Diamidino Yellow, Rhodamine-labeled latex microspheres, or wheat germ agglutinin conjugated with HRP) were injected into the solitary nucleus (NTS) and either the olfactory bulb (OB), periaqueductal gray (PAG) or superior colliculus (SC). The pattern of retrogradely labeled neurons in the medial frontal, insular and olfactory cortices was examined to determine the topographical organization of the cell populations projecting to these subcortical targets and the extent to which they overlapped. In the medial frontal cortex (MFC) SC projections originated most dorsally, while NTS and OB projections originated most ventrally and exhibited slight overlap. PAG projections originated from virtually the entire MFC and overlapped with cells projecting to the OB, NTS and SC. These results are consistent with the role of dorsal MFC as the rat's frontal eye field and the ventral MFC as a visceral motor area. Laterally, in the insular cortex there was virtually complete overlap between cells projecting to the NTS and PAG. The extensive overlap of PAG projections with NTS projections medially and laterally and with SC projections medially suggests the PAG is involved in a variety of brain visceral and somatic functions. In the piriform cortex there was overlap between cells projecting to the OB and cells projecting to the SC; the cells projecting to the SC were located in the endopiriform nucleus, and may provide a substrate for orienting responses to odors.     

Organization of medullary adrenergic and noradrenergic projections to the periaqueductal gray matter in the rat.

The periaqueductal or midbrain central gray matter (CG) in the rat contains a dense network of adrenergic and noradrenergic fibers. We examined the origin of this innervation by using retrograde and anterograde axonal tracers combined with immunohistochemistry for the catecholamine biosynthetic enzymes tyrosine hydroxylase (TH), dopamine beta-hydroxylase (DBH), and phenylethanolamine N-methyltransferase (PNMT). Following injections of the fluorescent tracers Fast Blue or Fluorogold into the CG, double-labeled neurons in the medulla were identified mainly in the noradrenergic A1 group in the caudal ventrolateral medulla (VLM) and A2 group in the medial part of the nucleus of the solitary tract (NTS); and in the adrenergic C1 group in the rostral ventrolateral medulla and C3 group in the rostral dorsomedial medulla. Injections of Phaseolus vulgaris-leucoagglutinin (PHA-L) into these cell groups resulted in a distinct pattern of axonal labeling in various subdivisions of the CG. Anterogradely labeled fibers originating in the medial NTS were predominantly found in the lateral portion of the dorsal raphe nucleus and in the adjacent part of the lateroventral CG (CGlv). Following PHA-L injections into the C3 region the anterogradely labeled fibers were diffusely distributed in the CGlv and the dorsal raphe nucleus at caudal levels, but rostrally tended to be located laterally in the CGlv. In contrast, ascending fibers from the caudal and rostral VLM terminated in the rostral dorsal part of the CGlv and in the dorsal nucleus of the CG, whereas ventral parts of the CG, including the dorsal raphe nucleus, contained few afferent fibers. Double-label studies with antisera against DBH and PNMT confirmed that noradrenergic neurons in the A1 and A2 groups and adrenergic neurons in the C1 and C3 groups contributed to these innervation patterns in the CGlv. Noradrenergic and adrenergic projections from the medulla to the CG may play an important role in a variety of autonomic, sensory and behavioral processes.     

Conditioned place aversion organized in the dorsal periaqueductal gray recruits the laterodorsal nucleus of the thalamus and the basolateral amygdala

The amygdala-ventral periaqueductal gray circuit is crucial for the expression of contextual conditioned fear. However, little is known about the neural circuits activated when the stimulation of the dorsal periaqueductal gray (dPAG) is used as unconditioned stimulus (US) in conditioned fear paradigms. The present paper examines the Fos-protein distribution in the brain of rats submitted to a conditioned place aversion (CPA) paradigm using the dPAG chemical stimulation with semicarbazide (SMC), an inhibitor of the GABA synthesizing enzyme, as US and the quadrant of an arena where the drug was injected as the paired neutral stimulus. Our results show that CPA associated with SMC injections caused a significant Fos labeling in the laterodorsal nucleus of the thalamus (LD), basolateral nucleus of amygdala (BLA) and in the dorsomedial PAG (dmPAG). This pattern of brain activation is clearly different from the neural substrates of the classical fear conditioning reported in the literature. Moreover, this paper shows that CPA with the use of chemical stimulation of the dPAG could be used as an experimental model of panic disorder with agoraphobia in the extent that panic attacks repeatedly associated with specific contexts may turn in this condition in the clinics. This condition activates the BLA probably through the LD. Besides, it indicates that the dPAG can be the link between amygdala and the brainstem motor regions that controls CPA when dPAG stimulation is used as US instead of footshocks. From this evidence we suggest that a loop dPAG-LD-BLA-dPAG is activated during the panic disorder with agoraphobia.     

Dorsal mesencephalic projections to pons, medulla, and spinal cord in the cat: limbic and non-limbic components.

The vertebrate dorsal mesencephalon consists of the superior colliculus, the dorsal portion of the periaqueductal gray, and the mesencephalic trigeminal neurons in between. These structures, via their descending pathways, take part in various behavioral responses to environmental stimuli. This study was undertaken to compare the origins and trajectories of these pathways in the cat. Injections of horseradish peroxidase into the cervical spinal cord and upper medullary medial tegmentum retrogradely labeled cells mainly in the contralateral intermediate and deep superior colliculus, and in the ipsilateral dorsal and lateral periaqueductal gray and adjacent tegmentum. Only injections in the medullary lateral tegmental field labeled mesencephalic trigeminal neurons ipsilaterally. Autoradiographic tracing results, based on injections across the dorsal mesencephalon, revealed three efferent fiberstreams. A massive first fiberstream (limbic pathway), consisting of thin fibers, descended ipsilaterally from the dorsal and lateral periaqueductal gray and adjacent superior colliculus through the mesencephalic and pontine lateral tegmentum, terminating in these areas as well as in the ventral third of the caudal pontine and medullary medial tegmentum. A few fibers from the dorsal periaqueductal gray matter (PAG) were distributed bilaterally to the dorsal vagal, solitary, and retroambiguus nuclei. The second fiberstream (the predorsal bundle) descended contralaterally from the superior colliculus (SC) and consisted of both thick and thin labeled fibers. The thin fibers terminated bilaterally in the dorsomedial nucleus reticularis tegmenti pontis and the medial half of the caudal medial accessory inferior olive. The thick fibers targeted the contralateral dorsal two thirds of the caudal pontine and medullary medial tegmental fields, and the facial, abducens, lateral reticular, subtrigeminal, and prepositus hypoglossi nuclei. A few fibers recrossed the midline to terminate in the ipsilateral medial tegmentum. Caudal to the obex, fibers terminated laterally in the tegmentum and upper cervical intermediate zone. From the lateral SC, fibers terminated bilaterally in the lateral tegmental fields of the pons and medulla and lateral facial subnuclei. The third fiberstream (mesencephalic trigeminal or Probst tract) terminated in the supratrigeminal and motor trigeminal nuclei, and laterally in the tegmentum and upper cervical intermediate zone. In summary, neurons in the PAG and in the deep layers of the SC give rise to a massive ipsilateral descending pathway, in which a medial-to-lateral organization exists. A similar topographical pattern occurs in the crossed SC projections. The possibility that these completely different descending systems cooperate in producing specific defensive behaviors is discussed.