Journal of Neurocytology 12, 299-316 (1983)

Thalamic inputs to identified commissural neurons in the monkey somatic sensory cortex

S. H . C. H E N D R Y a n d E. G . J O N E S

James L. O'Leary Division of Experimental Neurology and Neurological Surgery and McDonnell Center for the Study of Higher Brain Function, Washington Universi~ School of Medicine, 660 South Euclid Avenue, Saint Louis, Missouri 63110, USA

Received 9 July 1982; revised 22 September 1982; accepted 4 October 1982

S umm a r y

Commissurally projecting neurons were identified in the monkey first somatic sensory area (SI) by the retrograde axonaI transport of horseradish peroxidase (HRP) injected into the contralateral cortex. Neurons identified in this way have large pyramidal somata primarily in layer IIIB of the SI area. Their basal dendrites lie within the terminal plexus of thalamocortical afferents.

Electron microscopy was used to examine the synaptic relations of the labelled commissural cells, in particular to determine whether they receive monosynaptic thalamic connections. To do this, retrogradely labelled commissural cells and Golgi-impregnated large pyramidal neurons from layer IIIB were examined ultrastructurally in material in which thalamocortical terminals were degenerating due to a prior lesion of the thalamus. In a significant number of cases degenerating terminals were found to make synapses on the spines or shafts of labelled dendrites.

Injections of HRP into SI or into the white matter adjacent to the corpus callosum labelled callosal axons and terminals in the opposite SI. These axons terminated mainly near the somata of the layer IIlB pyramidal cells. Some of their terminals were found to synapse with dendrites receiving synaptic contacts from thalamocortical axon terminals.

Introduction

A major issue in the s tudy of cerebral cortical organization is the way in which afferent inputs, such as those of the thalamocortical and commissural systems, are relayed to the output , pyramidal cells. It is n o w clear that the long-held belief of a relay of thalamocortical input th rough a single popula t ion of granule or stellate cells in layer IV, the principal layer of thalamic terminations, is no longer tenable. First, it has been realized that several varieties of non-pyramida l cells exist, each with s tereotyped axon ramifications (Valverde, 1971; Szent~gothai, 1973; Jones, 19751 1981; Feldman & Peters, 1978; Gilbert & Wiesel, 1979) and having different synaptic relationships with other cortical cells (Somogyi, 1977; Peters & Fairen, 1978; Fair6n & Valverde, 1980; Somogyi & Cowey, 1981). Some of these cells also seem to use different transmitter

0300-4864/83 $03.00 + .12 �9 1983 Chapman and Hall Ltd.

300 HENDRY and JONES

agents (Hendry & Jones, 1981; Peters & Kimerer, 1981). Second, it appears that the somata or dendrites of any of these cells, provided they are within the layer of thalamic terminations, receive thalamocortical synapses (Peters & Feldman, 1976; Davis & Sterling, 1979; White, 1979). This implies that there are probably multiple circuits for the elaboration of sensory inputs in the cortex, not all of which, of course, need be in operation simultaneously.

It is also evident that even pyramidal cells, if they have dendrites within a zone of thalamic terminations, may receive monosynaptic thalamic inputs. The earlier indications of such inputs, based on suggestive light microscopic evidence (e.g. Globus & Scheibel, 1967; Valverde, 1967), and isolated electron microscopic observations (Jones & Powell, 1970; Strick & Sterling, 1974; Winfield & Powell, 1976) have now been confirmed for many types of pyramidal cells(Peters et al., 1977, 1979; White, 1978; Hersch & White, 1981) including several classes whose targets have also been identified, including the ipsilateral corticocortical (White & Hersch, 1981); the commissural cells (Hendry & Jones, 1:980; Hornung & Garey, 1980, 1981), and the corticothalamic cells (White & Hersch, 1982).

Although less extensively examined than the afferents of the thalamocortical system, commissural afferents have been shown to terminate, in part, in the same laminae and on similar neuronal elements as the thalamocortical afferents (Jones & Powell, 1970; Lund & Lund, 1970; Sloper & Powell, 1979). This raises the possibility that some neurons within the cortex may receive terminations from both types of afferent.

The morphological analysis of synaptic circuitry in the cerebral cortex has been greatly facilitated by the development of two methods that enable specific cells, identified light microscopically, to be examined electron microscopically. These are the method of labelling by the retrograde axoplasmic transport of horseradish peroxidase (LaVail & LaVail, 1974) and the Golgi-electron microscopic method of Fair6n et aI. (1977). When coupled with degenerative 'labelling' of thalamic afferents, the retrograde transport and Golgi-EM methods permit an identification of thalamocortical terminations on cells with a particular output connection and/or with a particular morphology. In the present study these methods have been used to examine the thalamic terminations on commissural neurons and to compare the termination of thalamocortical and commissural axons in the monkey first somatic sensory cortex. A preliminary account has appeared (Hendry & Jones, 1980).

Materials and methods

Eight cynomolgus monkeys (Macaca fascicularis) were used in this study. Under Nembutal anaesthesia, the left ventrobasal complex (VB) of the thalamus of each animal was identified by microelectrode recording of single unit responses to naturally applied peripheral stimuli. Multiple lesions were then made in VB by passing currents of 1-2 mA for 30 s through electrodes positioned at 12-15 previously identified sites. All recording and lesions were made in VB by

Tha lamic i n p u t to c o m m i s s u r a l n e u r o n s 301

introducing electrodes into the thalamus horizontally from behind, thus avoiding damage to the cortical area to be investigated.

After 2 or 3 days, a second operation was performed on four of the animals. Under Nembutal anaesthesia the right postcentral gyrus of each animal was exposed. Multiple injections of a 50% solution of horseradish peroxidase (HRP, Sigma, type VI) were made into the gyrus and into the white matter beneath it and adjacent to the corpus callosum through a 1/~1 Hamilton syringe. From 15 to 20, 0.5 to 1.0/xl injections were made in each animal. The HRP was delivered slowly to minimize damage to the cortex and dexamethasone was administered intravenously to help control cerebral oedema.

Following the second operation, the animals survived for an additional 2 days. They were then re-anaesthetized and perfused with a balanced saline solution followed by a fixative solution containing 1% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.2). After perfusion, the heads were removed and placed in fixative at 4 ~ C overnight.

The left first somatic sensory cortex (SI) of this group of animals was processed for HRP histochemistry in one of two ways. 1. The postcentral gyrus was sliced into parasagittally oriented slabs, 5-10 mm thick. Small blocks of tissue were taken at several points from each slab and were placed in 7% sucrose in phosphate buffer. The slabs were then frozen sectioned at 50 microns and both the sections and the blocks were reacted in 3,3'-diaminobenzidine tetrahydrochloride and hydrogen peroxide using the procedure of LaVail et aI. (1973). The blocks to be examined electron microscopically were chosen by reference to the localization of labelled neurons in the frozen sections. 2. The postcentral gyrus was sectioned parasagittally at 100/~m with a Vibratome (Oxford Instruments). The sections were collected, processed with a cobalt-intensified diaminobenzidine procedure (Adams, 1977) and examined under the light microscope. Small blocks of SI cortex containing groups of heavily labelled cells were cut out with the aid of a dissecting microscope and prepared for electron microscopy.

An additional four animals that were not injected with HRP were allowed to survive for 4 or 5 days following the placement of the thalamic lesion. They were then re-anaesthetized with Nembutal and perfused with a solution of 4% paraformaldehyde and 1% glutaratdehyde on 0.1 M phosphate buffer. Blocks of cortex containing the pre-central and post-central gyri were then stained with the Rapid Golgi method of Valverde (1970) and sectioned at a thickness of 100/xm on a Vibratome. Small wedges of tissue that included large pyramidal cells in layer IIIB of areas 3b, 1 and 2 were cut out of the sections and were processed with the method of Fairen et al. (1977) for electron microscopic examination of Golgi-impregnated neurons.

All tissue processed for electron microscopy was postfixed for one hour in 2% osmium tetroxide in phosphate buffer and dehydrated through ascending concentrations of ethanol to 70% ethanol. At this stage, the tissue was stained en bloc with 1% uranyl acetate in 70% ethanol for one hour. Following the staining, dehydration was completed and tissue was embedded in Spurr 's resin. Blocks were trimmed for electron microscopy after light microscopic inspection of 1/~m thick sections stained with methylene blue and Azure I1 (Richardson et al., 1960). Serial sections for electron microscopy were collected on Formvar-coated grids. The sections were stained on the grid either with lead citrate (Reynolds, 1963) alone or with uranyl acetate (Watson, 1958) and lead citrate. All sections were examined with a Zeiss EM-9S electron microscope. Serial reconstructions were made by photographically enlarging the electron micrographs and tracing a labelled element and all synapses it made or received onto paper. Only those terminals with recognizable membrane thickenings were accepted as synapsing on a labelled cell.

In a preliminary communication (Hendry & Jones, 1980), we reported on our examination of five HRP-labelled commissural neurons. The present paper is based upon 28 additional HRP labelled neurons and on four Golgi-impregnated neurons.

302 HENDRY and JONES

Results

Light microscopic appearance of commissural neurons Neurons giving rise to commissural axons in the monkey SI are large pyramidal cells (Fig. 1), m a n y of which have somata located in the deeper, large-celled part of layer 1II (layer IIIB). The somata of commissural cells are organized into radial columns (Fig. 1), the largest somata of which are present in the deepest part of layer lllB. Among them can be seen the labelled terminal ramifications of commissural axons (Figs. 1, 3).

Terminations of thalamocortical axons on labelled commissural neurons

Ultrastructurally, the pyramidal shape of the HRP-labelled commissural neurons is apparent (Fig. 2). The osmiophilic HRP reaction product is present in tubular and in lysosome-like structures such as membrane -bound and multivesicular bodies (Fig. 2); these are numerous in the somata, in the axon hillocks and axon initial segments and in the primary dendrites. While the majority of the labelled material is in the somata or larger dendrites of these cells, some small, probably distal dendritic branches and even dendritic spines are also found to be labelled.

Most degenerat ing thalamocortical terminals in the monkey SI cortex synapse with dendritic spines, and a small number contact dendritic shafts. In all cases synapses formed by thalamocortical axon terminals have asymmetric membrane thickenings. At the survival times used, all the degenerat ing terminals are very electron dense and are sur rounded by reactive astroglial processes (Figs. 4, 5 and 12). A proportion of the degenera t ing thalamocortical axon terminals are found to synapse w i th HRP-labelled commissural cell dendrites. The dendrites receiving such contacts, when traced through serial sections to their parent somata, are clearly the basal dendri tes of layer IIIB pyra-

midal cells (Figs. 5, 7). Degenerating thalamocortical terminals are present on the primary dendrites of some labelled cells and on more distal branches of other labelled cells. Some labelled dendrites receiving degenerat ing thalamocortical synapses o n their shafts could not be traced to their parent somata even through extensive series of sectioffs. Most of these dendri tes are likely to be distal braches of basal dendri tes (Fig. 4).

Fig. 1. Light micrograph of neurons in monkey somatic sensory cortex (SI) labelled by the retro- grade transport of HRP from the contralateral SI. Note the clustering of labelled pyramidal cells into two vertical columns. The largest cells (arrowheads) are in the deepest part of layer IIIB. Scale bar: 250/~m. Fig. 2. Electron micrograph of a pyramidal cell body in layer IIIB labelled by the retrograde transport of HRP from the contralateral cortex. A small number of axosomatic synapses (arrow- heads) display exclusively symmetric thickenings. A myelinated, degenerating thalamocortical axon (arrow) is present near the labelled cell body. Scale bar: 10 ~m. Inset: electron micrograph of HRP-labelled organelles in the celt body of a commissural neuron. The labelled structures include lysosome-like structures (arrowheads) and tubular organelles (arrow). Scale bar: 1 ~m.

Fig. 3. Camera lucida drawing of an HRP-filled commissural axon. The axon gives rise to numer- ous swellings in layers IIlB and IV and fewer in layers IliA, II and I. Scale bar: 100/~m.

304 HENDRY and JONES

Figures 8 a - d show a degenera t ing thalamocort ical axon terminal that was found to synapse on the dendrit ic spine o f an HRP-label led commissura l cell dendri te . The spine

arises, by w a y of a shor t , relatively thick stalk, f rom a labelled dendr i te (Fig. 8a) that

was identified as part of the basal dendrit ic sys tem of a labelled layer llIB pyramida l cell

(Fig. 7). The n u m b e r of degenera t ing thalamocort ical terminals o b s e r v e d to terminate o n the

shafts and spines of HRP-label led dendr i tes is ra ther small. The vast major i ty of basal dendr i tes of identified commissu ra l neurons , fo l lowed th rough lengthy series of sec-

tions, receive no synapses f rom degenera t ing axon terminals .

Fig. 4. Electron micrograph of an HRP-labelled commissural cell dendrite (labelled organelle indicated by curved arrow). A degenerating thalamocortical terminal forms an asymmetric synapse with the dendrite (straight arrow). Normal axon terminals synapse with the dendritic shaft and a dendritic spine (arrowheads). Scale bar: 3 ~m.

Fig. 5. (a and b) Electron micrographs of two sections from a series showing a basal dendrite of an HRP-labelled commissural neuron synapsing with a degenerating thalamocortical axon terminal (curved arrows). Labelled organelles are indicated by arrowheads. Scale bars: 5 ~m. Fig. 5c. A higher magnification electron micrograph of the dendrite in Fig. 5b showing the asymmetric synapse (arrows) between the commissural cell dendrite and a degenerating thalamocortical axon terminal. The curved arrow indicates an HRP-labelled organelle. Scale bar: 1 ~m.

Fig. 6. An electron micrograph of an HRP-labelled commissural cell dendrite. A terminal arising from a myelinated axon (arrow) forms a symmetric synapse with the dendrite. A labelled organelle is indicated by the curved arrow. Scale bar: 5 ~m.

Fig. 7. A reconstruction of serial electron micrographs of the HRP-labelled commissural neuron shown in Figs. 8 and 9. The arrow indicates the synapse between the degenerating thalamocorti- cal terminal and a spine of a secondary basal dendritic branch. Solid patches along the dendrite indicate asymmetric synapses. Stippling indicates symmetric synapses. Scale bar: 5 ~m.

Fig. 8. (a-d) Electron micrographs of four sections in a series showing a dendritic spine of a commissural neuron receiving a degenerating thalamocortical terminal (arrows), see Fig. 9. Scale bar: 2/~m.

Fig. 9. Low magnification electron micrograph of the HRP-labelled commissural cell dendrite (labelled organelle is indicated by curved arrow) that gave rise to the spine shown in Fig. 8a-d (large arrow). Asymmetric synapses on the shaft of the dendrite are indicated by arrowheads. The figure is rotated with respect to Fig~ 8a-d. Scale bar: 4 ~m.

Fig. 10. (a and b) Electron micrograph of two sections from a series showing basal dendrite of a Golgi-impregnated pyramidal cell of layer IIIB (see Fig. 11). A dendritic spine (arrow) receives an asymmetric synapse. Scale bar: 7/~m.

Fig. 11. Camera lucida drawing of the GolgMmpregnated, gold-toned layer IIIB pyramidal cell. Scale bar: 100 ~m.

Fig. 12. Electron micrograph of a synapse between a degenerating thalamocortical terminal and a dendritic spine (arrow) of the pyramidal cell in Fig. 11. The spine was found to arise from a thin dendritic branch (arrowhead). Scale bar: 4 ~m.

308 HENDRY and JONES

Terminations of degenerating thalamocortical terminals on Golgi-impregnated layer IIIB pyramidal cells

Examination of large layer IIIB pyramidal cells (Fig. 11) with the Golgi-EM method reveals the structure of even the smallest processes of these putatively commissurally projecting neurons. In particular, the distal dendritic branches and dendritic spines - processes difficult to connect with their parent dendrites with standard serial recon- structions - can be examined and the axon terminals synapsing with them identified. The vast majority of axon terminals ending on the dendritic shafts and spines of layer IIIB pyramidal cells are normal (Figs. 10a and b). Only a very small number of terminals, giving rise to axospinous contacts on these cells, are found to be degenerating and can thus be identified as arising from thalamocortical axons (Fig. 12).

Normal synapses on commissural neurons The dendrites of HRP-labelled commissural neurons and Golgi-impregnated layer IIIB pyramidal neurons receive many synapses from normal axon terminals of undetermined origin (Figs. 5, 6 and 9). Both symmetric synapses formed by terminals with flattened vesicles and asymmetric synapses formed by terminals with spherical vesicles are pres- ent on the shafts and spines of the dendrites. Terminals forming symmetric synapses are most common on dendritic branches close to the soma while those forming asymmetric synapses are present o n more distal branches and on dendritic spines. Although the

origin of all the axon terminals that end on commissural cell dendrites is not known, some terminals with flattened vesicles forming symmetric synaptic contacts are found to arise from myelinated axons (Fig. 6). Many symmetric synapses also occur on the initial segments of commissural cell axons.

The morphology of commissural axons Injections of HRP directly into the white matter of the postcentral gyrus result in the labelling of many axons in the contralateral areas 3b, 1 and 2. Labelled axons are found to ascend radially from the white matter to layer IV without branching (Figs. 3 and 14). In layer IV this type of axon divides several times, giving rise to thinner ascending branches. In layer IV and in the supragranular layers these branches give off a large number of even thinner branches, most of which ascend for a distance of 100/~m or more. However, some are either horizontal in orientation or ascend at an oblique angle. At irregular intervals along their lengths, each of these labelled axon branches gives rise

Fig. 13. Electron micrograph of a large aspiny dendrite in layer IV receiving asymmetric synapses from a large HRP-labelled commissural axon terminal (double arrows) and a degenerating thalamocortical axon terminal (single arrow). Scale bar: 2 ~m. Inset: electron micrograph of a densely labelled commissural axon terminal synapsing with a dendritic spine. Scale bar: 2 ~m. Fig. 14. A light micrograph of an HRP-labelled commissural axon. Numerous labelled swellings that are likely to be axon terminals are present in layers IV and IlIB. Scale bar: 90 ~m.

310 HENDRY and JONES

to swellings 1 to 3 #m in diameter (Fig. 14). From their size and distribution these swellings are likely to be the synaptic terminals of the labelled axons.

Most of the thinnest branches of these axons, especially those which are horizontally oriented, are in layers IIIB and IV, where they form a relatively dense and wide (200-300 #m) plexus. Ascending from this plexus, several thin branches and often one or two secondary or tertiary :branches of the main axon enter layer IIIB and the deep part of layer IliA. At this level tEey give rise to a second narrower and less dense plexus. A few labelled branches ascend from this second plexus to reach layer II and the lower part

of layer I. The lack of terminal branches from the labelled axons in layer VI, their predominantly

radial orientation and their terminations in the supragranular layers, correlates well with the pattern of anterograde labelling of the commissural projection previously demonstrated (Jones et al., 1975, 1979).

Termination of HRP-labelled commissural axons Labelled axon terminals are present among the somata and dendrites of retrogradely labelled neurons. The labelling in these terminals is either made up of one or more HRP-filled vesicular organelles (Fig. 13), similar to the anterograde labelling reported in other systems (LaVail & LaVail, 1974; Nauta et al., 1975; Winfield et al., 1975), or it is of the 'solid filling' variety (Fig. 13, inset) suggestive of transport in damaged axons. The absence of solid filling of retrogradely labelled commissural neurons and the very small number of HRP-filled organelles in the myelinated segments of these cells' axons, makes it unlikely that their intracortical axon collaterals contribute to the population of labelled terminals. We tentatively conclude, then, that the labelled terminals arise from commissural afferent axons. All labelled terminals form asymmetric synapses, most of which are on dendritic spines (Fig 13, inset) though a small number are on dendritic shafts.

Figure 13 shows a large unlabelled dendrite in layer IIIB which is contacted by several normal terminals forming asymmetric synapses as well as by a large HRP-labelled terminal and by a degenerating thalamocortical terminal. The labelled terminal shown in Fig. 13 was followed through an extensive series of sections and was found to arise from a large diameter axon, which forms an en passant synapse with the dendrite and, at some distance, with a dendritic spine. The dendrite also receives an asymmetric synapse from a second HRP-labelled terminal.

Discussion

The methods used in this study to identify thalamocortical axon terminals and commissural neurons are, for the most part, subject to few interpretative difficulties. Lesion-induced degeneration has been used for more than fifteen years to identify the axon terminals of cortical afferent systems (Colonnier, 1964) and has proven very successful in 'labelling' the terminals of thalamocortical axons (for review see White,

Thalamic input to commissural neurons 311

1979). The only afferents to the SI cortex destroyed by the lesions were those originating in the thalamus. It could be argued that in those animals that received injections of HRP into the contralateral SI, incidental damage could result in degeneration of commissural axon terminals which would confuse the interpretation. However, a survival of 2 days following the injections should result in little observable terminal degeneration or in only early phases of degeneration in any affected commissural axon terminals (Sloper & Powell, 1978). All degenerating axon terminals contacting identified commissural cell processes are at a late phase of degeneration, consistent with the survival of 4 or 5 days used following the thalamic lesions and inconsistent with any commissural degeneration.

The identity of the HRP-labelled neurons as commissural cells seems assured in view of the restriction of the injection to the contralateral postcentral gyrus. The organization of the labelled cells into columns, as seen in sections cut with the freezing microtome or a Vibratome, and their laminar distribution is consistent with that reported for the majority of commissural neurons of the monkey SI in previous studies (Jones et al., 1975, 1979; Jones & Wise, 1977).

The interpretation of Golgi-stained neurons as commissural neurons is indirect. The somata of commissural neurons in the monkey SI are of various sizes (Jones & Wise, 1977) and are present in laminae IliA and IIIB (Jones et aI., 1979). They resemble in size and in position the neurons giving rise to ipsilateral corticocortical axons (Jones & Wise, 1977; Jones et al., 1979). However, the pyramidal cells with the largest somata, present in the deepest part of layer IIIB, appear to project only to the contralateral hemisphere (Jones & Wise, 1977). It is possible, then, to assume provisionally that the large Golgi-stained pyramidal neurons of deep layer IIIB examined in this study are commissural neurons.

Degenerating thalamocortical axon terminals were found to synapse with the shafts or spines of commissural cell dendrites. Basal dendritic shafts of pyramidal cells have been reported previously to receive synapses from identified thalamocortical terminals in the cat visual cortex (Davis & Sterling, 1979). The small number of synapses between these terminals and the shafts of commissural cell dendrites was expected, since more than 90% of the thalamocortical terminals in the monkey SI synapse with dendritic spines. Most of the remainder contact dendrites of non-spiny, non-pyramidal neurons (Sloper & Powell, 1979). The finding that very few spines of HRP-labelled commissural cell dendrites receive thalamocortical axon terminations could be taken to indicate an extremely limited number of synapses between thalamic afferents and commissural neurons. It is possible, however, that a greater number are present on the spines of more distal dendritic branches which the HRP did not reach and which we could not reconstruct back to a labelled soma, or on spines with long, thin stalks which are difficult to trace to their parent dendrites (Davis & Sterling, 1979; White, 1979). The result is that the spines examined in the HRP-labelled material are mainly a restricted population - those connected by short stalks to relatively proximal branches of basal dendrites. The majority of synapses formed by degenerating thalamocortical terminals

312 HENDRY and JONES

on dendritic spines of the Golgi-stained, presumed commissural neurons are on spines connected by long, thin shafts to distal dendritic branches. Even here, however, the number of degenerating terminals was rather small when compared to the number of normal terminals on the cells.

The small number and the apparent location of the majority of thalamocortical

synapses on the distal branches of basal dendrites of commissural neurons may imply that the monosynaptic thalamic input to commissural neurons is not very powerful. However, commissural neurons of SI (Innocenti et al., 1974) and of the primary visual cortex (Hubel & Wiesel, 1967) of cats have receptive fields of the size and type expected of neurons driven strongly and directly by thalamic input. If the same synaptic relations exist for commissural cells in cat cortex as those found in SI monkeys, then the number and positions of the thalamocortical synapses, especially those on the shafts of primary dendrites, may be sufficient to drive effectively commissural neurons. There are also, of course, many polysynaptic intracortical routes whereby thalamic input may come to influence the commissural neurons via the terminations of thalamic fibres on the several varieties of cortical interneurons (Jones, 1981).

Monosynaptic thalamocortical connections with commissural neurons have also been demonstrated in area 17 of the cat visual cortex (Hornung & Garey, 1:980) and neurons in mouse SI projecting to the ipsilateral motor cortex (White & Hersch, 1981) and thalamus (White & Hersch, 1982) have been shown to receive thalamocortical contacts. Physiological evidence of monosynaptic thalamic input to corticotectal (Palmer & Rosenquist, 1974) and corticothalamic (Harvey, 1978; Bullier & Henry, 1979) neurons of cat area 17 has also been reported. In the cortical fields of SI in monkeys, thalamocor- tical axons terminate in layer IIIB, in areas 1 and 2 and, in layers IIIB and IV and in area 3. There is an additional zone of terminations in all the fields at the border of layers V and VI (Jones, 1981). This laminar pattern of thalamic terminations is such that either the apical or the basal dendrites of virtually all cortical efferent neurons in SI are apparently in a position to receive direct thalamic input, even though their somata are situated in different layers (Hendry & Jones, 1983).

Most synaptic terminals forming asymmetric contacts on commissural cell dendrites and dendritic spines show no signs of degeneration and almost certainly arise from sources otber than thalamocortical axons. These sources include axons of intrinsic corti- cal neurons, axon collaterals of pyramidal neurons, and axons of other afferent systems. The terminations of commissural axons exactly overlap the distribution of the majority of commissural neurons in SI of monkeys (Jones et al., 1979) and are in an obvious position to synapse with these neurons. However, positive evidence of synaptic con- tacts between commissural neurons and commissural axon terminals or between com- missural neurons and axons of any identified source other than thalamocortical has not yet been provided.

Though callosal and thalamocortical axon terminals were not observed synapsing with the same labelled callosal neurons, we did find callosal and thalamic terminations on the same unlabelled dendrites. The structure and synaptology of these dendrites

Thalamic input to commissural neurons 313

were similar to those repor ted for the large cortical basket cells (Sloper et al., 1979), an

apparent ly GABAergic in terneuron (Hendry & Jones, 1981). This synaptic a r rangement

is consis tent with reports of convergence of thalamic and callosal inputs onto cortical inhibi tory in terneurons in the cat visual cortex (Toyama & Matsunami, 1976).

In the visual cortex, commissural connect ions provide the basis for the extension of receptive fields near the vertical meridian into the ipsilateral half visual field

(Choudhury et aI., 1965; Berlucchiet al., 1967; Hubel & Wiesel, 1967; Berlucchi, 1972).

But in o ther areas the function of commissural neurons has not been firmly established. Therefore the function of direct thalamic inputs to commissural neurons is unclear. The necessity for a direct thalamic input to commissural neurons in the shaping of their

receptive fields in the visual cortex seems evident and there would seem to be some value in the future in assessing the somatic sensory cortex from a similar viewpoint .

Acknowledgements

We gratefully acknowledge the technical assistance of Mr Ronald Steiner and the sec- retarial assistance of Ms Margo Gross. The s tudy was suppor ted by Grant n u m b e r

NS10526 from the National Institutes of Health, United States Public Health Service.

References

ADAMS, J. C. (1977) Technical considerations in the use of horseradish peroxidase as a neuronal marker. Neuroscience 2, 141-7.

BERLUCCHI, G. (1972) Anatomical and physiological aspects of visual functions of the corpus callosum. Brain Research 37, 371-92.

BERLUCCHI , G. , G A Z Z I N I G A , M. S. & R I Z Z O L A T T I , G. (1967) Microelectrode analysis of transfer of visual information by the corpus callosum. Archives Italiennes de Biologie 105, 583-98.

BULL1ER, J. & HENRY, G. H. (1979) Laminar distribution of first order neurons and afferent terminals in cat striate cortex. Journal of Neurophysiology 42, 1271-81.

C H O U D H U R Y , B. P., W H I T T E R 1 D G E , D. & W I L S O N , M. E. (1965) The function of the callosal connections of the visual cortex. Quarterly Journal of Experimental Physiology 50, 214-9.

c o L O N N IE R, M. (1964) Experimental degeneration in the cerebral cortex. Journal of Anatomy 98, 47-53.

DAVIS, T, L. & STERLING, P. (1979) Microcircuitry of cat visual cortex: classification of neurons in layer IV of area 17, and identification of the patterns of lateral geniculate input. Journal of Comparative Neurology 188, 599-628.

FAIRI~N, A. , PETERS, A. & S A L D A N H A , J. (1977) A new procedure for examining Golgi-impregnated neurons by light and electron microscopy. Journal of Neurocytology 6, 311-37.

FAIREN, A. & VALVERDE, F. (1980) A specialized type of neuron in the visual cortex of the cat: a Golgi and electron microscopic study of chandelier cells. Journal of Comparative Neurology 194, 761-80.

FELD MAN, M. L. & PETERS, A. (1978) The forms of non-pyramidal neurons in the visual cortex of the rat. Journal of Comparative Neurology 179, 761-94.

314 H E N D R Y a n d JO N ES

GILBERT, C. D. & WIESEL, T. N. (1979) Morphology and intracortical projections of functionally characterised neurones in the cat visual cortex. Nature 280, 120-5.

GLOBUS, A. & SCHEIBEL, A. B. (1967) Synaptic loci on visual cortical neurons of the rabbit: the specific afferent radiation. Experimental Neurology 18, 116-31.

HARVEY, A. R. (1978) Characteristics of corticothalamic neurons in area 17 of the cat. Neuroscience Letters 7, 177-81.

HENDRY, S. H. C. & JONES, E. G. (1980) Electron microscopic demonstration of thalamic axon terminations on identified commissural neurons in monkey somatic sensory cortex. Brain Research 196, 253-7.

HENDRY, S. H. C. & JONES, E. G. (1981) Sizes and distributions of intrinsic neurons incorporating tritiated GABA in monkey sensory-motor cortex. Journal of Neuroscience 1, 390-408.

HENDRY, S. H. C. & JONES, E. G. (1983) The organization of pyramidal and non-pyramidal cell dendrites in relation to thalamic afferent terminations in the monkey somatic sensory cortex. Journal of Neuro~tology 12, 278-98.

HERSCH, S. M. & WHITE, E. L. (1981) Thalamocortical synapses involving identified neurons in mouse primary somatosensory cortex: a terminal degeneration and Golgi/EM study. Journal of Comparative Neurology 195, 253-63.

H O R N U N G, J~ P. & G A R E Y, L. J. (1980) A direct pathway from thalamus to visual callosal neurons in cat. Experimental Brain Research 38, 121-3.

HORNUNG, l- P~ & GAREY, L. J. (1981) The thalamic projection to cat visual cortex: ultrastructure of neurons identified by Golgi impregnation or retrograde horseradish peroxidase transport.

Neuroscience 6, 1053-68. HUBEL, D. H. & WIESEL, T. N. (1967) Cortical and callosal connections concerned with the vertical

meridian of visual fields in the cat. Journal of Neurophysiology 30, 1561-73. INNOCENTI, G. M., MANZON1, J. & SP1DALIERI, G. (1974)Patterns of the somesthetic messages

transferred through the corpus callosum. Experimental Brain Research 19, 447-66. JONES, E. G. (1975)Varieties and distribution of non-pyramidal cells in the somatic sensory

cortex of the squirrel monkey. Journal of Comparative Neurology 160, 205-68. JONES, E. G. (1981) Anatomy of cerebral cortex: columnar inpu t -ou tpu t organization. In The

Cerebral Cortex (edited by SCHMITT, F. O., WORDEN, F. G., ADELMAN, G. and DENNIS, M.), pp. 199-235. Cambridge: MIT Press.

JONES, E. G., BURTON, H. & PORTER, R. (1975)Commissural and cortico-cortical 'columns' in the somatic sensory cortex of primates. Science 190, 572-4.

JONES, E. G., COULTER, J. D. & WISE, S. P. (1979) Commissural columns in the sensory-motor cortex of monkeys. Journal of Comparative Neurology 188, 113-6.

JONES, E. G. & POWELL, T. P. S. (1970) An electron microscopic study of the laminar pattern and mode of termination of afferent fibre pathways in the somatic sensory cortex of the cat. Philosophical Transactions of the Royal Society of London Series B 257, 45-62.

JONES, E. G. & WISE, S. P. (1977) Size, laminar and columnar distribution of efferent cells in the sensory-motor of monkeys. Journal of Comparative Neurology 175, 391-438.

LAVAIL, J. H. & LAVAIL, M. M. (1974) The retrograde intra-axonal transport of horseradish peroxidase in the chick visual system: a light and electron microscopic study. Journal of Comparative Neurology 157, 303-57.

LAVAIL, J. H., WINSTON, K. R. & TISH, A. (1973) A method based on retrograde intra-axonal transport of protein for identification of cell bodies of origin of axons terminating within the CNS. Brain Research 58, 470-7.

LUND, S. & LUND, R. D. (1970) The termination of callosal fibers in the paravisual cortex of the rat. Brain Research 17, 25-45.

Thalarnic input to commissural neurons 315

NAUTA, H. J. W., KAISERMAN-ABRAMOF, I. & LASEK, R. J. (1975) Electron microscopic observat ions of horseradish peroxidase t ranspor ted from the cauda toputamen to the substantia nigra in the rat: possible involvement of the agranular reticulum. Brain Research 85, 373-84.

PALMER, L. A. & ROSENQUIST, A. C. (1974) Visual receptive fields of single striate cortical units projecting to the superior colliculus in the cat. Brain Research 67, 27-42.

PETERS, A. & FAIREN, A. (1978) Smooth and sparse ly-spined stellate cells in the visual cortex of the rat: a s tudy using combined Golgi -e lec t ron microscopic technique. Journal of Comparative Neurology 181, 129-72.

PETERS, A. & FELDMA N, M. L. (1976) The projection of the lateral geniculate nucleus to area 17 of the rat cerebral cortex. I. General description. Journal of Neurocytology 5, 63-84.

PETERS, A. & KIMERER, M. (1.981) Bipolar neurons in rat visual cortex: a combined Golgi -e lec t ron microscope study. Journal of Neurocytology 10, 921-46,

PETERS, A., PROSKAUER, C. C., FELDMAN, M. L. & KIMERER, L. (1979) The projection of the lateral geniculate nucleus to area 17 of the rat cerebral cortex. V. Degenerat ing axon terminals synapsing with Golgi - impregnated neurons. Journal of Neurocytology 8, 331-57.

PETERS, A., WHITE, E. L. & FAIRI~N, A. (1977) Synapses be tween identified neuronal elements. An electron microscopic demonstra t ion of degenerat ing axon terminals synapsing with Golgi - impregnated neurons. Neuroscience Letters 6, 171-5.

RE Y N OLD S, E. S. (1963) The use of lead citrate at high pH as an electron-opaque stain in electron microscopy. Journal of Ceil Biology 17, 208-12.

RICHARDSON, K. D., J ARETT, L. & FINKE, E. H. (1960) Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technology 35, 313-23.

SLOPER, J. J. & POWELL, T. P. S. (1978) Observat ions on the processes of degenerat ion of the afferent connections to the sensory-motor cortex of the monkey. Neuroscience 3, 1031-44.

SLOPER, J. J. & POWELE, T. P. S. (1979) An experimental electron microscopic study of afferent connections to the primate motor and somatic sensory cortices. Philosophical Transactions of the Royal Society of London Series B 285, 199-226.

SLOPER, I. J., HIORNS, R. W. & POWELL, T. P. S. (1979) A qualitative and quantitative electron microscopic s tudy of the neurons in the primate motor and somatic sensory cortices. Philosophical Transactions of the Royal Society of London Series B 285, 141-71.

SOMOGYI, P, (1977) A specific axo-axonal interneuron in the visual cortex of the rat. Brain Research 136, 345-50.

SOMOGYI, P. & COWEY, A. (1981) Combined Golgi and electron microscopic s tudy on the synapses formed by double bouquet cells in the visual cortex of the cat and monkey. Journal of Comparative Neurology 195, 547-66.

STRICK, P. & STERLING, L. P. (1974) Synaptic termination of afferents from the ventrolateral nucleus of the thalamus in the cat motor cortex. A light and electron microscopic study. Brain Research 61, 218-41.

SZENTAGOTHAI, J. (1973) Synaptology of the visual cortex. In Handbook of Sensory Physiology, Vol. VII~3 Central Processing of Visual Information, Part B, Visual Centers of the Brain (edited by JUNG, R.), pp. 269-324, Berlin: Springer-Verlag.

T O Y A M A, K. & M A T S U N A M 1, K. (1976) Convergence of specific visual and commissural impulses upon inhibitory interneurons in cat 's visual cortex. Neuroscience 1, 107-12.

VALVERDE, F. (1967) Apical dendrit ic spines of the visual cortex and light deprivation in the mouse. Experimental Brain Research 3, 337-52.

VALVERDE, F. (1970) The Golgi method. A tool for comparative structural analyses. In Contemporary Research Methods in Neuroanatomy (edited by NAUTA, W. J. H. and EBBESO N, S. O. E.), pp. 11-31. New York: Springer-Verlag.

316 H E N D R Y a n d J O N E S

VALVERDE, F. (1971) Short axon neuronal subsys tems in the visual cortex of the monkey. International Journal of Neuroscience 1, 181-97.

WATSON, M. L. (1958) Staining of tissue sections for electron microscopy with heavy metals. Journal of Biophysical and Biochemical Cytology 4, 475-8.

WHITE, E. L. (1978) Identified neurons in mouse Sml cortex which are post-synaptic to thalamocortical axon terminals: a combined Golgi-e lect ron microscopic and degenerat ion study. Journal of Comparative Neurology 181, 627-62.

WHITE, E. L. (1979) Thalamocortical synaptic relations: a review with emphasis on the projection of specific thalamic nuclei to the primary sensory areas of the neocortex. Brain Research Reviews 1, 275-311.

WHITE, E. L. & HERSCH, S. M. (1981) Thalamocortical synapses of pyramidal cells which project from Sml to MsI cortex in the mouse. Journal of Comparative Neurology 198, 167-81.

w H I TE, E. L. & H E R S C H, S. M. (1982) A quantitative s tudy of thalamocortical and other synapses involving apical dendri tes of corticothalamic projection cells in mouse SmI cortex. Journal of Neurocytology 11, 137-57.

WINFIELD, D. A., GATTER, K. C. & POWELL, T. P. S. (1975) An electron microscopic study of retrograde and or thograde t ransport of horseradish peroxidase to the lateral geniculate nucleus of the monkey. Brain Research 92, 462-7.

WINFIELD, D. A. & POWELL, T. P. S. (1976) The termination of thalamocortical fibres in the visual cortex of the cat. Journal of Neurocytology 5, 269-81.