"Tactile" stimulus intensity: information transmission by relay neurons in different trigeminal nuclei.

Comparison of the information transmitted about the intensity of a steady "tactile" stimulus applied to facial skin by single trigemino-thalamic neurons in nucleus oralis and nucleus caudalis indicates that little information loss occurs at the medial lemniscal synaptic relay (nucleus oralis), but that it is gross within the nucleus caudalis.

Touching textured surfaces: cells in somatosensory cortex respond both to finger movement and to surface features.

Single neurons in Brodmann's areas 3b and 1 of the macaque postcentral gyrus discharge when the monkey rubs the contralateral finger pads across a textured surface. Both the finger movement and the spatial pattern of the surface determine this discharge in each cell. The spatial features of the surface are represented unambiguously only in the responses of populations of these neurons, and not in the responses of the constitutent cells.

Neural basis of the sense of flutter-vibration.

Comparison of human detection thresholds for oscillatory movement of the skin of the hand with response properties of first-order myelinated mechanoreceptive afferents from the monkey's hand, activated in an identical stimulus pattern, indicates that flutter-vibration is a dual form of mechanical sensibility, served peripherally by two different sets of fibers.

Manual dexterity and corticospinal connectivity following unilateral section of the cervical spinal cord in the macaque monkey.

The macaque recovers quite rapidly from the immediate severe flaccid hemiparesis that results from unilateral section of the cervical spinal cord (between C3 and C6) and starts to use the impaired hand to pick up objects within about 30 days following the surgery. Within another 60 days, the monkey is quite dexterous; nonetheless, there is a persisting deficit. We used video recording to study the long-term recovery of manual dexterity following unilateral section of the cervical cord in newborn and juvenile monkeys. A reach-and-retrieve manual task was examined. By using a preset oppositional force, opposition of the pads of the index finger and thumb in the vertical plane was needed to retrieve the desired target object. The corticospinal connectivity of each monkey was also examined by using retrograde or anterograde tracers at the end of the experimental period (Galea and Darian-Smith [1997] J. Comp. Neurol., this issue) and was correlated with the manual performance. Manually retrieving an object depends on the coordination of several control processes acting in parallel, including 1) visually guided components, such as directing the arm toward the object, aligning the digits with the target object by pronating the forearm, and preshaping the index/thumb separation to match with the size and shape of the target, and 2) manipulative components that depend on tactual input and that also include independent movements of the digits and the application of the appropriate oppositional forces. The impairment of manual dexterity that persisted after a cervical section, although it was small, involved these processes and was evident in 1) the less direct trajectory used in reaching, 2) the loss of preshaping of the separated index finger and thumb prior to grasping the target object, and 3) a weakening of the oppositional forces that could be developed between the pads of the index finger and thumb. Although, in the accompanying paper, we did not preclude some regeneration of severed corticospinal connections, we did show that, if any such reconstruction occurred, then it was limited. The remarkable but incomplete recovery of dexterity over a period of 6-12 months, therefore, must be achieved by 1) optimizing the transmission of information from the cortex to the spinal cord by the substantially reduced populations of corticospinal neurons and corticobulbospinal projections and/or 2) the effective use of spinal circuitry in regulating the more stereotyped elements of the manual task.

Corticospinal projection patterns following unilateral section of the cervical spinal cord in the newborn and juvenile macaque monkey.

Immediately following a unilateral section of the midcervical spinal cord that interrupts the dorsolateral, lateral, and ventral columns, the macaque monkey has a severe flaccid paralysis on the side of the lesion. Recovery of hand function is rapid, and, although it is incomplete, within a few months, the monkey uses the initially disabled hand and fingers with considerable skill. We examined the accompanying changes in the pattern of projection of corticospinal neurons to the cervical spinal cord that occurred following such a lesion. Spinal section was done both in newborn and juvenile macaques, and the postlesion period was followed for up to 150 weeks. Corticospinal neuron populations were visualized by using both anterogradely and retrogradely transported labels, and their origins, spinal pathways, and terminations were examined at intervals during the period of recovery of hand function. Immediately following unilateral section of the spinal cord at C3, sampled counts of soma profiles of retrogradely labeled neurons indicated that there was a profound reduction in the corticospinal projection to the hemicord caudal to the lesion. The few labeled corticospinal axons spared by the lesion bypassed the spinal lesion by descending in the contralateral cord and then crossing the midline caudal to the lesion. A few corticospinal axons may also have bypassed the lesion in the ipsilateral ventromedial column when this was not fully interrupted by the lesion. In every monkey, we observed a similar, profound reduction in the corticospinal (and rubrospinal) projections to the hemicord caudal to the lesion: This pattern did not alter significantly over an extended recovery period. An unchanging corticospinal projection to the cervical spinal cord contralateral to the lesion was also visualized in each monkey and resembled that seen in the normal macaque. Although the resolution of the labeling and counting procedures used precluded the identification of small increases in the numbers of corticospinal neurons projecting to the hemicord caudal to the lesion, we concluded that there was no substantial reconstruction of this projection over a recovery period of more than 2 years.

Thalamic projections to areas 3a, 3b, and 4 in the sensorimotor cortex of the mature and infant macaque monkey.

Area 3a in the macaque monkey, located in the fundus of the central sulcus, separates motor and somatosensory cortical areas 4 and 3b. The known connections of areas 4 and 3b differ substantially, as does the information which they receive, process, and transfer to other parts of the central nervous system. In this analysis the thalamic projections to each of these three cortical fields were examined and compared by using retrogradely transported fluorescent dyes (Fast Blue, Diamidino Yellow, Rhodamine and Green latex microspheres) as neuron labels. Coincident labeling of projections to 2-3 cortical sites in each monkey allowed the direct comparison of the soma distributions within the thalamic space of the different neuron populations projecting to areas 3a, 3b, and 4, as well as to boundary zones between these cortical fields. The soma distribution of thalamic neurons projecting to a small circumscribed zone (diameter = 0.5-1.0 mm) strictly within cortical area 3a (in region of hand representation) filled out a "territory" traversing the dorsal half of the cytoarchitectonically defined thalamic nucleus, VPLc (abbreviations as in Olszewski [1952] The Thalamus of the Macaca mulatta. Basel: Karger). This elongate, rather cylindrical, territory extended caudally into the anterior pulvinar nucleus, but not forward into VPLo. The rostrocaudal extent of the thalamic territory defining the soma distribution of neurons projecting to small zones of cortical area 3b was similar, but typically extended into the ventral part of VPLc, filling out a medially concavo-convex laminar space. Two such territories projecting to adjacent zones of areas 3a and 3b, respectively, overlapped and shared thalamic space, but not thalamic neurons. Contrasting with the 3a and 3b thalamic territories, the soma distribution of thalamic neurons projecting to a circumscribed zone in the nearby motor cortex (area 4) did not penetrate into VPLc, but instead filled out a mediolaterally flattened territory extending from rostral VLo, VLm, VPLo to caudal and dorsal VLc, LP, and Pul.o. These territories skirted around VPLc. All three cortical areas 4, 3a, and 3b) also received input from distinctive clusters of cells in the intralaminar Cn.Md. It is inferred that, in combination, the thalamic territories enveloping those neuron somas projecting to, say, the sensorimotor hand representation in areas 3a, 3b, and 4 (and also areas 1 and 2), which would be coactive during the execution of a manual task, constituted a lamellar space extending from VLo rostrally to Pul.o caudally.(ABSTRACT TRUNCATED AT 400 WORDS)

Macaque red nucleus: origins of spinal and olivary projections and terminations of cortical inputs.

The cerebellar, spinal, bulbar, and cortical connections of the mammalian red nucleus imply a motor role. However, what information the red nucleus receives, processes, and distributes is poorly understood, partly because the rubral microcircuitry, especially in primates, remains incompletely defined. Multiple retrogradely transported fluorescent tracers were injected into the spinal cord and inferior olive of the macaque to label rubrospinal and rubroolivary neuron populations, respectively. Anterograde dextran amines were used to label the terminals of corticorubral neurons. These data provided the topographic framework for examining the morphology of rubral neurons in the accompanying paper (Burman et al. [2000]). Soma profiles of rubrospinal and rubro-olivary neurons were respectively segregated in the magnocellular and parvocellular nuclei. A subpopulation of neurons (DL-spinal cells) with their somas immediately dorsolateral to the rostral magnocellular nucleus and its capsule, also projected to the spinal cord, as did clusters of neurons in the periaqueductal grey matter. Terminals of corticorubral axons originating from ipsilateral primary motor area 4 (the densest projection), the supplementary motor area, cingulate area 24, area 8, and posterior parietal area 5, were each mapped in the parvocellular red nucleus. Only area 4 projected to the magnocellular red nucleus, and this projection as small. DL-spinal neurons had no cortical input. The somatotopic organization of rubral connections was examined only in (a) the corticorubral input from motor area 4, and (b) the rubrospinal and DL-spinal projections. These connections and their somatotopic alignment, were mapped in a 3-dimensional reconstruction of the red nucleus.

Geometry of rubrospinal, rubroolivary, and local circuit neurons in the macaque red nucleus.

The primate red nucleus consists of three main neuron subpopulations, namely, rubrospinal neurons in the magnocellular nucleus, rubroolivary cells in the parvocellular nucleus, and local circuit neurons in both subnuclei: Each subpopulation has unique cerebellar and neocortical inputs. The structural framework for the interactions of these rubral subpopulations remains poorly defined and was the focus of this study in six macaques. Somata of rubrospinal neurons, dorsolateral-spinal (DL-spinal) neurons, as defined in the accompanying paper (Burman et al. [2000] J. Comp. Neurol., this issue), and rubroolivary neurons were labeled retrogradely first with Fast Blue injected either into the cervical spinal cord or the inferior olive. The soma/dendrite profiles of selected cells (53 rubrospinal, 19 DL-spinal, and 17 rubroolivary cells) were visualized by the intracellular injection of Lucifer Yellow/biocytin in fixed slices (400 microm thick) of midbrain. The descriptive statistics of the somata and the dendritic arborization of each rubral neuron type were established. Projection neuron subpopulations had similar but differentiable soma/dendrite profiles, with four to six slender, spine-bearing dendritic trees radiating out approximately 400 microm from the soma. Twelve presumed interneurons, all in the parvocellular nucleus, differed from projection neurons in that they had smaller somata and many slender, spine-bearing segments that constituted the multibranching dendrite profile that radiated out approximately 250 microm from the soma. A tentative model of the macaque rubral microcircuitry was developed, and its functional implications were explored. It incorporated 1) the known topography of the nucleus and its connections, 2) our data specifying the soma/dendrite morphology of the three main rubral neuron types, and 3) the ultrastructure reported by other laboratories of intrarubral synaptic connections.

Ipsilateral cortical projections to areas 3a, 3b, and 4 in the macaque monkey.

In the macaque monkey area 3a of the cerebral cortex separates area 4, a primary motor cortical field, from somatosensory area 3b, which has a subcortical input mainly from cutaneous mechanoreceptive neurons. That each of these cortical areas has a unique thalamic input was illustrated in the preceding paper. In the present experiments the cortical afferent projections to these 3 areas of the sensorimotor cortex monkey were visualized and compared, using 4 differentiable fluorescent dyes as axonal retrogradely transported labels. The cortical projection patterns to areas 3a, 3b, and 4 were similar in that they each consisted of (a) a "halo" of input from the immediately surrounding cortex, and (b) discrete projections from one or more remote cortical areas. However, the pattern of remote inputs from precentral, mesial, and posterior parietal cortex was different for each of the 3 cortical target areas. The cortical input configuration was least complex for area 3b, its remote input projecting mainly from insular cortex. The pattern of discrete cortical inputs to the motor area 4, however, was more complex, with projections from the cingulate motor area (24c/d), the supplementary motor area, postarcuate cortex, insular cortex, and postcentral areas 2/5. Area 3a, in addition to the proximal projections from the immediately surrounding cortex, also received input from the supplementary motor area, cingulate motor cortex, insular cortex, and areas 2/5. Thus, this pattern of cortical input to area 3a resembled more closely that of the adjacent motor rather than that of the somatosensory area 3b. Contrasting with this, however, the thalamic input to area 3a was largely from somatosensory VPLc (abbreviations from Olszewski [1952] The Thalamus of the Macaca mulatta. Basel: Karger) and not from VPLo (with input from cerebellum, and projecting to precentral motor areas).

Thalamic projections to sensorimotor cortex in the macaque monkey: use of multiple retrograde fluorescent tracers.

We used several fluorescent dyes (Fast Blue, Diamidino Yellow, Rhodamine Latex Microspheres, Evans Blue, and Fluoro-Gold) in each of eight macaques, to examine the patterns of thalamic input to the sensorimotor cortex of macaques 12 months or older. Inputs to different zones of motor, premotor, and postarcuate cortex, supplementary motor area, and areas 3b/1 and 2/5 in the postcentral cortex, were examined. Coincident labeling of thalamocortical neuron populations with different dyes (1) increased the precision with which their soma distributions could be related within thalamic space, and (2) enabled the detection by double labeling, of individual thalamic neurons that were common to the thalamic soma distributions projecting to separate, dye-injected cortical zones. Double-labeled thalamic neurons projecting to sensorimotor cortex were rarely seen in mature macaques, even when the injection sites were only 1-1.5 mm apart, implying that their terminal arborizations were quite restricted horizontally. By contrast, separate neuron populations in each thalamic nucleus with input to sensorimotor cortex projected to more than one cytoarchitecturally distinct cortical area. In ventral posterior lateral (oral) (VPLo), for example, separate populations of cells sent axons to precentral medial, and lateral area 4, medial premotor, and postarcuate cortex, as well as to supplementary motor area. Extensive convergence of thalamic input even to the smallest zones of dye uptake in the cortex (approximately 0.5 mm3) characterized the sensorimotor cortex. The complex forms of these projection territories were explored using 3-dimensional reconstructions from coronal maps. These projection territories, while highly ordered, were not contained by the cytoarchitectonic boundaries of individual thalamic nuclei. Their organization suggests that the integration of the diverse information from spinal cord, cerebellum, and basal ganglia that is needed in the execution of complex sensorimotor tasks begins in the thalamus.

Thalamic projections to sensorimotor cortex in the newborn macaque.

In the present experiments thalamocortical projections to different functional areas of the newborn (or prematurely delivered) macaque's sensorimotor cortex were labeled using retrogradely transported fluorescent dyes. Several dyes were used in each animal to (1) enable the direct comparison of the soma distributions of different thalamocortical projections within thalamic space, and (2) identify by double labeling neurons shared between these distributions. The projection patterns in the newborn macaque were compared with those of the mature animal reported by Darian-Smith et al. (J. Comp. Neurol. 1990;298:000-000). The main observations were (1) all thalamocortical projections to the sensorimotor cortex of the mature macaque are well established by embryonic days 146-150, as was shown by labeling these pathways in infants delivered by cesarean section, (2) a significant number of thalamocortical neurons in the newborn were double-labeled following dye injections into different pre- or postcentral areas, and where the margins of the dye uptake zones were separated by 3-8 mm, and (3) extensive projections from the anterior pulvinar nucleus to the motor and premotor cortex, and to the supplementary motor cortex were labeled in the newborn macaque. Both the exuberant terminal arborizations, and the precentral pulvinar projections were diminished by the 6th postnatal month, and absent in the mature macaque. The role of epigenetic determinants of these postnatal events is briefly considered.

Transcutaneous recording of single neuron activity in the cerebral cortex of the monkey.

A new method is described for recording responses of single neurons in the monkey's cerebral cortex by passing a microelectrode directly through the overlying skin and dura. Previous to this recording a window had been cut in the calvarium, and the bone deficit repaired using a full-thickness skin graft. Stable unitary recordings have been made over a period of one year following craniotomy, the only skull attachment being a small stainless steel peg, on which the microdrive was mounted when required.

A stimulator for moving textured surfaces sinusoidally across the skin.

The stimulator allows textured surfaces to be moved sinusoidally across the skin of the fingerpad. Sinusoidal motion is produced by a "scotch yolk" driven by a DC motor. The amplitude of movement is adjustable up to a maximum of 80 mm peak to peak and the frequency is continuously adjustable from 0.1 Hz to 2.0 Hz. Movement of the surface is monitored by an optical transducer and contact force between the finger and the surface is monitored by a strain gauge bridge. The stimulator is simple and robust and is suitable for both neurophysiological and psychophysical experiments in animals and humans.

Peripheral neural representation of the spatial frequency of a grating moving across the monkey's finger pad.

1. Responses in mechanoreceptive afferent fibres innervating the monkey's finger pads were examined when a ridged surface ("grating') was moved across the fibre's receptive field with a specified velocity and applied force. 2. The stimulus feature represented in single fibre responses was the temporal frequency of the moving grating (stimulus temporal frequency = velocity of moving surface/spatial period); information about the spatial period of the grating was represented equivocally. 3. Peripheral neural representation of the grating's spatial period (or spatial frequency) depended on information signalled by the responding fibre population rather than by individual fibres. 4. The three mechanoreceptive fibre populations responded differentially to a grating moving across the finger pad. Slowly adapting fibres coded best those stimulus combinations with a stimulus temporal frequency in the range 20-60 Hz, rapidly adapting fibres coded best those with frequencies of 60-200 Hz, and Pacinian fibres best defined those stimuli with a high temporal frequency (100-300 Hz). 5. Applying the moving grating to the skin with varying radial forces in the range 20-60 g wt. did not greatly modify the pattern of discharge in the responding fibre populations.

Innervation density of mechanoreceptive fibres supplying glabrous skin of the monkey's index finger.

1. The innervation densities of mechanoreceptive fibres supplying the ridged glabrous skin of the middle and terminal phalanges of the monkey's (Macaca nemestrina) index finger were estimated using a combination of histological and neurophysiological procedures. 2. This estimate was based on (a) a count of the total number of A beta myelinated fibres in the palmar digital nerve at the level of the proximal phalanx, (b) the demonstration that the majority of A beta fibres in the monkey's palmar digital nerve are mechanoreceptive afferents, (c) the estimation, based on a sample of 398 fibres, of the fractions of rapidly adapting, slowly adapting and Pacinian mechanoreceptive fibres in the palmar digital nerve, and (d) the estimation of the area of glabrous skin innervated by the palmar digital nerve. 3. The estimated innervation density of the finger pad and the skin of the middle phalanx were: rapidly adapting fibres, 178 and 80/cm2; slowly adapting fibres, 134 and 46/cm2; and Pacinian fibres, 13/cm2 for both phalanges.

Multiple corticospinal neuron populations in the macaque monkey are specified by their unique cortical origins, spinal terminations, and connections.

In primates, multiple corticospinal projections from the sensorimotor cortex operate in concert to regulate voluntary action. We examined the soma distributions of all those corticospinal neuron populations projecting to different zones in the cervical and more caudal spinal segments in the macaque that are labeled with retrogradely transported fluorescent tracers; 2-4 differentiable dyes were injected into different sites in the cervical spinal cord of each of 11 monkeys. Lamina V of the cerebral cortex, in which all corticospinal neuron somas were located, was unfolded with computer assistance to form a flat surface, and local soma densities were displayed on this plane as contour and 3-D maps. At least nine discrete, somatotopically organized corticospinal projections were identified. Three separate corticospinal projections originated in frontal cortex. The first projected mostly from area 4 (approximately 35% of the total contralateral neuron population), but also from the adjacent dorsolateral area 6a alpha (approximately 6% of total). The second large corticospinal projection (approximately 15% of total) originated in the supplementary motor area and a third small projection (approximately 2.6% of total) projected from the "postarcuate" cortex. Two separate corticospinal neuron populations were identified in areas 24 (approximately 6% of total) and 23 (approximately 4% of total) of the cingulate cortex. Thus, nearly 70% of the contralateral corticospinal projection originated in frontal and cingulate cortex. At the boundary between the primary motor and somatosensory cortex there was a sharp change in the pattern of projections. Only approximately 2.2% of the contralateral corticospinal projection originated in area 3a, rising to approximately 9% in areas 3b/1, and approximately 13% in areas 2/5. The projections from SII and insula totaled 3.4%. Ipsilateral and contralateral corticospinal projection patterns were similar, but the ipsilateral projection was only approximately 8.1% of that from the contralateral cortex. Each corticospinal neuron population had terminals in the intermediate zone of all spinal segments; additionally, there were ventral horn projections from the primary motor and cingulate cortex, and dorsal horn projections from the somatosensory cortex. Recognizing a number of separate populations of corticospinal neurons in the frontal, parietal, and insular cortex, each with unique thalamic and cortical inputs, and each of which has continuous access to all spinal motoneuron populations, underlines the importance of cortical and spinal connections linking them and coordinating their action. No coherent model of the cortical control of limb movements that incorporates this functional anatomy yet exists.

Postnatal maturation of the direct corticospinal projections in the macaque monkey.

Postnatal changes in the topography of the multiple corticospinal projections in the macaque monkey were followed using retrogradely transported fluorescent tracers, and related to the monkey's acquisition of manual dexterity; both behavioral and anatomical maturation were completed by about 8 postnatal months. Cortical origins of the corticospinal projections were examined by constructing planar projection maps of the distributions of labeled corticospinal neuron somas; these somas were found only in lamina V. At birth elaborate somatotopically organized corticospinal projections from primary motor cortex (area 4), the mesial supplementary motor area and cingulate areas 23 and 24, area 12, dorsolateral area 6a beta, the dorsolateral and ventral area 6a alpha (area F4), parietal areas 2/5, 7b and the peri-insular cortex (including area SII), were clearly defined, with axons extending to all spinal cord segments. While this pattern of regional projections broadly resembled that of the mature macaque, there were, however, substantial maturational changes during the 8 months after birth. These included (1) a halving of the area of cerebral cortex from which the contralateral corticospinal projection originated and (2) a threefold reduction in the number of labeled corticospinal neurons projecting to all segments of the cord. Collateral elimination rather than neuronal cell death was the likely mechanism for this reduction in the population and areal extent of corticospinal neurons in the maturing macaque. The surviving corticospinal axon terminals also developed substantially during the postnatal period. At birth some terminals had invaded the intermediate zone in each spinal segment, but few had penetrated the dorsal and ventral horns. By 6 postnatal months, however, many corticospinal neurons were retrogradely labeled following the injection of fluorescent labels into each of these spinal zones in cervical and lumbar spinal segments. These data demonstrate a considerable postnatal reduction in corticospinal neurons projecting to the contralateral spinal cord, and imply that many of the axons that are eliminated never synapse on spinal neurons. It is suggested that during the middle fetal period the axons of many of the cortical neurons in lamina V that in the mature monkey will terminate on particular neuron populations in the thalamus, brainstem, or spinal cord, traverse a common pathway down through the internal capsule into the spinal cord, passing close to these successive targets, and possibly forming collaterals at these levels. In the postnatal period each such neuron establishes a stable, effective synaptic input to only one or a few of these subcortical target populations, and the remaining collateral branches regress. The postnatal maturation of corticospinal neurons, examined in this study, is compatible with such a model.