Search

The spinothalamic tract (STT), well known for its role in the relay of information about noxe, temperature, and crude touch, is usually associated with projections from lamina 1, but spinothalamic neurons in other laminae have also been reported. In cat, no complete overview exists of the precise location and number of spinal cells that project to the thalamus In the present study the laminar distribution of retrogradely labeled cells in all spinal segments (C1-Coc2) was investigated after large WGA-HRP injections in the thalamus. The results show that this distribution of STT cells differed greatly between the different spinal segments. Quantitative analysis showed that there exist at least five separate clusters of spinothalamic neurons. Lamina I neurons in cluster A and lamina V neurons in cluster B are mainly found contralaterally throughout the length of the spinal cord. Cluster C neurons are located bilaterally in the ventrolateral part of laminae VI-VII and lamina VIII of the C1-C3 spinal cord. Cluster D neurons were found contralaterally in lamina VI in the C1-C2 segments, and cluster E neurons were located mainly contralaterally in the medial part of laminae VI-VII and lamina VIII of the lumbosacral cord. Most spinothalamic neurons are not located in the enlargements and most spinothalamic neurons are not located in lamina 1, as suggested by several other authors. The location of the spinothalamic neurons shows remarkable similarities, but also differences, with the location of spino-periaqueductal gray neurons.

By far, the strongest spinal cord projections to periaqueductal gray (PAG) and thalamus originate from the upper three cervical segments, but their precise organization and function are not known. In the present study in cat, tracer injections in PAG or in thalamus resulted in more than 2400 labeled cells, mainly contralaterally, in the first three cervical segments (C1-C3), in a 1:4 series of sections, excluding cells in the dorsal column and lateral cervical nuclei. These cells represent about 30% of all neurons in the entire spinal cord projecting to PAG and about 45% of all spinothalamic neurons. About half of the C1-C3 PAG and C1-C3 thalamic neurons were clustered laterally in the ventral horn (C1-3v1), bilaterally, with a slight ipsilateral preponderance. The highest numbers Of C1-3v1-PAG and C1-3v1-thalamic cells were found in C1, with the greatest density rostrocaudal ly in the middle part of C 1. A concept is put forward that C1-3v1 cells relay information from all levels of the cord to PAG and/or thalamus, although the processing of specific information from upper neck muscles and tendons or facet joints might also play a role. (c) 2005 Elsevier B.V. All rights reserved.

The spinothalamic tract, well known for its role in nociception, is the most frequently studied ascending pathway originating from the spinal cord. It is known that spinothalamic neurons are located in all segments of the spinal cord, but in most mammals the total number of spinothalamic neurons is not known. In three cats, after large wheat germ agglutinin-conjugated horseradish peroxidase injections involving all parts (one case) or almost all parts of the thalamus (two cases), the number of retrogradely labeled profiles was counted in a 1:4 series of sections of all spinal segments from C I to Coc2. After applying the correction factor of Abercrombie (Anat. Rec. 94 (1946) 239), it appears that a total of 12,000 cells in the spinal cord project to the thalamus. (C) 2004 Elsevier Ireland Ltd. All rights reserved.

The possibility of collateral segmental actions of spinocervical tract (SCT) neurones upon interneurones with input from cutaneous and group II muscle afferents was investigated in deeply anaesthetized cats. To this end, intracellular and/or extracellular recordings were made from 35 dorsal horn and 15 intermediate zone interneurones in midlumbar segments of the spinal cord and effects of stimulation of the ipsilateral dorso-lateral funiculus (DLF) at C3 and C1 levels, i.e. below and above the lateral cervical nucleus where axons of SCT cells terminate, were compared. The stimuli applied at the C3 segment were within the range of stimuli (50-100 microA) required for antidromic activation of SCT neurones in the same experiment. Those applied at the C segment (200-500 microA) were at least 3 times stronger than C3 stimuli. Under the same experimental conditions, long ascending and descending tract neurones (dorsal spino-cerebellar and rubro-spinal tract neurones) with axons in the DLF were activated at similar thresholds from the C and C3 segments. Intracellular recordings were made from 29 interneurones of which 19 (65%) were dorsal horn and 10 (35%) were intermediate zone interneurones. Excitatory postsynaptic potentials (EPSPs) evoked by single stimuli applied at the C3 segment, but not the C segment, were found in 14 (48%) of those interneurones; their latencies (3.0-5.7 ms) and frequency following with only minimal temporal facilitation were as required for potentials being evoked monosynaptically by the fastest conducting SCT neurones. Extracellular recordings were made from 30 interneurones (24 dorsal horn and 6 intermediate zone interneurones), and in these neurones spike potentials induced from the C3, but not from the C segment, were evoked only by short trains of stimuli. However, their latencies from the first effective stimulus (4.3-5.4 ms) were compatible with mono- or oligosynaptically mediated collateral actions of SCT neurones. They were found in 10 (33%) of the 30 investigated interneurones. Similar effects of C3 stimuli were found in similar proportions of dorsal horn interneurones and intermediate zone interneurones. Indications were also found for synaptic actions evoked by C3 stimuli that could not be attributed to direct collateral actions of SCT neurones. In some intracellularly recorded dorsal horn interneurones, short-latency EPSPs were evoked from the C3 segment by the 2nd or 3rd stimulus in the train, but not by single stimuli. In other dorsal horn and intermediate zone interneurones, inhibitory postsynaptic potentials (IPSPs) were evoked from the C3 segment at minimal latencies (2.7-3.2 ms), which might be too short to allow their mediation via SCT neurones. We conclude that SCT neurones might be used to forward information from muscle group II and cutaneous afferents not only to neurones in the lateral cervical nucleus and via them to thalamus and cerebral cortex but also to interneurones in spinal reflex pathways. Thereby reflex actions evoked from group II and cutaneous afferents might be co-ordinated with responses mediated by supraspinal neurones. We conclude also that dorsal horn and intermediate zone mid-lumbar interneurones might contribute to the previously reported di-and poly-synaptic excitation or inhibition of postsynaptic dorsal column (PSDC), spinothalamic tract (STT) and spinomesencephalic tract (SMT) neurones by collateral actions of SCT cells. Thereby these interneurones might contribute to the co-ordination of responses mediated by various populations of supraspinal neurones.

Terminal fibers anterogradely labeld with Phaseolus vulgaris leucoagglutinin (PHA-L) were observed in the locus coeruleus in the brainstem of the cat and the cynomolgus monkey following injections in lamina I of the spinal or medullary dorsal horn. Thus, thermoreceptive- or nociceptive-specific lamina I cells that project to the locus coeruleus could directly influence arousal, vigilance, and the descending control of spinal integration.

The lateral ventral horn. of the upper cervical (C(1-3vl)) cord in rat, cat and monkey contains many cells that project to the periaqueductal gray matter (PAG). Until now it was assumed that these cells only project to the ventrolateral part of the caudal PAG. Because the ventrolateral caudal PAG is involved in quiescence and hypotension, it was hypothesized that the C(1-3vl)-PAG projecting cells play a role in immobility behavior, possibly activated by neck muscle afferents. However, in the present anterograde and retrograde study in cat we showed that C(1-3vl) cells do not only target the caudal PAG, but terminate even more abundantly in the intermediate and rostral parts of the PAG. There, projections target the ventrolateral column, like in the caudal PAG, but also terminate in the lateral and dorsomedial columns. This finding, combined with the current, albeit limited, physiological data on C(1-3vl)-PAG and C(1-3vl)-thalamic projecting cells, sheds a new light on the possible functions of C(1-3vl) cells. it might be that the C(1-3vl) cells with complex response properties and large receptive fields have a relay function similar to cells in the dorsal column nuclei, lateral cervical or central cervical nucleus. Other C(1-3vl) cells might receive somatic or visceral input that was never tested before, or perhaps other input, like vestibular information. It might also be that these cells project to the more caudal cord to modulate visceral input, with ascending collaterals to more rostrally, located structures, including mesencephalon and thalamus. (C) 2009 Elsevier B.V. All rights reserved.

The location of the ascending axons of spinal lamina I cells was studied in cats that received injections of Phaseolus vulgaris leucoagglutinin (PHA-L) in the superficial dorsal horn of the cervical or lumbosacral enlargement. Lamina I axons that could be ascribed to the spinothalamic tract (STT) were of particular interest. The cases were divided into three sets: in seven optimal cases the injections were restricted to lamina I; in ten nominal cases the injections involved laminae I-II or laminae I-III and occasionally lamina IV; and in eight mixed cases laminae I-V were injected. Since ipsilateral propriospinal and bilateral supraspinal axons originate from laminae I and V, but only ipsilateral propriospinal axons from laminae II-IV, this categorization facilitated a comparative analysis. Ascending axons labeled immunohistochemically with avidin/Texas Red were observed in oblique transverse sections from the C1, C3/4, T6, T12, and L3/4 levels. Incidental axonal labeling occurred in the ipsilateral dorsal columns because of passing primary afferent fiber uptake and, in nominal and mixed cases with involvement of laminae III-IV, in the superficial dorsolateral funiculus at the location of the spinocervical tract. Ipsilateral ascending lamina I axons in optimal cases were located in Lissauer's tract and in the white matter adjacent to the dorsal horn. Since these appeared to terminate in lamina I, and few remained at C1, they were ascribed to propriospinal projections. Contralateral ascending lamina I axons in optimal and nominal cases were distributed throughout the dorsal and ventral portions of the lateral funiculus (LF), but, despite considerable variability between animals in their location and dispersion, they were consistently concentrated in the middle of the LF (i.e., at the level of the central canal). This concentration was observed in a slightly more ventral location at C1, and a similar but weaker concentration of lamina I axons was located slightly more dorsally in C1 on the ipsilateral side. These supraspinal lamina I projections were ascribed to the spinomesencephalic tract (SMT) and to the STT. In mixed cases, additional ascending axons ascribed to lamina V cells were labeled in the ventrolateral and ventral funiculi. Many labeled axons were found in this region following a large injection of biocytin into lumbosacral laminae V-VIII in a supplementary case. These results thus together support previous descriptions of a dorsoventral distribution of STT axons according to laminar origin, but they contradict recent reports that lamina I axons ascend in the dorsolateral funiculus.(ABSTRACT TRUNCATED AT 400 WORDS)

1. The effects of electrical stimulation at different rostrocaudal levels of the midbrain, and at sites in the rostral medulla ipsilateral and contralateral to spinal recording sites, were evaluated against the responses of 46 cells belonging to the cat spinomesencephalic tract (SMT). 2. Inhibitory and/or excitatory effects of brain stem stimulation were observed on SMT cells that responded best (26 cells) or exclusively (12 cells) to noxious mechanical or thermal stimuli, as well as on 7 cells responding only to tap and/or stimulation of deep tissues. Recording sites for 32 cells were located in laminae V-VIII (27 cells) and laminae I-III (5 cells). 3. Midbrain stimulation sites were located in the superior colliculus, central gray (CG), red nucleus, and the midbrain reticular formation. Both inhibitory-only and excitatory-only effects were observed, although the most common effect of midbrain stimulation was excitation followed by inhibition (mixed effects). The effects of stimulation at different midbrain levels were determined for each cell. Stimulation in the caudal, middle, or rostral midbrain was often found to exert different effects on the same SMT cell. 4. Stimulation in the rostral medulla at sites located in nucleus raphe magnus (NRM), nucleus reticularis gigantocellularis, and nucleus reticularis magnocellularis produced the same complement of effects observed with midbrain stimulation. Excitation followed by inhibition was the most common effect observed. 5. Stimulus intensities required to produce excitatory or inhibitory effects from midbrain were 114 +/- 85 (SD) microA and 210 +/- 91 microA, respectively. Stimulus currents required to produce excitatory or inhibitory effects from medullary stimulation sites were 124 +/- 56 microA and 70 +/- 60 microA, respectively. The mean currents required to produce mixed effects were 221 +/- 120 microA (midbrain) and 127 +/- 71 microA (medulla). Increasing the stimulus intensity used to evaluate brain stem effects increased the magnitude and duration of effects for 33 cells. Mixed effects were observed on 11 cells at stimulus intensities greater than those required to produce inhibitory-only effects. 6. Significant differences were found between the latencies of excitation and inhibition produced from different brain stem levels. These differences suggest that midbrain and medullary stimulation activate descending pathways with a wide range of conduction velocities and/or supraspinal and spinal connectivities. 7. The spinal trajectory of pathways contributing to the varied effects of brain stem stimulation as well as the complex receptive fields (RFs) of SMT cells were evaluated by the placement of lesions in the cervical spinal cord.(ABSTRACT TRUNCATED AT 400 WORDS)

The spinothalamic tract, and especially its fibers originating in lamina 1, is the best known pathway for transmission of nociceptive information. On the other hand, different studies have suggested that more lamina I cells project to the parabrachial nuclei (PBN) and periaqueductal gray (PAG) than to the thalamus. The exact ratio of the number of lamina I projections to PBN, PAG and thalamus is not known, because comprehensive studies examining these three projections from all spinal segments, using the same tracers and counting methods, do not exist.In the present study, the differences in number and distribution of retrogradely labeled lamina I cells in each segment of the cat spinal cord (C1-Coc2) were determined after large wheat germ agglutinin-conjugated horseradish peroxidase (WGA-HRP) injections in either PBN, PAG or thalamus. We estimate that approximately 6000 lamina I cells project to PBN, 3000 to PAG and less than 1500 to the thalamus. Of the lamina I cells projecting to thalamus or PAG more than 80%, and of the lamina I-PBN cells approximately 60%, were located on the contralateral side. In all cases, most labeled lamina I cells were found in the upper two cervical segments and in the cervical and lumbar enlargements. (c) 2005 IBRO. Published by Elsevier Ltd. All rights reserved.

The midbrain periaqueductal gray matter (PAG) integrates the basic responses necessary for survival of individuals and species. Examples are defense behaviors such as fight, flight, and freezing, but also sexual behavior, vocalization, and micturition. To control these behaviors the PAG depends on strong input from more rostrally located limbic structures, as well as from afferent input from the lower brainstem and spinal cord. Mouton and Holstege (2000, J Comp Neurol 428:389-410) showed that there exist at least five different groups of spino-PAG neurons, each of which is thought to subserve a specific function. The lateral cervical nucleus (LCN) in the upper cervical cord is not among these five groups. The LCN relays information from hair receptors and noxious information and projects strongly to the contralateral ventroposterior and posterior regions of thalamus and to intermediate and deep tectal layers. The question is whether the LCN also projects to the PAG. The present study in cat, using retrograde and anterograde tracing techniques, showed that neurons located in the lateral two-thirds of the LCN send fibers to the lateral part of the PAG, predominantly at rostrocaudal levels A0.6-P0.2. This part of the PAG is known to be involved in flight behavior. A concept is put forward according to which the LCN-PAG pathway alerts the animal about the presence of cutaneous stimuli that might represent danger, necessitating flight. J. Comp. Neurol. 471:434-445, 2004.

The periaqueductal gray is well known for its involvement in nociception control, but it also plays an important role in the emotional motor system. To accomplish these functions the periaqueductal gray receives input from the limbic system and from the caudal brainstem and spinal cord. Earlier studies gave the impression that the majority of the periaqueductal gray projecting cells in caudal brainstem and spinal cord are located in the contralateral lamina I, which is involved in nociception. The present study in the cat, however, demonstrates that of all periaqueductal gray projecting neurons in the contralateral caudal medulla less than 7% was located in lamina I. Of the spinal periaqueductal gray projecting neurons less than 29% was located in lamina I. However, within the spinal cord large segmental differences exist: in few segments of the enlargements the lamina I-periaqueductal gray projecting neurons represent a majority. In conclusion, although the lamina I-periaqueductal gray projection is a very important nociceptive pathway, it constitutes only a limited part of the total projection from the caudal medulla and spinal cord to the periaqueductal gray. These results suggest that a targe portion of the medullo- and spino-periaqueductal gray pathways conveys information other than nociception. (C) 2001 Elsevier Science Inc.

The periaqueductal gray (FAG) is involved in motor activities, such as movements of the neck, back and hind limbs, cardiovascular regulation, micturition, vocalization, and mating behavior, as well as in nociception control. To accomplish these functions the FAG uses information from other parts of the limbic system, from the lower brainstem, and from the spinal cord. To study the ascending projections from the spinal cord to the PAG, tracer was injected in different parts of the FAG, and the number of retrogradely labeled neurons were counted for each spinal segment. Results show that large segmental differences exist in the number of FAG projecting neurons throughout the length of the spinal cord and that different parts of the spinal cord project to specific areas in the FAG. (C) 1997 Elsevier Science Ireland Ltd.

Extracellular microelectrode recordings were made from spinomesencephalic tract (SMT) neurons in the lumbosacral spinal cord of cats anaesthetized with chloralose and paralysed with gallamine triethiodide. The SMT cells were antidromically fired from the posterolateral parts of the superior colliculus and the intercollicular region, were located in laminae IV to VIII, and had response properties and axonal conduction velocities similar to those described previously. The effects of stimulating the dorsolateral funiculus of the cervical cord at C3 and rostral C1, below and above the termination of spinocervical tract (SCT) axons in the lateral cervical nucleus, were examined on 33 SMT cells. The strength of stimulation was adjusted so that at C3 it was above threshold for antidromic activation of SCT cells and at C1 was below threshold for activation of the same cells. Seven (21%) SMT neurons were excited from C3 but not from C1. The remaining 26 (79%) were excited from both C3 and rostral C1 and 23 (70% of these) were excited significantly more from C3. That is, 91% of the total sample were either excited only from C3 or more strongly from C3 than from rostral C1. We discuss the possible neuronal systems involved and conclude that the greater excitatory effects from C3 are most likely due to antidromic activation of the SCT. The shortest latency effects from C3 indicate a monosynaptic linkage between SCT cells with the fastest axons and the SMT. The longer latency actions may be due to monosynaptic connexions from SCT cells with slower conducting axons, to di- or polysynaptic actions from SCT cells with fast axons, or a combination of both. SMT cells are another population of spinal neurons, in addition to postsynaptic dorsal column, spinothalamic and dorsal horn spinocerebellar neurons, which receive excitation via SCT collaterals.

The periaqueductal gray (FAG) plays a crucial role in reproductive behavior. The present study investigates whether lumbosacral FAG-projecting neurons contain estrogen receptors. In four ovariectomized adult female cats, injections with cholera toxin subunit (CTb) were made into the FAG to retrogradely label FAG projecting neurons in the lumbosacral cord. Estrogen receptor immunoreactive ER-IR neurons in the lumbosacral cord were identified immunohistochemically using the antibody H222, PAG-projecting neurons that were immunoreactive for the estrogen receptor were very scarce, and predominantly present in the medial part of the ventral horn. The results indicate that only very few of the neurons relaying information from the urogenital organs to the FAG contain estrogen receptors. (C) 1997 Elsevier Science Ireland Ltd.

The nucleus retroambiguus (NRA) projects to distinct brainstem and cervical and thoracic cord motoneuronal cell groups. The present paper describes NRA projections to distinct motoneuronal cell groups in the lumbar enlargement. Lumbosacral injections of wheat germ agglutinin-horseradish peroxidase (WGA-HRP) were made to localize and quantify the retrogradely labeled neurons in the caudal medullary lateral tegmentum. These injections were combined with spinal hemisections to distinguish between neurons having ipsi- or contralaterally descending axons. The NRA-lumbosacral fibers descend almost exclusively contralaterally, but neurons in areas surrounding the NRA project mainly ipsilaterally.In an anterograde tracing study, injections of WGA-HRP or tritiated leucine were made in the region of the NRA to determine the NRA. targets in the lumbosacral cord. Hemisections in C2 made it possible to distinguish between NRA projections and projections from neurons in the adjoining lateral tegmentum. The results show delicate NRA projections to distinct lumbosacral motoneuronal cell groups innervating specific hindlimb muscles (iliopsoas, adductors, and hamstrings) as well as axial muscles (medial longissimus and proximal tail muscles). The projection is bilateral, with a contralateral predominance. Ipsilaterally terminating fibers are derived from NRA neurons whose axons cross the midline at the level of the obex, descend through the contralateral spinal white matter, and recross at the level of termination. A conceptual description is presented in which the periaqueductal gray-NRA-lumbosacral projections form the final common pathway for lordosis in the cat. (C) 1995 Wiley-Liss, Inc.

Information concerning the rate of bladder filling is determined by receptors in the bladder wall and conveyed via afferent fibers in the pelvic nerve to sensory neurons in the lumbosacral cord. It was assumed that this information is relayed from the lumbosacral cord to a medial cell group in the dorsolateral pontine tegmentum, called the M-region, the pontine micturition center, or Barrington's nucleus. The M-region, in turn, projects via long descending pathways to the sacral parasympathetic motoneurons. In the present electron microscopic study, it was investigated in cats whether monosynaptic projections from lumbosacral neurons to the M-region indeed exist. Wheat-germ agglutinin-horseradish peroxidase injections were made into the lumbosacral, cord. Many retrogradely labeled dendrites and somata were found in the M-region, but no labeled terminals were found on retrogradely labeled dendrites or somata. Only a small number of anterogradely labeled terminals, which were filled with mainly round vesicles, contacted unlabeled dendrites in the M-region. In contrast, many more anterogradely labeled terminals, which were filled with mainly round and, to a limited extent, dense core vesicles and with asymmetrical synapses, were found on dendrites in the lateral part of the periaqueductal gray (PAG). Previously (Blok and Holstege [1994] Neurosci. Lett. 166:93-96), it was demonstrated that the lateral part of the FAG contains neurons projecting to the M-region. A concept for the central organization of the micturition reflex is presented in which ascending projections from the lumbosacral cord convey information on bladder filling to the FAG. When the bladder contains so much urine that voiding is necessary, the FAG, in turn, triggers the M-region. The M-region, however, also receives afferents from the preoptic area, which might be involved in the final decision to start micturition. (C) 1995 Wiley-Liss, Inc.

Response and receptive field properties were evaluated for 62 spinomesencephalic tract cells in the upper cervical spinal cord (C1-C3) of cats anesthetized with sodium pentobarbital and alpha-chloralose. Recordings were made from cells in laminae I-VIII and X contralateral to antidromic stimulating electrodes positioned in the rostral, caudal and intercollicular region of the midbrain. The mean antidromic threshold for all cells was 185 +/- 132 microA, and conduction velocities ranged from 2.3 to 38.6 m/sec. Twelve cells were backfired from both midbrain and diencephalic stimulation sites. Receptive fields ranged from simple, i.e., ipsilateral forelimb or face, to complex, i.e., excitatory and/or inhibitory responses from large portions of the body. Peripheral receptive fields included muscles, joints, cornea, dura, forelimbs, hind limbs, tail, and/or testicles. Five functional classes of cells were observed: (a) wide dynamic range (14 cells); (b) high threshold (2 cells): (c) low threshold (4 cells); (d) deep/tap (11 cells); and (e) non-responsive (31 cells). Eight cells were evaluated for responses to different doses (5-150 micrograms) of intravenous (i.v.) serotonin. Two of the 8 cells exhibited excitatory effects, whereas 2 cells classified as deep/tap and 4 cells classified as non-responsive were not affected. The results of this study have shown the upper cervical component of the spinomesencephalic tract is made up of a heterogenous population of cells involved in the integration of varied inputs from large portions of the body. It is proposed that the large population of SMT cells in the upper cervical spinal cord may be involved in pain mechanisms, especially those related to the affective consequences of acute and chronic pain.

The cells of origin and terminal areas of the feline spinomesencephalic tract were investigated by the intraaxonal transport method. Following injection of wheat germ agglutinin-horseradish peroxidase conjugate into the cervical and lumbar enlargements, anterograde labelling was observed in several regions of the dorsal midbrain. The main terminal areas were the periaqueductal gray matter, the intercollicular nucleus, the posterior pretectal nucleus and the nucleus of Darkschewitsch. In addition, there was a sparse projection to the cuneiform nucleus and the anterior pretectal nucleus. The superior colliculus was virtually devoid of labelling except for a weak termination in the caudal part of the deep layers. Although there was a considerable overlap, the projection from the cervical spinal cord to the periaqueductal gray matter terminated more rostrally than that from the lumbar segments, indicating the presence of a somatotopic organization. The retrograde labelling seen after tracer injection into the midbrain terminal areas revealed that the cells of origin were located mainly in the upper cervical segments and in the cervical and lumbar enlargements; in the latter parts of the cord an overwhelming majority were situated in lamina I, with smaller fractions in laminae IV and V, whereas in the upper cervical segments and in the less densely labelled thoracic and sacral segments a much larger proportion of the peroxidase-positive neurons were found in the deep laminae. About 75% of the labelled cells were located contralateral to the injection site. The functional implication of the present results are discussed in relation to somatosensory activity in the mesencephalon. It is suggested that several regions of the dorsal midbrain have an important somesthetic function including that of pain.

Spinothalamic tract cells in the lumbar, sacral and caudal segments of the primate spinal cord were labelled by the retrograde transport of horseradish peroxidase (HRP) injected into the thalamus. The laminar distribution of stained spinothalamic cells in the lumbosacral enlargement differed according to whether the HRP was injected into the lateral or the medial thalamus. Lateral injections labelled cells in most laminae, but the largest numbers of cells were in laminae I and V. The highest concentrations of cells labelled from the medial thalamus were in laminae VI-VIII. Ninety percent or more of the stained spinothalamic cells in the lumbosacral enlargement were contralateral to the injection site. In the conus medullaris stained spinothalamic cells were most numerous in laminae I, V and VI following lateral thalamic injections of HRP. Many of the cells of the conus were in Stilling's nucleus. Twenty-three percent of the cells in the conus were ipsilateral to the injection site in the lateral thalamus. Only a few cells in the conus were labelled by medial thalamic injections. The total number of spinothalamic cells from L5 caudally was estimated to be at least 1,200-2,500. An injection of HRP into the midbrain resulted in a laminar distribution of labelled cells much like that produced by a lateral thalamic injection. The types of spinothalamic tract cells and the sizes of their somata were determined for different laminae. The cells types resemble those already described from Golgi and other studies of the spinal cord gray matter. The spinothalamic tract cells in lamina I included Waldeyer cells and numerous small fusiform, pyriform or triangular cells. Those in lamina II included limitrophe and central cells. Spinothalamic cells in lamina III were central cells. Most of the labelled cells in laminae IV-X were polygonal, although there were also flattened cells in these layers. The smallest spinothalamic cells were in laminae I-III, while the largest were in laminae V and VII-IX. Spinothalamic cells in the conus medullaris included cells like those in the lumbosacral enlargement, but also a special cell type in Stilling's nucleus. Some cells in the conus had dendrites that crossed the midline. Spinothalamic axons could sometimes be traced to the ventral white commissure within one or a few sections. In longitudinal sections, most labelled axons were in the ventral part of the lateral funiculus on the side of the injection, although a few were in the ventral funiculus or on the contralateral side. The axons were widely dispersed, and a few were located adjacent to the pia-glial membrane.

We have examined the physiological and morphological characteristics of spinal dorsal horn lamina I neurons with projections to the midbrain in the cat by combining physiological recording of neurons with the intracellular injection of HRP. Lamina I spinomesencephalic neurons were antidromically activated from the region that included the cuneiform nucleus and lateral periaqueductal gray at the intercollicular level. The majority of mesencephalic projection neurons (50 of 55) responded exclusively to noxious stimulation (nociceptive-specific) of their peripheral receptive fields. Lamina I spinomesencephalic neurons were activated from both the ipsilateral and contralateral midbrain and had slow antidromic conduction velocities (1 to 18 m/second). We identified eight cells with projections to both the midbrain and the thalamus and eight cells that were antidromically activated only from the thalamus. Intracellular injection of HRP revealed that lamina I spinomesencephalic neurons were of diverse morphological types, but generally had extensive, rostrocaudally oriented, dendritic arbors confined to lamina I and the overlying white matter. Axons were observed on nine of the HRP-filled spinomesencephalic neurons; five of the axons issued collateral branches. The morphological characteristics of these neurons did not appear to correlate with functional categories (i.e., wide-dynamic-range- or nociceptive-specific-type neurons). The large number of nociceptive-specific neurons with projections to the midbrain and the interconnections of these midbrain sites with hypothalamic and limbic structures suggest that the lamina I spinomesencephalic pathway plays an important role in the autonomic and affective responses to pain.

Anterograde transport of wheat germ agglutinin conjugated to horseradish peroxidase was used to determine the terminal domain of the projection from the lumbosacral spinal cord to the midbrain in the rat, cat, and monkey. Results have shown that several midbrain regions receiving afferent input from this level of the spinal cord are common to the three species examined. Structures innervated by this projection were located throughout the full rostrocaudal extent of the midbrain. The strongest projections were to the intercollicular region and caudal midbrain contralateral to injection sites in the spinal cord. Terminal labeling in the rostral midbrain, except that observed in the nucleus of Darkschewitsch, was substantially less than that observed at more caudal midbrain levels. Structures receiving the strongest input from the spinal cord included the central gray, nucleus cuneiformis, the deep and intermediate layers of the superior colliculus, and the intercollicular nucleus. Other structures receiving afferent input from the lumbosacral spinal cord included the anterior and posterior pretectal nuclei, red nucleus, Edinger-Westphal nucleus, interstitial nucleus of Cajal, and the mesencephalic reticular formation. It is concluded that the spinal projection to the midbrain is a multicomponent projection consisting of several pathways terminating in discrete midbrain regions. Considering the diverse functions associated with midbrain regions receiving spinal input and the response and receptive field properties of cells belonging to this pathway, the results of the present study are discussed in relation to the potential role of the spinomesencephalic tract in somatic, visceral, and motor function.

The method of retrograde axonal transport of horseradish peroxidase (HRP) was used to identify the locations of cells of origin of the spinothalamic tract in the cat. Injections of from 0.2–3.0 μl of 30% HRP were made unilaterally in various regions of the somatosensory thalamus. Massive injections of the caudal thalamus in several cats showed the spinothalamic cells of origin to be located mainly in laminae I, VII and VIII in the lumbar enlargement, and in laminae I, V and VII–VIII in the cervical enlargement. Small injections of HRP were made into the three major spinothalamic terminal zones in the thalamus, to determine the laminar origin(s) of the spinal projections to each zone. Neurons in lamina I in both cervical and lumbar enlargements were found to project almost exclusively to the rostral VB-caudal VL border zone. A small number of neurons in laminae VII and VIII also project there but a larger number project to the intralaminar region. Neurons projecting to the PO regions were located mainly in laminae IV and V. This anatomical segregation of thalamic afferents probably reflects a functional segregation of input, since the functional properties of spinal neurons vary according to their laminar location. Comparison of these data with the differential projection spinothalamic neurons in the rat and monkey indicate that it is unlikely that the proposed “paleo-” and “neospinothalamic” systems would arise from anatomically separate groups of spinal neurons.

The projections to the midbrain from the spinal cord have been investigated in the cat with the degeneration technique and by using horseradish peroxidase (HRP) as an anterograde tracer. Two types of spinal cord lesions were performed: 1) Cordotomies at cervical or thoracic levels transecting the ventral and lateral funiculi. 2) Transections of the ventral, ventrolateral, dorsolateral or dorsal funiculus, respectively, at cervical levels. In the anterograde tracing experiments HRP was injected into the spinal cord at cervical, lumbar or sacral levels. The results show large projections to the lateral and ventrolateral parts of the periaqueductal gray (PAG1), the posterior pretectal nucleus (PP) and the nucleus of Darkschewitsch (D). More moderate projections go to the medial division of the periaqueductal gray (PAGm), the cuneiform nucleus (CF), the mesencephalic reticular formation (MRF), lateral part of the deep layer of the superior colliculus (SP) and magnocellular medial geniculate nucleus (GMmc), while scattered spinal fibers are present in the dorsal part of the periaqueductal gray (PAGd), the external inferior collicular nucleus (IX), the intermediate layer of the superior colliculus (SI), the lateral part of the red nucleus (NR) and in the Edinger-Westphal portion of the oculomotor nucleus (3). In addition a few fibers are present in the interstitial nucleus of Cajal (CA) and anterior pretectal nucleus (PAc). The results indicate that at midcervical levels most of the spinomesencephalic fibers ascend in the ventral funiculus, with only a moderate fraction ascending in the ventral half of the lateral funiculus. Almost no fibers ascend in the dorso-lateral funiculus and none appear to pass in the dorsal funiculus. No distinct somatotopic pattern was found in the spinomesencephalic projections, but more fibers from cervicobrachial segments terminate in the rostral than in the caudal parts of the terminal fields in PAG, CF, SP and IX, while the lumbar fibers were more numberous in the caudal parts. PP seems to receive spinal fibers mainly from the caudal half of the cord.

Spinomesencephalic tract neurons in the cat spinal cord were retrogradely labeled following injections of wheatgerm agglutinin conjugated to horseradish peroxidase into the region of the midbrain parabrachial area. Labeled cell bodies were concentrated in lamina I, bilaterally. A more scattered distribution was observed in lamina V and deeper laminae. Bilateral lesions of the dorsolateral funiculus (DLF) at thoracic levels eliminated labeling of lamina I neurons below the lesions, but had no effect on the labeling of the neurons in deeper laminae. Injections of colchicine into the spinal white matter caused the label to accumulate intra-axonally and revealed labeled axons bilaterally in the DLF and ipsilaterally in the ventrolateral and ventral funiculi.

An anatomical technique based on the retrograde transport of horseradish peroxidase (HRP) was used to investigate the projections of spinal cord neurons to the reticular formations in the rat. Both large and restricted injections were staggered all along the bulbar and pontine levels, involving the nucleus gigantocellularis, the nuclei reticularis pontis, pars oralis and caudalis and in some cases the nucleus raphe magnus. Labeled cells were constantly encountered in the reticular part of the neck of the dorsal horn throughout the whole length of the cord, mainly contralateral to the central core of the injection site. This area was taken as the equivalent of lamina V in the cat. Other labeled cells were observed in the medial parts of the intermediate and ventral horns, in areas considered similar to laminae VII and VIII in the cat. The two most rostral cervical segments were characterized by an additional bilateral projection originating from the dorsolateral part of ventral horns. Thus, this study is a clear confirmation that the bulbopontine reticular formations constitute a target for various somatosensory inputs originating in spinal cord. It demonstrates that the medial spinoreticular tract (mSRT) differs from the other main ascending tracts by the absence of projections from (1) superficial layers and nucleus of the dorsolateral funiculus contrary to the spinomesencephalic tract; (2) ventromedial zone of the lumbar dorsal horn unlike the spinothalamic tract; (3) the neck of the dorsal horn in its medial portion contrary to the spinoreticular component reaching the lateral reticular nucleus; and (4) central cervical nucleus and Clarke's columns, unlike the spinocerebellar tracts. The difficulty in demonstrating retrograde labeling from discrete injections could result from the fact that mSRT neurons have sparsely ramified collaterals on their terminal zones.

Recordings were made from 90 identified spinomesencephalic tract (SMT) cells in the lumbosacral spinal cord of cats anesthetized with alpha-chloralose and pentobarbital sodium. Recording sites were located in laminae I-VIII. Antidromic stimulation sites were located in different regions of the rostral and caudal midbrain including the periaqueductal gray, midbrain reticular formation, and the deep layers of the superior colliculus. Twelve SMT cells were antidromically activated from more than one midbrain level or from sites in the medial thalamus. The mean conduction velocity for the population of cells sampled was 45.2 +/- 21.4 m/s. Cells were categorized based on their responses to graded intensities of mechanical stimuli and the location of excitatory and/or inhibitory receptive fields. Four major categories of cells were encountered: wide dynamic range (WDR); high threshold (HT); deep/tap; and nonresponsive. WDR and HT cells had excitatory and/or inhibitory receptive fields restricted to the ipsilateral hindlimb or extending to other parts of the body including the tail, forelimbs, and face. Some cells had long afterdischarges following noxious stimulation, whereas others had high rates of background activity that was depressed by nonnoxious and noxious stimuli. Deep/tap cells received convergent input from muscle, joint, or visceral primary afferent fibers. The placement of mechanical lesions at different rostrocaudal levels of the cervical spinal cord provided information related to the spinal trajectory of SMT axons. Six axons were located contralateral to the recording electrode in the ventrolateral/medial or lateral funiculi while two were located in the ventrolateral funiculus of the ipsilateral spinal cord. Stimulation at sites used to antidromically activate SMT cells resulted in the inhibition of background and evoked responses for 22 of 25 cells tested. Inhibitory effects were observed on responses evoked by low/high intensity cutaneous stimuli and by the activation of joint or muscle primary afferent fibers. Based on the response and receptive-field properties of SMT cells it is suggested that the SMT may have an important role in somatosensory mechanisms, particularly those related to nociception.

Anatomical and electrophysiological methods were used to investigate the projections and response properties of neurons in the second cervical (C2) spinal segment of the cat giving origin to a previously undescribed projection to the ipsilateral thalamus. The method of retrograde axonal transport of horseradish peroxidase (HRP) was used to identify neurons in C2 giving rise to thalamic projections. Following large (3.0 microliter) thalamic HRP injections, a large number of labeled neurons was observed in lateral laminae VII-VIII of C2 ipsilateral to the injections. They occurred as small clusters of cells along the longitudinal axis of C2. Labeled neurons were also observed contralaterally in the lateral cervical nucleus, dorsal horn (especially medial lamina VI), and loosely distributed in the ventral horn. The ipsilaterally projecting neurons were also labeled following small (0.2--0.5 microliter) HRP injections restricted to individual spinothalamic terminal zones (intralaminar nuclei, ventrobasal complex-nucleus ventralis lateralis border zone, medial division of the posterior nuclei), indicating that as a group they project widely throughout the thalamus. Single unit recording methods were used to obtain complementary information on the functional properties of these neurons. The antidromic stimulation method was applied to identify units in C2 projecting to the ipsilateral thalamus in anesthetized, paralyzed cats. Three categories of ipsilaterally projecting C2 units were identified: (1) units not driven by any type of natural stimulation; (2) units having large cutaneous receptive fields (RFs) and wide dynamic response ranges ("widefield"), and (3) units with smaller RFs and varied properties ("other"). Widefield units with bilaterally symmetrical and asymmetrical RFs were observed. Co-stimulation of different portions of an excitatory RF produced summation of the unit response. Inhibitory RF components were identified in one-third of the widefield units. Unit recordings after spinal tract lesions revealed that the afferent input passed via the ipsilateral lateral and/or ventral funiculi. Widefield unit responses to somatosensory stimuli could be inhibited by dorsal column conditioning stimulation. Several "other" units resembled widefield units, while a second group had small RFs restricted to the C2 dermatome. Possible functional roles of the projecting C2 neurons in somatosensory and non-specific systems are discussed.

HIGH-THRESHOLD MECHANORECEPTORS and their centrally projecting myelinated fibers make up a functionally distinct group of cutaneous sensory units that have been suggested as part of the afferent apparatus for pain resulting from mechanical damage to the skin (3, 17). The argument for their relation to pain was based on two points: 1) the ability of such afferent elements, unique among those with medullated fibers, to provide signals differentiating noxious from innocuous mechanical events affecting the skin; and 2) the well-established correlation between pain and activity in thin myelinated afferent fibers