PubMed Health⌕ Search

Biomedical subjects

H P Killackey

Publications and source records attributed to H P Killackey.

At least 37 records · Page 2Linked to original sources
PubMed1991

Early ingrowth of thalamocortical afferents to the neocortex of the prenatal rat.

The initial ingrowth of thalamocortical afferents into the presumptive somatosensory cortex was examined in the fetal rat. Thalamic fibers were labeled in fixed brains with the carbocyanine dye 1,1'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate (DiI). On embryonic day 16, thalamocortical afferents arrive in the neocortex and course tangentially within the intermediate zone immediately underneath the cortical plate. By embryonic day 17, thalamocortical fibers have begun their radial growth into cortex and their arbors span the cell-sparse zone between layer VIb and the bottom of the cortical plate. By the day of birth (embryonic day 21), thalamocortical fibers from a dense plexus within layers VI and V below the dense cortical plate. Our observations indicate that in the rat thalamic afferents arrive in the cortex at a very early age and arborize within the forming cortical layers without an apparent "waiting" period.

Afferent Pathways↗
PubMed1990

Areal and laminar organization of corticocortical projections in the rat somatosensory cortex.

The present study examines patterns of connectivity between the primary somatosensory cortex of the rat (SI) and surrounding cortical areas also implicated in the processing of somatosensory information. The impetus for the study was the recent reports of major differences in the organization of cortex lateral and caudal to the SI in two other rodent species; the mouse (Carvell and Simons, '86: Somatosens. Res. 3:213-237; '87: J. Comp. Neurol. 265:409-427) and the grey squirrel (Krubitzer et al., '86: J. Comp. Neurol 250: 403-430). Corticocortical connections between the somatosensory areas of the rat parietal cortex were examined by using the combined retrograde and anterograde transport of horseradish peroxidase as well as the retrograde transport of fluorescent tracers. Tracer injections were made into different locations within SI and dysgranular cortex as well as into more lateral regions of parietal cortex. The tangential patterns of distribution both of callosal connections and of cytochrome oxidase activity together provided points of reference in determining the relation between injection sites and the resultant patterns of label. The results indicate that two distinct somatosensory areas, SI and the dysgranular cortex, are interconnected with a further lateral somatosensory area referred to as the second somatosensory area (SII). These projections are organized in a topographic fashion, which we interpret as evidence for a single representation of the body surface in SII. The three somatosensory areas each exhibit unique laminar patterns of ipsilateral corticocortical projection neurons and terminations. In SI, projection neurons are found mainly in layers II, III, and Va, and terminations are largely restricted to the infragranular layers. In the dysgranular cortex, projection neurons and terminations are found in all layers except layer I in which only terminal label is detectable and layer Vb in which notably fewer neurons are labelled. In SII, projection neurons and terminations are found in all layers except layer I and are particularly dense in lower layer III and layer IV. Further, whereas the laminar and areal distributions of ipsilateral and contralateral corticocortical projections largely overlap in both SI and the dysgranular cortex, in SII they tend to be areally segregated. Neurons projecting bilaterally to both ipsilateral and contralateral somatosensory cortex were equally rare in all three somatosensory areas. These results are discussed in relation to the organization of SII in other rodent species, and it is concluded that in the rat, like the mouse, cortex lateral and caudal to SI contains a single representation of the body surface.

Animals↗
PubMed1990

Callosal projections in rat somatosensory cortex are altered by early removal of afferent input.

During the first postnatal week, the distribution of callosal projection neurons in the rat somatosensory cortex changes from a uniform to a discontinuous pattern. To determine if this change is influenced by afferent inputs to the somatosensory cortex, the effect of both early unilateral infraorbital nerve section and unilateral removal of the dorsal thalamus on the distribution of callosal projections in rat somatosensory cortex was examined. One month after either of the above manipulations at birth, the tangential distribution of callosal projections in the somatosensory cortex was examined using the combined retrograde and anterograde transport of horseradish peroxidase. Both manipulations alter the distribution of callosal projection neurons and terminations in the somatosensory cortex. After infraorbital nerve section, the distribution of callosal projections is altered in the contralateral primary somatosensory cortex. The abnormalities observed are consistent with the altered distribution of thalamocortical projections. In addition, consistent abnormalities were observed in the pattern of callosal projections of the second somatosensory area of both hemispheres. Most notably, they are absent in a portion of the region that contains the representation of the mystacial vibrissae and sinus hairs in this area. Thalamic ablation resulted in highly aberrant patterns of callosal projections in the somatosensory cortex on the operated side, where abnormal bands and clusters of callosal projections were observed in apparently random locations. These results are interpreted as evidence that both peripheral and central inputs influence the maturational changes in the distribution of callosal projection neurons.

Afferent Pathways↗
PubMed1989

Laminar and areal differences in the origin of the subcortical projection neurons of the rat somatosensory cortex.

Fluorescent retrograde tracing techniques were employed in a double-labelling paradigm to determine the distribution of corticospinal, corticotectal, and corticotrigeminal projection neurons in layer Vb of the adult and neonatal rat somatosensory cortex. The double-labelling paradigm allowed a direct comparison of the cortical distribution of neurons projecting to each target and identification of neurons projecting to more than one target. In the adult rat, each population of projection neurons was found to have a unique laminar and/or areal distribution. Corticospinal projection neurons were located throughout the width of layer Vb in the medial granular portion of somatosensory cortex, while corticotrigeminal projection neurons were distributed throughout the width of layer Vb in the more laterally located dysgranular portion of somatosensory cortex. Corticotectal projection neurons were located more superficially in layer Vb than either corticospinal or corticotrigeminal projection neurons and found scattered throughout both dysgranular and granular somatosensory cortex. Each combination of subcortical injections also resulted in double labelling a small percentage of uniquely distributed neurons. These distribution differences coupled with measurements of cell size allowed us to identify the parent population of the dual projection neurons. Subpopulations of corticotectal neurons also project to the brainstem trigeminal complex and to the spinal cord. Subpopulations of corticotrigeminal neurons also project to the spinal cord, and a proportion of corticotrigeminal neurons projects to at least two targets within the brainstem trigeminal complex (nucleus principalis and subnucleus interpolaris). In the adult rat, corticospinal neurons (as defined by either laminar position or somal size) did not appear to give off collaterals to either the superior colliculus or brainstem trigeminal complex. In the neonatal rat, double-labelled neurons which project to both the spinal cord and the tectum are distributed throughout the full width of layer Vb, rather than restricted to the superficial portion of the layer as in the adult rat. Further, it appears as if the ontogenetic change in the laminar distribution of corticospinal and tectal projection neurons is achieved by mechanisms of selective process elimination rather than cell death. These results are discussed in terms of both the developmental factor which may contribute to the discrete distribution of cortical projection neurons found in the adult and the functional significance of bifurcating projection neurons.

Animals↗
PubMed1989

Callosal projection neurons in area 17 of the fetal rhesus monkey.

We have studied the distribution of callosal projection neurons in area 17 of a fetal rhesus monkey which received large injections of horseradish peroxidase into the contralateral occipital cortex. In comparison to other cortical areas, area 17 contains few callosal projection neurons. Most of these cells are confined to a region extending tangentially about 2.5 mm from the 17/18 border, although a few neurons were noted as much as 5 mm from the border. Comparing the distribution of callosal projection neurons in the fetal monkey with what has been described in newborn and adult macaques, it is apparent that although some degree of refinement in striate callosal connections may occur during in utero development, the prenatal development of callosal connections in the macaque is inherently adult-like.

Animals↗
PubMed1989

Process elimination underlies ontogenetic change in the distribution of callosal projection neurons in the postcentral gyrus of the fetal rhesus monkey.

During fetal development, the regional distribution of callosal projection neurons in the rhesus monkey's postcentral gyrus changes from a uniform to a discontinuous pattern. To determine if this developmental change reflects the retraction of transient callosal projections, two different fluorescent tracers were injected into the brain of fetal monkeys of known gestational ages. Fast blue was injected into the entire postcentral gyrus of one hemisphere, whereas a second tracer (rhodamine latex beads or diamidino yellow) was injected into the caudal portion of the postcentral gyrus of the other hemisphere. The rostral portion of the postcentral gyrus (contralateral to the hemisphere injected with fast blue) was subsequently examined for the presence of labeled cells. In animals injected early in fetal development, on embryonic day 110 or younger and sacrificed 4 weeks later, there were numerous cells labeled with both tracers. In contrast, very few double-labeled cells were found in fetuses injected at an older age, embryonic day 135. We interpret these findings as showing that early in fetal development, when callosal projection neurons in the postcentral gyrus show a continuous distribution pattern, single cells in the rostral portion of this gyrus possess at least two collaterals, one projecting to the contralateral hemisphere and the other to the caudal portion of the gyrus. Subsequently, many of these neurons retract callosal collaterals while maintaining ipsilateral projections. Thus, process elimination accounts for the establishment of the discontinuous distribution of callosal neurons found in the postcentral gyrus of the mature primate.

Animals↗
PubMed1988

Temporal sequence of neurotransmitter expression by developing neurons of fetal monkey visual cortex.

The developing fetal monkey visual cortex was studied immunocytochemically from 110-155 days post-conception in order to localize cell populations immunoreactive (ir) for gamma-aminobutyric acid, Substance P, cholecystokinin-octapeptide, somatostatin, neuropeptide Y, and proenkephalin A peptide (BAM-18). The area 17/18 border and all cortical laminae identified in the adult visual cortex were discernible from the youngest age examined. All ir-cell populations studied were present at each fetal age. However, despite a relatively adult-like cytoarchitecture, all ir-cell populations studied displayed patterns of immunostaining which were unlike those described in adult visual cortex, and showed significant changes in laminar distribution, morphology, and numbers over the time course of gestation examined. Despite the differences in the patterns of immunostaining between the fetal and adult visual cortex, ir-cell populations intrinsic to the developing visual cortex exhibited adult-like combinations of co-localized transmitters and peptides. The developing monkey cortex also contains ir-cell populations, particularly BAM-18-ir cells, which have not been detected immunocytochemically in the adult monkey cortex. Differences between the fetal and the adult ir-cell populations might be accounted for by cell death, morphological transformation, secondary migration or changes in gene expression for neurotransmitters and neuropeptides.

Animals↗
PubMed1987

Choline acetyltransferase-immunoreactive neurons in fetal monkey cerebral cortex.

Two monoclonal antibodies to choline acetyltransferase (ChAT) were used to stain the cerebral cortex of fetal monkeys at 110-150 days post-conception. In addition to a small number of immunostained fibers, cells resembling typical non-pyramidal neurons were immunostained in developing layers V and VI and in the subjacent white matter of each area examined (sensory-motor and visual areas). ChAT-immunoreactive neurons have not been described in the cerebral cortex of adult primates, but the present observations indicate such neurons exist in the developing primate cortex.

Animals↗
PubMed1987

The organization and mutability of the forepaw and hindpaw representations in the somatosensory cortex of the neonatal rat.

The present study demonstrates that the primary somatosensory cortex of the rat contains a map of the entire body surface that is discernible with a routine anatomical staining technique, the succinic dehydrogenase reaction. The overall proportions of this map are relatively constant from rat to rat and very similar to those reported in previous physiological investigations (Welker: Brain Res. 26:259-275, '71, J. Comp. Neurol. 166:173-190, '76). We found 67% of the map to be related to the head of the rat, 15% to the forelimb, 14% to the trunk, and 4% to the hindlimb. Within the forelimb and hindlimb representations, there is a consistent internal organization that can be related to specific peripheral structures (digits or palm pads). Further, damage to either the periphery or the nerves innervating these regions on the day of birth produces disruptions in the normal pattern, but damage on day 6 or later does not. We interpret these results as indicating that the role of the periphery in organizing central neuronal structures during development previously demonstrated for the trigeminal system extends to the entire rat somatosensory system. Comparison of the present results with physiological studies of adult cortical maps after peripheral damage suggests to us that different substrates underlie the changes reported in the adult.

Age Factors↗
PubMed1987

Receptive-field properties of rat ventral posterior medial neurons before and after selective kainic acid lesions of the trigeminal brain stem complex.

Single neurons were recorded from the ventral posteromedial thalamic nucleus (VPM) of urethan-anesthetized rats. Six of these animals were intact, 28 sustained kainic acid (KA) lesions of trigeminal nucleus principalis (PrV), and 9 received similar lesions of trigeminal subnucleus interpolaris (SpVi). Four animals sustained PrV lesions that were followed, at an interval of 1-3 mo, by KA injections into SpVi. Special attention was paid to the receptive-field characteristics of neurons that were sensitive to deflection of the mystacial vibrissae. In normal animals, we recorded a total of 167 VPM neurons, 85% (n = 142) of which were vibrissa sensitive. The remaining VPM cells were excited by either guard hair deflection (8.4%), indentation of the skin (0.6%), or deflection of either vibrissae or guard hairs (1.8%). Seven cells (4.2%) were unresponsive. The topography of the trigeminal representation in VPM was similar to that reported previously by Waite (59). Vibrissa-sensitive neurons in intact rats generally gave rapidly adapting responses (84.5%), and only 16.2% were directionally selective. The vast majority (80.3%) of the vibrissa-sensitive cells were activated by deflection of only one whisker (1.2 +/- 0.5, mean +/- SD); none were excited by deflection of more than four vibrissae. Injections of KA into SpVi of otherwise intact rats (n = 9) had no appreciable effect on the receptive-field characteristics of vibrissa-sensitive VPM neurons. Injections of KA into PrV markedly altered the receptive-field properties of VPM cells. Recordings were made from 45 VPM neurons over a period extending from 0 to 10 h after KA injections into PrV in five rats. Of these cells, 4.4% were excited by vibrissa deflection and the remainder were unresponsive. Additional recordings from SpVi and the superior colliculus of these same animals indicated that the neurotoxin probably did not damage interpolaris neurons or their axons. Recordings were made from 394 VPM cells in 22 rats that survived 1-6 days after KA lesions of PrV. These experiments demonstrated an increase in the number of thalamic cells that were responsive to peripheral stimulation over this period. By 6 days after the lesion (4 animals), 52.8% of the 73 VPM neurons we recorded were excited by somatosensory stimuli. Of these, 89.5% were activated by deflection of one or more mystacial vibrissae. The average number of whiskers that excited a given VPM cell in these rats was 6.3 +/- 2.0 (SD). Recordings were made from VPM in five rats that survived 30-90 days after KA injections in PrV.(ABSTRACT TRUNCATED AT 400 WORDS)

Animals↗
PubMed1987

Terminal arbors of axons projecting to the somatosensory cortex of the adult rat. I. The normal morphology of specific thalamocortical afferents.

A characteristic feature of the rat somatosensory neocortex is a discrete topographic representation of the facial whiskers. Afferent fibers projecting to this vibrissae representation were "bulk-labeled" by injecting horseradish peroxidase into the white matter. Terminal arbors with the morphological characteristics of Lorente de No's (1949) "specific" thalamocortical afferents were then reconstructed through serial sections. These terminal arbors, characterized by the discrete organization of their dense plexus in layer IV, have a laminar distribution of boutons that parallels the laminar pattern of terminal degeneration resulting from lesions of the ventral posterior nucleus of the thalamus. The regional distribution of different-sized arbors corresponds to the distribution of vibrissae-related clusters of different sizes. Larger arbors were found in the posteromedial region corresponding to the mystacial vibrissae representation, while smaller arbors were found in the anterolateral region corresponding to the representation of the anterior sinus hairs. Terminal arbors were also reconstructed from sections stained simultaneously to demonstrate the pattern of vibrissae-related clusters. The greatest concentration of boutons on these axons occurred within a single vibrissae-related cluster. Furthermore, when 2 fibers terminated within a single cluster, their terminal arbors appeared to be largely coextensive. The morphology, size, and distribution of these terminal arbors support the hypothesis that the layer IV plexus of a single specific thalamocortical afferent tends to fill a vibrissae-related cluster. Thus, the organization of specific thalamocortical afferents may be responsible for clustered organization within the somatotopic map of the rodent neocortex.

Animals↗
PubMed1987

Terminal arbors of axons projecting to the somatosensory cortex of the adult rat. II. The altered morphology of thalamocortical afferents following neonatal infraorbital nerve cut.

The organization of the whisker representation within the neocortex of the rat is dependent on an intact periphery during development. To further investigate how alterations in this cortical map arise we examined the organization of thalamocortical afferents to the whisker representation in adult animals in which the infraorbital branch of the trigeminal nerve was cut on the day of birth. The disrupted pattern of thalamocortical projections to the vibrissae representation was apparent in the abnormal pattern of the anterograde transport of horseradish peroxidase from the thalamus, as well as in the abnormal pattern of succinate dehydrogenase activity. To determine the morphology of individual thalamocortical axons associated with this disrupted pattern, terminal arbors were "bulk-labeled" by injections of horseradish peroxidase into the white matter beneath the somatosensory cortex. Terminal arbors were identified by their laminar distribution of boutons corresponding to the specific thalamocortical afferent. The medial to lateral extent of these terminal arbors varied dramatically, from 350 to 1500 microns. In addition, terminal arbors innervating the same local area of cortex appeared to have varying degrees of overlap. Thus, the disruption of the neocortical vibrissae representation appears to involve the abnormal arborization of individual thalamocortical afferents. This finding supports the hypothesis that the fine-grain organization of the somatotopic map is dependent on the morphology and organization of individual thalamocortical arbors, which, in turn, are dependent on the periphery during development.

Animals↗
PubMed1986

Ontogenetic change in the distribution of callosal projection neurons in the postcentral gyrus of the fetal rhesus monkey.

In the postcentral gyrus of the mature rhesus monkey the distribution of callosal projection neurons is discontinuous. The density of callosal projection neurons, which are mainly located in the supragranular layers, varies both within and across cytoarchitectonic areas (Killackey et al., '83). In the present study, we investigated the ontogeny of corpus callosum projections of the postcentral gyrus in five fetal rhesus monkeys, ranging in age from embryonic day (E) 108 to E 133. Multiple large injections of horseradish peroxidase that involved the underlying white matter were made into the postcentral gyrus of one hemisphere and the distribution of labeled neurons in the ipsilateral thalamus and the other hemisphere was determined. The pattern of thalamic label indicated that the tracer was effectively transported from all portions of the postcentral gyrus. We found that the areal distribution pattern of labeled callosal projection neurons varied at the different fetal ages. At early fetal ages (E 108, E 111, and E 119) callosal projection neurons were continuously distributed throughout the postcentral gyrus. As in the adult animal, the vast majority of labeled callosal projection neurons were found in the supragranular layers, although a few labeled cells were located in the infragranular layers. From the earliest age, there was regional variation in the width of the band of labeled supragranular callosal projection neurons. The difference between the precentral and postcentral gyrus was most obvious, but there was also a difference between anterior and posterior portions of the postcentral gyrus. The first indication of some discontinuity in the distribution of callosal projection neurons was noted at E 126. By E 133, approximately 1 month before birth, the distribution of callosal projection neurons appeared remarkably mature. On E 119 aggregations of anterograde label could be detected in restricted portions of the posterior postcentral gyrus beneath the cortical layers. By E 133 anterograde label was found within the cortical layers (most densely in layer IV) in these regions of the postcentral gyrus. Thus, the emergence of the discrete pattern of callosal projection neurons appears to be temporally correlated with the ingrowth of callosal afferents. On the basis of these observations, as well as those of others (discussed in the text), we propose that the ontogenetic changes in the distribution of callosal projection neurons reflect the unique strategy employed by cortical projection neurons in establishing their patterns of connectivity. It is hypothesized that this strategy may involve multiple processes.

Animals↗
PubMed1985

Transganglionic transport of HRP from the circumvallate papilla of the rat.

To learn whether horseradish peroxidase (HRP) injections in gustatory papillae on the tongue can be used to study central topographical projections of taste buds and papillae, injections were made into the circumvallate papilla in rats. Labeled central projections after papilla injections were compared to projections after applying HRP to the cut glossopharyngeal nerve. Papilla injections result in HRP transport by afferent and efferent fibers of the glossopharyngeal nerve, and the pattern of central projections is similar to that after labeling the cut nerve. Projections include a separation in the brainstem of afferent, dorsally located fibers and efferent, ventrally located fibers. Afferent fibers project to the solitary nucleus and the trigeminal system. Efferent projections label muscle motorneurons in the nucleus ambiguus and the cells of origin of parasympathetic preganglionic fibers, which from the inferior salivatory nucleus. The parasympathetic neurons labeled after papilla projections are preganglionic fibers to Remak's ganglia in the tongue; post-ganglionic fibers of these ganglia are the secretomotor supply to the von Ebner's glands. In summary, injections of HRP into gustatory papillae reliably label central projections of the papilla and can be used for studies to discern topography in central projections of the taste system. Injections into the circumvallate papilla also have demonstrated that the parasympathetic neurons innervating von Ebner's glands are located in the inferior salivatory nucleus.

Animals↗
PubMed1985

The organization of the neonatal rat's brainstem trigeminal complex and its role in the formation of central trigeminal patterns.

The present study delimits the relationship of primary trigeminal afferents to their targets, the brainstem trigeminal nuclei of the neonatal rat. Previously, the brainstem trigeminal complex of the rat has been subdivided on the basis of either cytoarchitectonics or patterns of succinic dehydrogenase activity into the principal sensory nucleus and the three subnuclei of the spinal trigeminal nucleus, oralis, interpolaris, and caudalis. In this paper, we demonstrate that each of these subdivisions can also be identified by its pattern of primary trigeminal afferents. In addition, we demonstrate that the terminations of these afferents are distributed in a punctate fashion which correlates with vibrissae-related patterns of histochemical staining. Further, vibrissae removal in the neonatal rat at any age studied results in a corresponding deafferentation of both the principal sensory nucleus and all subnuclei of the spinal trigeminal nucleus. This same procedure has a graded, age-dependent effect on the vibrissae-related pattern of cytochrome oxidase staining in somatosensory cortex. On this basis, we conclude that vibrissae-related pattern formation in the central trigeminal system can be best understood in terms of a single "sensitive" period for the entire system. We hypothesize that this is the period during which an interaction normally occurs between primary trigeminal afferents and target neurons of the principal sensory nucleus.

Afferent Pathways↗
PubMed1985

Fine structural survey of the rat's brainstem sensory trigeminal complex.

The fine structural organization of the principal sensory trigeminal nucleus was compared with that of the spinal trigeminal nucleus (subnuclei oralis, interpolaris, and the deep layers of caudalis) in adult albino rats. Direct comparisons indicate similarities between all of the subdivisions of the brainstem trigeminal complex both in the major morphological classes of neurons present and in basic patterns of synaptic connections. Major differences between the several subdivisions occur in the relative numbers and distribution of the different cell types. The spinal trigeminal nucleus is distinguished by more numerous large (22-40 micron) polygonal neurons which give rise to long straight primary dendrites. Both the perikaryal surface and the thick primary dendrites of many of these cells are densely innervated by synaptic terminals. Especially large cells of this type are a prominent feature of subnucleus oralis. By contrast, the principal sensory nucleus is distinguished by its high density of small to medium-sized (8-20 micron) round or ovoid neurons. These smaller neurons tend to receive a sparse axosomatic innervation. In addition to these differences the spinal trigeminal neuropil is distinguished by the striking manner in which it is broken up by large rostrocaudally oriented bundles of myelinated axons. Proximal dendrites of polygonal and fusiform neurons often wrap around these large axon bundles. Morphologically heterogeneous populations of synaptic terminals with round vesicles (R terminals) and terminals with predominantly flattened vesicles (F terminals) occur in all of the subdivisions of the trigeminal complex. Both types of terminal make primarily axodendritic synapses, but both also make axosomatic synapses, and axospinous synapses with somatic as well as dendritic spines. In addition, axoaxonic synaptic contacts from F terminals onto large R terminals are seen in all subdivisions. Convincing examples of presynaptic dendrites were not observed in any of the brainstem subdivisions. Synaptic glomeruli, characteristic groupings of dendrites and synaptic terminals, are found throughout the brainstem trigeminal complex. The dendritic elements in these glomeruli tend to be small-diameter dendrites, spines, and large, spinelike appendages. Within the glomerulus these elements are postsynaptic to a single large R terminal and may also be postsynaptic to smaller F terminals. In addition, axoaxonic synaptic contacts from the F terminals onto the R terminal are a consistent feature of trigeminal synaptic glomeruli.(ABSTRACT TRUNCATED AT 400 WORDS)

Animals↗
PubMed1985

The role of the principal sensory nucleus in central trigeminal pattern formation.

Complete lesions of the principal sensory nucleus in the neonatal rat disrupts vibrissae-related pattern formation in the ventral posterior nucleus of the dorsal thalamus. Similar lesions of the spinal trigeminal nucleus do not effect pattern formation in the ventral posterior nucleus. The results are interpreted as suggesting that the principal sensory nucleus provides a template for pattern formation in central trigeminal structures.

Animals↗