THE JOURNAL OF COMPARATIVE NEUROLOGY 321:373-386 (1992)

Area 3a in the Cat. 11. Projections to the Motor Cortex and Their Relations to Other Corticocortical Connections CARLOS AVENDANO, ALBERTO J. ISLA, AND ESTRELLA RAUSELL Department of Morphology, School of Medicine, Autonoma University, 28029 Madrid, Spain

ABSTRACT It is well known that area 3a in the cat may monosynaptically influence the activity of neurons in the motor cortex. Much less information is available, however, on the anatomy of these connections. By using single or combined injections of different retrograde axonal tracers, we investigated the topography (horizontal and laminar) of area 3a neurons projecting to the motor cortex, and the anatomical relationships between these neurons and those projecting to other areas (2, 5, and SII) which, in turn, project to the motor cortex. Area 3a projects to all parts of area 4y, to area 46, and to the agranular area 6 in the lateral bank of the presylvian sulcus (urea 6ay), but not to other parts of areas 4 and 6. This projection exhibits a loose topographic organization along the mediolateral dimension of area 3a, and, in many cases, arises predominantly from the rostra1 half of this area. Although single small injections in the motor cortex produced two or more separate patches of retrograde labeling in 3a, after simultaneous injections of fluorochromes in two separate loci there often appeared in area 3a overlapping populations of neurons which were labeled retrogradely by each of the dyes, but with very few double-labeled neurons. In horseradish peroxidase (HRP) cases, 72% of area 3a neurons projecting to area 4y were distributed in supragranular layers (mainly layer 1111, although the proportion of labeling in infragranular layers was larger when using fluorescent dyes. Double-labeled cells predominated in infragranular layers. These results have a bearing upon the functional roles that have been attributed to area 3a, as a cortical locus involved in muscle sensation, and a cortical relay to the motor cortex of rapid feedback information from muscle activity during movement. o 1992 WiIey-Liss, Inc. Key words: sensory cortex, parietal cortex, transcortical reflex, double labeling

The role played by area 3a in mediating the transfer of sensory information from deep tissues to the motor cortex has been a subject of considerable attention since Thompson et al. ('70, '73) demonstrated that the stimulation of area 3a could excite or inhibit motor cortex neurons in the cat. More recent studies confirmed these findings and further demonstrated that a variety of cells in the motor cortex (including pyramidal tract and non-pyramidal tract neurons) were monosynaptically influenced by group I muscle afferent-driven cells in area 3a (Zarzecki et al., '78; Asanuma et al., '82; Herman et al., '85; Zarzecki, '89). The anatomical support for these physiological findings is, however, scarce: Grant et al. ('75) produced small lesions in area 3a in the focus of maximal amplitude of potentials evoked by stimulation of low-threshold afferents in the deep radial nerve, and observed with silver impregnation techniques discontinuous patches of degenerating terminals in area 4y in the posterior (GSP) and lateral sigmoid gyrus (GSL). More recently, Yumiya and Ghez ('841, using the retrograde transport of HRP, reported a topographically .

O

1992 WILEY-LISS, INC.

organized projection from area 3a to area 4y in the sigmoid gyrus, mostly or exclusively arising from supragranular layers. The existence of muscle afferent input to area 3a and the motor cortex in monkey is as well established as in cat (refs. in Jones and Porter, '80 and Hepp-Reymond, '88). However, the transcortical routes through which this input may gain access to the motor cortex are not clear. The first physiological studies which dealt with this issue offered contradictory suggestions as to the role played by area 3a in relaying group I muscle input to the motor cortex (Phillips et al., '71; Wiesendanger, '73). On the contrary, the earliest anatomical studies which used sensitive axoplasmic tracers seemed to rule out the existence of area 3a projections to motor cortex in the monkey (Jones et al., '78; Jones and Porter, '801, and this view prevailed for several years. More Accepted March 5,1992 Carlos Avendafio's present address is Department of Physiology, University of Montreal, C P 6128, Succ. A, Montreal, Quebec, Canada H3C 357.

c. AVENDANO ET AL.

374

recent observations, however, contradicted this view (Ghosh MATERIALS AND METHODS et al., '87; Huerta and Pons, '90; see Discussion). Twenty-three adult cats of both sexes were used for this An alternative route for group I muscle afferents to reach study. They received small, single or multiple injections of the motor cortex might be the cortical projections from area retrograde tracers in different zones of the motor cortex. 3a to various somatosensory areas which, in turn, project to Seven animals received simultaneous injections of the the motor cortex. Area 3a projects in a topographically fluorescent dyes Fast Blue (FB) and Diamidino Yellow organized fashion to areas 1and 2, SII, and area 5 , in both dihydrochloride (DY;both from Dr. Illing GmbH & Go. KG) cat and monkey (Grant et al., '75; Jones et al., '78; Burton in selected portions of the motor cortex, or in the motor and Kopf, '84; Yumiya and Ghez, '84; Pons and G a s , '86; cortex and other cortical areas. The remaining 16 animals Avendado et al., '88; Manzoni et al., '90). Projections to received single injections of horseradish peroxidase (HRP, areas 3b, SIV,and the granular insular cortex have also Sigma VI) in the motor cortex. Only those cases (15) in been described in cat (Grant et al., '75; Burton and Kopf, which retrograde labeling occurred in area 3a will be '84; Esteky and Schwark, '90; Amaya, '91). The distribu- described in the present report (Table 1;Fig. 1). tion to different cortical areas of the muscle afferent input HRP injections arriving in area 3a might subserve different functions, such as a rapid transcortical transfer to the motor cortex of Under Nembutal anesthesia, the animals were mounted information of muscle length during movement (Phillips, on a stereotaxic apparatus, the skull was opened, and the '69; Jones and Porter, '801, or the elaboration of conscious dura mater was removed over the sigmoid gyrus to expose kinesthetic percepts (McCloskey, '78). Little is known on the motor cortex. The injections were carried out under the latter, and the contribution of the somesthetic areas to direct visual approach with a 1pl Hamilton syringe, or, in a the function of the motor cortex is only beginning to be few cases when the target was hidden in the depth of the explored. Moreover, few data have been provided about the cruciate sulcus (SC), with a thin-tipped pipette coupled to a fine details of these projections. It is known, however, that manually operated system that was able to deliver brief air motor cortical projections from areas 3a and 2 differ, in that pulses. After a 2 day survival, the animals were deeply the latter end mostly upon supragranular neurons, and the anesthetized and sequentially perfused through the ascendformer drive monosynaptically neurons located in all layers ing aorta with cold saline, a mixture of aldehydes in of the motor cortex (Kosar et al., '85; Herman et al., '85; phosphate buffer, and the same buffer with 10% sucrose. Porter and Sakamoto, '88; Porter et al., '90). Also, it has The brains were then removed, blocked, photographed, and been shown that the SII neurons projecting monosynapti- cryoprotected with phosphate-buffered sucrose for an addically to the motor cortex are responsive to cutaneous tional 1-3 days. Two series of coronal, 50 pm thick sections stimulation (Mori et al.,'891, whereas neurons in areas 3a were obtained, at 200-400 pm intervals. They were treated and 2 respond to both superficial and deep somatic stimuli according to the tetramethylbenzidineprotocol of Mesulam ('78) for HRP; one was left without counterstaining, and (Asanuma et al., '82; Waters et al., '82). the other was counterstained with thionin. In order to advance our knowledge of the functional meaning of area 3a, a basic issue to be resolved is how this Fluorochrome injections area distributes its projections to the motor cortex and to The anesthesia and the surgery in these cases were other cortical regions which, in turn, project to the motor cortex, With this aim, in the present study we have used similar to those in HRP cases. FB and DY were suspended different retrograde tracers to analyze the organization of in distilled water or phosphate buffer a t 2-3%, and the the neurons in area 3a which project to the various suspension was sonicated right before use. In all animals subdivisions of the cat motor cortex, and to compare these the injections of each tracer were made with separate 1 pl projections with those directed to other cortical targets of Hamilton syringes in two different areas, as indicated in area 3a. Since the very definition of area 3a is controversial, Table 1and Figure 1. The animals were allowed to survive 7-9 days, and then we have previously carried out a study of this area with a number of morphological techniques, which is presented in were deeply anesthetized and perfused through the ascending aorta with 500 ml of saline, followed by 3,000 ml of 4% the accompanying paper (Avendado and Verdu, '92). paraformaldehyde in 0.1 M phosphate buffer; then the same buffer with 10% sucrose was added. The brains were removed, blocked, photographed, and kept overnight in the same buffer with 25% sucrose. Frozen sections 30 pm thick Abbreviations were cut in the coronal plane, and two series of one-in-five DY Diamidino Yellow dihydrochloride sections each were mounted, air-dried, and, in all but one FB Fast Blue case (GA 163, in which sections were left uncovered), coronal gyrus GCo anterior ectosylvian gyrus GEcSA quickly dehydrated, defatted, and coverslipped. Additional lateral gyms GL series were stained with Nissl, Gallyas, or acetylcholinestanterior sigmoid g y r u s GSA erase histochemistry techniques (see Avendafio and Verdu, lateral sigmoid gyrus GSL '92). posterior sigmoid gyrus GSP ~~

GSs

HRP PB SAn

sc sco SL SPS SSP

suprasylvian gyrus horseradish peroxidase phosphate buffer ansate sulcus cruciate sulcus coronal sulcus lateral sulcus presylvian sulcus splenial sulcus

Microscopic study and areal localization The counterstained sections in the HRP cases were projected onto paper with a Leitz-Prado Universal projector. The cortical surface and the borders between layers 1-11, and between layer VI and subcortical white matter were outlined. Blood vessels and tissue folds were also

CORTICAL AFFERENTS TO MOTOR CORTEX FROM AREA 3a

375

TABLE 1. Injection Data on 15 Cases _ _ _ _ _ ~ ~

Case'

Tracer solution

G 223 G 224 G216 G 266 G 356 G 385 G 405 G 415 G 528 GA 163

HRP 50% (DH2O) HRP 50% (DHlO) HRP 50%(DH20) HRP 50% (DHIO) HRP 50% (DH2O) HRP 50%(DH2O) HRP 50% (DH20) HRP 50% (DHzO) HRP 50% (DHzO) FB 2% (DH201 DY 2%(PB) FB 2% (DHiO) DY 2%(PB) FB 3% (DHzOj DY 2% (PB) FB 3% (DHgO) DY 2%(PB) FB 2% (DHAO) DY 2% (PB) FB 2% (DHzO) DY 2% (PB)

GA 175 GA 184 GA 186 GA 189 GA 191

No of injections 1 1

1 1 1 1 1 1 1 1 3

1 2 1 1 4 4 2

2 3 2

Total volume injected (nl) 60 60 40 60 70 60 70 60

80 200 600 200 300 100 100 400 400 300 300 450 400

Location of the injection(s1 areah-topography

Sumval time 51 h 48 h 44 h 44 h 46 h 45 h 44 h 48 h 46 h

Rd 8d 9d

4y (lateral part of GSP, bottom of SCI

4y (lateral part of GSP) 6ay tlateral bank of SPsl 47 + 6au (GSA and lower hank of SC) 4y (medial part of GSPI 4.1 + 6au (lower bank of SC) 46 + 3a (upper hank and dorudl lip ufscl 4y (intermediate part of GSP) 4y (GSL) 47 (GSL) 5a 5b (GSs) 4y (GSL and lateral part of GSA) 4y (GSP) 3a (medial part of GSP)

+

4y (medial part of GSP)

7d

5a

+ 5b (GSsand GLI

4y (GSL and lateral parts of GSP and GSAI

9d 8d

2 (GSPj 4y (GSL and lateral parts of GSP and GSAI

SII (GEcSAl 4y (GSL and lateral party of GSP and GSA)

'The HRP injections were carried out by Drs. A. Maran and F. Reinoso-Suarez,to whom we are gnteful for kindly providing us with this materlal

drawn, in order to use them later as fiduciary marks to place the labeled neurons more accurately. These sections were studied under brightfield, darkfield, and polarized illumination in a Dialux 20 Leitz microscope. Fluorochrome sections were studied with a Leitz-Ploemopak fluorescence system, with an A excitation filter (which provides a 360 nm wavelength excitation light), and 25 x , 4 0 x , and 63 x Leitz fluorescence objectives. The profiles of the sections and the location of the retrogradely labeled neurons were fed into a computer through an MD2 digitizer (Minnesota Datametrics) and plotted onto paper (Hewlett-Packard 7475A). After the labeled cortical neurons were plotted on the drawings of the sections, the architectonic areas were identified and their borders were also indicated on the same drawings. The terminology used to define the areas in the present report (see Fig. 1) follows that of Hassler and Muhs-Clement ('64) and Avendaiio and Verdu ('92) for the somatosensory cortex, Burton et al. ('82) for SII, and Avendaiio et al. ('88) for area 5. The motor cortical areas have been identified following Hassler and Muhs-Clement ('64), with modifications discussed below. The nomenclature of sulci and gyri was taken mostly from ReinosoSuarez ('61).

called this area 6uy, and its architectonic differences with area 6ap may be easily appreciated in Figure 2. Area 6ay more closely resembles area 6am, although it is narrower than the latter, and its layer VI is less densely populated.

Projections from area 3a to the motor cortex

Area 3a exhibited labeled neurons in all cases with injections in any portion of area 47, and also in the cases injected in areas 46 and 6ay. These projections are topographically organized along the mediolateral extent of area 3a [i.e., from the upper lip of the cruciate sulcus (SC) in the medial face of the hemisphere, through the posterior (GSP) and lateral sigmoid gyrus (GSL), to the cortex surrounding the rostroventral end of the coronal sulcus (SCo); see Avendano and Verdii, '921, and, less clearly, along its narrow rostrocaudal dimension. The projections directed to the medial portions of area 4y in the GSP and upper bank of the SC, as well as those to area 46, arise only from medial portions of area 3a, in the GSP and the upper lip of the SC (Figs. 3,4). After injecting more lateral portions of area 4y, in the lateral part of the GSP, retrograde labeling was widely distributed along the mediolateral extent of area 3a (Figs. 4,5). Injections in area 4y in the GSL and around the bottom of the SC gave rise to labeled neurons mostly in the intermediate one-third of area 3a, with a few neurons RESULTS reaching its lateral end around the SCo (Figs. 6, 7). Finally, Architectonic divisions of the motor cortex injections in area 4y in the GSA or the lower bank of the SC Our observations basically confirm Hassler and Muhs- (Fig. 8 ) , and injections in area 6ay (Figs. 2B, 8) produced Clement's ('64) description of two major divisions, areas 4 retrograde labeling only in the lateral extreme of area 3a, and 6, of the motor cortex of the cat. Each division namely in the coronal gyms (GCo) and in the cortex around comprises several subareas, all of which have in common the rostroventral end of the SCo. Our material did not provide a clear-cut answer to the the lack of a granular layer IV, and, in general, area 6 could be distinguished from area 4 by its relatively uniform question of whether neurons projecting to area 4 exhibited cellularity and its larger cortical thickness. However, we any topographical preference regarding the rostrocaudal have found consistently that area 6ap is more restricted dimension of area 3a, i.e., between its rostral (with area 47) than previously described: this area occupies only the and caudal (with area 3b) borders. It seemed, nevertheless, medial one-third of the GSA and ends at about the rostrodor- that in most cases the neurons showed a tendency to sal end of the SPs (Fig. 1).The cortex intervening between concentrate rostrally in area 3a, near the 3a-4y border, area 6ap and area 4y in the GSA is, to our view, a rostral regardless of their mediolateral location within area 3a extension of area 6aa. More ventrally, near the rostrodorsal (Figs. 4,6). Exceptions happened, particularly in the caudal end of the SPs, a new area appears which extends ventrally "pocket" of area 3a around the bottom of the SCo (Avenin the lateral bank of the SPs, where Hassler and Muhs- dano and Verdii, '92), where labeled neurons approached in Clement placed the rostral part of their area 6ap. We have some cases the area 3a-3b border (Figs. 7,9), or, in the case

c. AVENDANOET AL.

376

SL

Fig. 1. A: Surface map of motor and parietal areas of the cat’s brain, represented on lateral (left), dorsal (center),and medial (right) views of the rostral portion of the left hemisphere taken from the standard map of Reinoso-Suarez (’61).Areal borders, indicated by dashed lines, have been drawn according to Hassler and Muhs-Clement (’641, with modifications (Avendano et al., ’88; Avendano and Verdu, ’92).The corpus callosum and the anterior commissure are shown in jet black. B: Schematic representation of the location of tracer injections in the cat motor cortex. The diagram on the left shows an oblique, rostral, dorsal, and medial view of the rostral portions of the left hemisphere. The

medial face of the hemisphere is shaded (dotted area). Curved arrows indicate the direction along which the upper hank of the SC and the lateral hank of the SPs are pulled, so that the cortex buried in these sulci may be exposed. The diagram on the right shows these sulci already opened, and the approximate extent of the injections carried out in the motor cortex. Short, curved rungs mark the bottom of the sulci. Open injection profiles correspond to cases which did not have labeled neurons in area 3a. Combined injections in distant areas are indicated within circles.

of area 3a projections to area 46 (Fig. 4),where labeled neurons distributed evenly across the rostrocaudal extent of area 3a. Among 698 labeled cells counted in area 3a in five animals injected in area 4y, 503 (72%) were distributed in

supragranular layers, mainly layer 111. Most infragranular neurons appeared around the bottom of the SCo, and, for reasons that are not clear (but see Discussion),the proportion of supragranular neurons projecting to the motor cortex was larger in HRP than in fluorochrome cases.

CORTICAL AFFERENTS TO MOTOR CORTEX FROM AREA 3a

Fig. 2. A Photomicrograph of area 6ay in the lateral bank of the presylvian sulcus. B: Low-power photomicrograph of a coronal section of the frontal lobe in case G 216, showing an HRP injection in area 6a-y. C: Photomicrograph of area 6aB in the medial part of the GSA. The cytoarchitecture of this area is clearly different from that of area 6a. Calibration bars, A and C: 250 pm; B: 1 mm.

1

Fig. 3. Coronal sections through the frontoparietal cortex in a case that received a FB injection in area 3a in the upper lip of the SC, and a DY injection in an adjacent portion of area 47. FB-labeled neurons (circles), and DY-labeled neurons (dots) overlap in area 3a, but no double-labeled cells (asterisks) appear in this area. However, a few double-labeled neurons appear in other areas, such as 7m, 5, and 4,

mainly in infragranular layers. Although the motor cortex injection was located in the most medial and posterior sector of area 47, labeled neurons in area 3a distributed in several patches between the medialmost part of this area, in the upper lip of the SC, and the lateral part of the GSP.

c. AVENDANO ET AL.

378

G

405

Fig. 4. Single HRP injections in area 46 in the upper hank of the SC (top), and in area 4y in the lateral part of the GSP (bottom).Labeled neurons in area 3a appeared in both cases in the medial portions of this area, in the upper lip of the SC. However, some projections to area 4 also originated from more lateral portions of area 3a, up to the dorsal bank of the SCo (bottom-leftdiagram).

Results from double-labeling experiments In one case (GA 184, Figs. 3, 10) DY was injected in the medial portion of area 4y in the GSP, and FB in the medialmost portion of the GSP, involving area 3a in the upper lip of the SC, and the adjacent 3a-4y border zone. Neurons containing either of the dyes were found in area 3a in the GSP, and caudally along the upper lip of the SC, also invading area 7m (Avendaiio and Verdu, '92). Despite a manifest overlap in several parts of area 3a of the neurons labeled by each dye, no double-labeled cells were found. When simultaneous injections of different fluorescent dyes were placed in area 4y and in areas 2 (Fig. 6), 5 (Figs. 7, 91, or SII (Fig. 91, topographical overlap of retrogradely labeled cells was also evident, yet few double-labeled neurons were found. In case GA 189, which had the largest

number of double-labeled neurons, these accounted for less than 4% of the cells projecting to area 4y and about 10% of those projecting to area 2. In the cases injected in areas 4 and 5, these figures remained below 3%. Moreover, double labeling in these cells was weak, one of the dyes being hardly visible when objectives below 40x were used (Fig. 10). An interesting feature of the double-labeled cells is that their laminar distribution almost reversed that of the overall population of retrogradely labeled neurons in area 3a, showing a remarkable preference for appearing in layer V (Fig. 6). This feature was not confined to neurons in area 3a, since similar findings were obtained for neurons in area 4 projecting to other sectors of the same area (case GA 175, not shown in figures), and neurons in area 7m projecting to areas 3a and 4y (case GA 184, Fig. 3).

CORTICAL AFFERENTS TO MOTOR CORTEX FROM AREA 3a

379

Fig. 5. Photomicrographs of the HRP injection site (A),and of labeled neurons in medial (B)and lateral (C) parts of area 3a in case G 224, also shown in previous figure. Calibration bars, A: 1mm; B and C: 200 um.

DISCUSSION The present study confirms the existence of direct projections from area 3a to the motor cortex in the cat, and analyzes various aspects of the horizontal and radial (laminar) distribution of the neurons of oriein " of these Droiections. Before discussing the specific observations in their anatomical and functional context, however, it seems conve1

nient to comment on some conceptual issues pertaining to the motor cortex and area 3a

Definition of motor cortex and area 3a in the cat

0

A large body of knowledge has accumulated in the last 100 years on the localization, structure, and function of the

c. AVENDANO ET AL.

380

GA 189 Fig. 6. Coronal sections taken at the levels indicated in the diagram at the lower right corner of the figure, from a case that received two FB injections in area 2, and two DY injections in area 4y in the GSL. Dashed line within the cortex marks the upper boundary of layer V. Other conventions as in Figure 3.

motor cortex in primates (Hepp-Reymond, ’88; Asanuma, ’89). Much less is known, however, about the equivalent cortical region of carnivores, including cat. In attempting to establish areal equivalences across species, it is usually hopeless to rely only on direct comparisons between cortical architectonics. This may not be true, however, for those cortical areas with some remarkable features, as is the case of area 4y in cat and area 4 in primates, characterized in both species by the presence of large and giant pyramidal cells in layer V and the absence of a granular layer IV.In addition to cytoarchitecture, these areas exhibit in cat and monkey comparable physiological properties (Nieoullon and Rispal-Padel, ’76;Asanuma, ’891, and thalamic connections (Moran and Reinoso-Suarez, ’88; Matelli et al., ’89;

Darian-Smith et al., ’90). These similarities allow for extending to area 4y in the cat the functional terms “primary motor cortex” or “MI” applied-at least in a restricted sense (see Wise, ’85)-to area 4 in the monkey. The situation is less clear regarding the functional meaning of the various divisions of area 6, and divisions of area 4 other than 4y in the cat, mainly because of the paucity of physiological data on cat’s “motor” cortex outside area 4y. Two sets of findings, however, suggest that gross equivalences may be established between these areas and the premotor and perhaps part of the motor areas in the monkey. Firstly, Hassler and Muhs-Clement’s (’64)areas 4 and 6 encompassed all cortical regions lacking an internal granular layer, and situated between the overtly granular

c. AVENDANO ET AL.

382

G 216

Fig. 8. Coronal sections from a case that received a single HRP injection in area 4y and the 4y-6aa border region in the lower bank of the SC (top),and a case injected in area 6ay in the lateral bank of the SPs (bottom).Labeled neurons are represented by dots.

The identification of area 3a in cat, and its correspondence to area 3a in monkey, have been discussed in the accompanying paper (Avendano and Verdii, '92) and will not be further commented on here.

Connections from area 3a to motor cortex: a species-specific projection? The existence of direct projections from area 3a to the motor cortex in the cat was first proposed by Jones and Powell ('68) using silver impregnation methods, although they could not prove it conclusively, since none of the lesions carried out in the somatic sensory cortex was

restricted to area 3a. With the same techniques, but performing minute needle stitch lesions in an electrophysiologically defined portion of area 3a in the GSP, Grant et al. ("75) demonstrated two or three separate foci of terminal projections in area 4y. This divergent pattern of the area 3a to area 4 projection is reciprocated by the mediolateral dispersion of patches of labeled neurons projecting to restricted foci of the motor cortex in our findings. The only previous retrograde tracer study dealing with this issue (Yumiya and Ghez, '84) reported a more restricted topographic relationship between areas 3a and 4.The presence of a significant degree of convergence and divergence in this

381

CORTICAL AFFERENTS TO MOTOR CORTEX FROM AREA 3a

GA 163

Fig. 7. Coronal sections taken at the levels indicated in the diagram at the lower right corner of the figure, from a case that received three FB injections in area 5, and one DY injection in area 4y in the GSL. Conventions as in Figure 6.

parietal (or postcentral) cortex and the thinly granular frontal cortex. Monkey's motor and most of the premotor cortices are likewise agranular, and are situated between granular postcentral and granular frontal cortical areas. Secondly, the main thalamic afferents to areas 4, to 6aa and the rostral part of 6ap in cat (Moran and Reinoso-Suarez, '88), and to motor and premotor areas in monkey (Matelli et al., '89; Darian-Smith et al., '90) arise in a topographically organized fashion from the ventrolateral complex (including the oral division of the ventral posterior nucleus in monkey), the ventral anterior nucleus, and the lateral portion of the medial dorsal nucleus. Finer details of the correspondence between areal subdivisions in cat and monkey are not readily apparent. Areas 4sf and 4fu (Hassler and Muhs-Clement, '64) have not been explored physiologically, and their thalamic connections, as

well as those directed to the caudal part of 6ap, arise from a complex mixture of ventral, lateral, medial, and anterior thalamic nuclei (Moran and Reinoso-Suarez, '88). Also, it has not been proved that areas 6aa and 6ap in cat somehow correspond to Vogt and Vogt's ('19) homonymous areas in monkey, whose relationship to the physiologically defined premotor fields in monkey is uncertain (Wise, '85). Likewise, no satisfactory counterpart in the cat of the monkey supplementary motor cortex has been found as yet (Isla, '91). On the other hand, the rostral part of Hassler and Muhs-Clement's area 6ap, which we have defined as area 6ay (see above), may correspond to a lateral ventral sector of the motor, and perhaps also the premotor, monkey cortex, considering its motor representation (Nieoullon and Rispal-Padel, '76) and thalamic connections (Moran and Reinoso-Suarez, '88; Matelli et al.,'89).

383

CORTICAL AFFERENTS TO MOTOR CORTEX FROM AREA 3a

GA 191

GA 186

Fig. 9. Coronal sections from a case that received three FB injections in SII and two DY injections in area 4y (top), and another case that received four FB injections in area 5 and four DY injections in area 4y (bottom). The sections selected show levels where the overlap of FB- and DY-labeled cells was maximal. Conventions as in Figures 6 and 7.

projection may be related to the existence of multiple representations of the same body parts in area 4 (Pappas and Strick, ’79, ’81),and also to the fact that two-thirds of motor cortex-projecting neurons of area 3a have receptive fields which are different from those of the area 4 neurons upon which they project monosynaptically (Asanuma et al., ’82). Experiments in monkey were less conclusive than those in cat, and for several years the dominant idea was that area 3a did not project directly to area 4 in the monkey. This idea was incorporated in recent reviews on the motor cortex (Hepp-Reymond, ’88; Asanuma, ’89). However, there is only one well-controlled anatomical study which supported this view, and probably led to its acceptance (Jones et al.,

’78). In this study and in a later review (Jones and Porter,

’801, nevertheless, the rejection of direct motor cortical connections from area 3a was contingent upon a “strict” definition of area 3a, i.e., the rostra1 portion of the postcentral cortex in which layer IV thins, but which lacks giant pyramidal cells in layer V. In fact, Jones and Porter (’80) admitted that, if the region containing scattered giant cells in layer V were included in area 3a, then the presence of “nonspecific,” short intracortical connections would make it true that area 3a projects to area 4. No other anterograde studies are available which could shed light on this issue, but a few recent retrograde studies provide sound evidence in favor of the existence of direct connections from area 3a to the motor cortex. Ghosh et al. (’87) observed a substan-

c. AVENDARO ET AL.

384

Fig. 10. Fluorescence photomicrographs from cases GA 184, injected with FB in area 3a and with DY in area 4y (A,E), and GA 189, injected with FB in area 2 and with DY in area 4y (C,D). A: FB (left) and DY (right) injections in the upper bank of the SC. E: Numerous DY-labeled neurons in all layers, and a few FB-labeled neurons (mostly in deep layers), in area Tin, slightly behind the 7m-3a border. C:

High-power view of a dense population of DY-labeled neurons in layer IIIa of area 3a in the GSL; arrow points to a pyramidal cell weakly labeled also by FB. D:Double-labeled cell in layer Ilk a little behind the cells shown in C. Calibration bars: 500 pm (A), 250 +m (B), and 40 pm

tial number of labeled cells in area 3a after placing a small injection of HRP in area 4 of a cynomolgous monkey. The injection was visually guided and the architectonic boundaries were determined on the basis of Nissl staining. Labeled neurons were all supragranular and, although spreading along the rostrocaudal extent of area 3a, were more abundant in its rostral half. Huerta and Pons ('90) in the macaque monkey, and Stepniewska et al. ('90) in the owl monkey confirmed these projections with fluorescent tracers and using intracortical microstimulation to guide the injections. The preference for a rostral location in 3a of the labeled neurons, which we noticed in several of our cats, was also shown in monkey by Huerta and Pons ('901, who'

likewise observed a small proportion of labeled neurons in layers V and IV, in addition to a substantial population of neurons in layer 111. No other species have been investigated in search of motor cortex afferents from area 3a, mainly because the latter has not been clearly defined outside cats and primates. However, some findings in the rat merit a brief comment in this respect. It is known that interspersed within the densely granular cortex of SI, and partially surrounding it, there is a "dysgranular" cortex responsive to muscle and joint stimulation (Chapin and Lin, '84). When small HRP injections are placed in the forelimb and head region of the rat motor cortex (i.e., in a region where

(C,D).

CORTICAL AFFERENTS TO MOTOR CORTEX FROM AREA 3a somatic and motor areas do not overlap), it is precisely in the dysgranular cortex where more retrogradely labeled neurons appear (Donoghue and Parham, '83). Still, the detection of a great cutaneous-muscular convergence over large areas of the somatosensory and motor regions has led other authors to reject the existence in rat of an area 3a comparable to that in cat or monkey (Gioanni, '87).

Do projections to motor cortex and to other areas arise from separate sets of cells in area 3a? It is a common observation in anatomical studies of corticocortical connections that small lesions or deposits of anterograde tracers in a given focus usually produce a number of separate patches of anterograde degeneration or labeling. This projection pattern may be uncovering in the cortex either a close intermingling of cells with separate projection targets, a frequent occurrence of branching axons, or both. Our double-labeling studies indicate that very few neurons project to two separate areas, despite a large degree of overlap in some places of cells projecting to each area. Morphological studies at the single cell level, however, have shown that the axons from different kinds of cells in primary sensory cortices often give off collaterals along their course which form several restricted terminal clusters even at long distances from the soma (Gilbert and Wiesel, '83; De Felipe et al., '86). Although there is a relatively high degree of divergence in the corticocortical connectivity, there is indeed a preference within the somatosensory and motor cortex for establishing corticocortical connections between somatotopically matching regions (Waters et al., '82; Burton and Kopf, '84; Avendafio et al., '88; Mori et al., '89; Manzoni et al., '90). Therefore, an explanation for the scarcity of double-labeled cells in our material might be found in the possibility of having injected noncorresponding areal sectors within the same animal. To some extent, this could have happened in the case injected in areas 4y and SII (GA 1911, since most of the cortical and thalamic neurons projecting to SII appeared in lower limb sectors of, respectively, SI and the ventroposterior thalamic complex, whereas the locus of the motor cortex injection corresponded mainly to the representation of the forelimb. Yet, neurons projecting to each injection site overlapped in area 3a in the GSP in this case. In the cases injected in areas 5 or 2, however, the predominant location of cortical and thalamic neurons was in forelimb sectors of somatic sensory and motor cortical areas and thalamic nuclei, as was that of the neurons projecting to area 4y in the same cases. Moreover, in these cases there was a remarkable topographical overlap in area 3a of neurons single-labeled by either of the dyes. Also, case GA 184 received two closely spaced injections in the medial portions of areas 3a and 4y, where the same body parts are represented (Nieoullon and Rispal-Padel, '76; Verdu, '89), and retrogradely labeled neurons projecting to either-but not both-of the injection sites appeared in the same discontinuous patches in area 3a in the GSP and the upper lip of the SC. These findings suggest that multifocal, multiareal projections arising from the same neuron may not be a general property of a majority of corticocortical neurons, or else that the various terminal patches arising from the same axon are small and restricted in spread, so that there is a low probability of several of them being simultaneously labeled by relatively restricted, separate injections of retrograde tracers.

385

The presence of many more fluorochrome-labeled than HRP-labeled neurons in deep cortical layers in our material may reflect a greater sensitivity of these dyes as retrograde tracers with respect to free HRP (Kennedy and Bullier, '85). These authors also showed that, in a number of cortical areas in the monkey, neurons projecting to both V1 and V2 were more frequent in infragranular layers. These findings, however, were obtained in cortical areas whose projections to V1 and V2 arise mainly from infragranular layers. In our material double-labeled neurons in area 3a appeared in an increased proportion in infragranular layers, in spite of the clear preeminence in absolute and relative terms of the supragranular projection. Although the interpretation of this finding is not straightforward and would demand additional electrophysiological studies, it is noteworthy that the multiareal simultaneous projection arises from subcortically projecting cells in area 3a, some of which could be pyramidal tract neurons (Zarzecki et al., '78). Therefore, these projections might simultaneously deliver "efference copies" of area 3a activity to a number of motor-sensory cortical areas from the same layer V cells which project to motoneurons.

Concluding remarks The presence of direct low threshold muscle afferent input to both motor cortex and area 3a led Jones and Porter ('80) to suggest that the latter would be mainly involved in muscle sensation, rather than in relaying deep afferent information to the motor cortex for regulating its outflow. The widespread distribution of area 3a connections to other somatosensory areas actually favors its role in relaying muscle input for conscious perception. However, the characteristics of activation of motor cortex and area 3a neurons by muscle stretches differ in threshold (lower in area 3a) and latency (shorter in area 3a: see Hepp-Reymond, '88, for review). In addition, the inactivation of area 3a impairs, although not eliminates, motor cortex responses to joint and muscle afferent fibers stimulation (Gioanni et al., '83). All of this suggests that area 3a provides the motor cortex with information which is different from that accessing it directly (probably through the thalamus: Asanuma et al., '79). Therefore, it may well be the case that the direct, topographically organized projection from 3a to the motor cortex shown in the present study is instrumental for long transcortical loops, providing the motor cortex with rapid feedback information on the ongoing movement.

ACKNOWLEDGMENTS The authors thank Ms. Maria Teresa Fernandez-Yuste for her help in preparing part of the photographic material, and Dr. Reinoso-Suarez for helpful comments on the manuscript. This work was supported by CICyT Grant PB87-0130.

LITERATURE CITED Amaya, C. (1991) Conexiones Aferentes Corticales de la Corteza Insular. Estudio Anatomic0 en el Gato. Madrid: Doctoral Thesis, Autonoma University. Asanuma, H. (1989) The Motor Cortex. New York Raven Press. Asanuma, H., K.D. Larsen, and P. Zarzecki (1979) Peripheral input pathways projecting to the motor cortex in the cat. Brain Res. 172:197-208.

386 Asanuma, H., R.S. Waters, and H. Yumiya (1982) Physiological properties of neurons projecting from area 3a to area 4y of feline cerebral cortex. J. Neurophysiol. 48: 1048-1057. Avendado, C., and A. Verdu (1992) Area 3a in the cat. I. A reevaluation of its location and architecture on the basis of Nissl, myelin, acetylcholinesterase and cytochrome oxidase staining. J. Comp. Neurol. 321:357-372. Avendano, C., E. Rausell, D. Perez-Aguilar, and S. Isorna (1988) Organization of the association cortical afferent connections of area 5: A retrograde tracer study in the cat. J. Comp. Neurol. 278:l-33. Burton, H., and E.M. Kopf (1984) Ipsilateral cortical connections from the second and fourth somatic sensory areas in the cat. J. Comp. Neurol. 225:527-553. Burton, H., G. Mitchell, and D. Brent (1982) Second somatic sensory area in the cerebral cortex of cats: Somatotopic organization and cytoarchitecture. J. Comp. Neurol. 210:109-135. Chapin, J.K., and C . 3 Lin (1984) Mapping the body representation in the SI cortex of anesthetized and awake rats. J. Comp. Neurol. 229:199-213. Darian-Smith, C., I. Darian-Smith, and S.S. Cheema (1990) Thalamic projections to sensorimotor cortex in the macaque monkey: Use of multiple retrograde fluorescent tracers. J. Comp. Neurol. 299: 17-46. De Felipe, J., M. Conley, and E.G. Jones (1986) Long-range focal collateralization of axons arising from corticocortical cells in monkey sensorymotor cortex. J. Neurosci. 6:3749-3766. Donoghue, J.P., and C. Parham (1983) Afferent connections of the lateral agranular field of the rat motor cortex. J. Comp. Neurol. 217t390404. Esteky, H., and H.D. Schwark (1990) Intrinsic corticocortical connections of cat primary somatosensory cortex. SOC. Neurosci. Abstr. 16:228. Ghosh, S., C. Brinkman, and R. Porter (1987) A quantitative study of the distribution of neurons projecting to the precentral motor cortex in the monkey (M.fusccculurzs).J. Comp. Neurol. 259,424444. Gilbert, C.D., and T.N. Wiesel (1983) Clustered intrinsic connections in cat visual cortex. J. Neurosci. 3t1116-1133. Gioanni, Y. (1987) Cortical mapping and laminar analysis of the cutaneous and proprioceptive inputs from the lat foreleg: An extra- and intracellular study. Exp. Brain Res. 67t510-522. Gioanni, Y., J. Everett, and M. Lamarche (1983) The transcortical reflex triggered by cutaneous or muscle stimulation in the cat with a penicillin epileptic focus: Relative importance of regions 3a and 4. Exp. Brain Res. 51:57-64. Grant, G., S. Landgren, and H. Silfvenius (1975) Columnar distribution of U-fibers from the postcruciate cerebral projection area of the cat’s group I muscle afferents. Exp. Brain Res. 24:57-74. Hassler, R., and K. Muhs-Clement (1964) Architektonischer Aufbau des sensomotorischen und parietalen Cortex der Katze. J. Hirnforsch. 6377420. Hepp-Reymond, M.-C. (1988) Functional organization of motor cortex and its participation in voluntary movements. In H.D. Stecklis and J. Erwin (eds): Comparative Primate Biology, Vol. 4: Neurosciences. New York: Alan R. Liss, pp. 501-624. Herman, D., R. Kang, M. MacGillis, and P. Zarzecki (1985) Responses of cat motor cortex neurons to cortico-cortical and somatosensory inputs. Exp. Brain Res. 57:598-604. Huerta, M.F., and T.P. Pons (1990) Primary motor cortex receives input from area 3a in macaques. Brain Res. 537:367-371. Isla, A.J. (1991) Conexiones Cortico-corticales al as Areas 4 y 6 de la Corteza Motora en el Gato. Madrid: Doctoral Thesis, Autonoma University. Jones, E.G., and R. Porter (1980) What is area 3a? Brain Res. Rev. 21-43, Jones, E.G., and T.P.S. Powell (1968) The ipsilateral cortical connexions of the somatic sensory areas in the cat. Brain Res. 9t71-94. Jones, E.G., J.D. Coulter, and S.H.C. Hendry (1978) Intracortical connectivity of architectonic fields in the somatic sensory, motor and parietal cortex of monkeys. J. Comp. Neurol. 181:291-348. Kennedy, H., and J. Bullier (1985) A double-labeling investigation of the afferent connectivity to cortical areas V1 and V2 of the macaque monkey. J. Neurosci. 5:2815-2830. Kosar, E., R.S. Waters, N. Tsukahara, and H. Asanuma (1985) Anatomical and physiological properties of the projection from the sensory cortex to the motor cortex in normal cats: The difference between corticocortical and thalamocortical projections. Brain Res. 345t68-78. Manzoni, T., P . Barbaresi, and S. Bernardi (1990) Matching of receptive fields in the association projections from SI to SII of cats. J. Comp. Neurol. 300:331-345.

c. AVENDA&O ET AL. Matelli, M., G. Luppino, L. Fogassi, and G. Rizzolatti (1989) Thalamic input to inferior area 6 and area 4 in the macaque monkey. J. Comp. Neurol. 280:468488. McCloskey, D.I. (1978) Kinesthetic sensibility. Physiol. Rev. 58t763-820. Mesulam, M.-M. (1978) Tetramethylbenzidine for horseradish peroxidase neurohistochemistry. A non-carcinogenic blue-reaction product with superior sensitivity for visualizing neuronal afferents and efferents. J. Histochem. Cytochem. 26: 106-117. Moran, M.A., and F. Reinoso-SuArez (1988) Topographical organization of the thalamic afferent connections to the motor cortex in the cat. J. Comp. Neurol. 270:64-85. Mori, A., R.S. Waters, and H. Asanuma (1989) Physiological properties and patterns of projection in the cortico-cortical connections from the second somatosensory cortex to the motor cortex, area 47, in the cat. Brain Res. 504:206-2 10. Nieoullon, A., and L. Rispal-Padel (1976) Somatotopic localization in cat motor cortex. Brain Res. 105:405-422. Pappas, C.L., and P.L. Strick (1979) Double representation of the distal forelimb in cat motor cortex. Brain Res. 167:412-416. Pappas, C.L., and P.L. Strick (1981)Physiological demonstration of multiple representations in the forelimb region of the cat motor cortex. J. Comp. Neurol. 200:481-490. Phillips, C.G. (1969) Motor apparatus of the baboon’s hand. Proc. R. SOC. Lond. [Biol.] 173:141-174. Phillips, C.G., T.P.S. Powell, and M. Wiesendanger (1971) Projection from low-threshold muscle afferents of hand and forearm to area 3a of baboon’s cortex. J. Physiol. 21 7:419446. Pons, T.P., and J.H. Kaas (1986) Corticocortical connections of area 2 of somatosensory cortex in macaque monkeys: A correlative anatomical and electrophysiological study. J. Comp. Neurol. 248t313-335. Porter, L.L., and K. Sakamoto (1988) Organization and synaptic relationships of the projection from the primary sensory to the primary motor cortex in the cat. J. Comp. Neurol. 271:387-396. Porter, L.L., T. Sakamoto, and H. Asanuma (1990) Morphological and physiological identification of neurons in the cat motor cortex which receive direct input from the somatic sensory cortex. Exp. Brain Res. 8Ot209-212. Reinoso-Su&ez, F. (1961) Topographischer Hirnatlas der Katze. Darmstadt: Merck, A.G. Stepniewska, I., A. Morel, and J. Kaas (1990) Connections of electrophysiologically defined locations in primary motor cortex, MI, of owl monkeys (Aotus triuirgutus).SOC.Neurosci. Abstr. 16:240. Thompson, F.J., J. Fernandez, H. Asanuma, and K. Kubota (1973) Relationship between 3a sensory cortex and motor cortex in the cat. Fed. Proc. 32340. Thompson, W.D., S.D. Stoney, Jr., and H. Asanuma (1970) Characteristics of projections from primary sensory cortex to motorsensory cortex in cats. Brain Res. 2215-27. Verdu, A. (1989) Estructura y conectividad de la corteza parietal anteromedial: Un estudio anatomic0 y electrofisiol6gico en el gato. Madrid: Doctoral Thesis. Autonoma University. Vogt, C., and 0. Vogt (1919) Allgemeinere Ergebnisse unserer Hirnforschung. J. Psychol. Neurol. (Lpz.)25.279-462. Waters, R.S., 0. Favorov, and H. Asanuma (1982) Physiological properties and pattern of projection of cortico-cortical connections from the anterior bank of the ansate sulcus to the motor cortex, area 4y, in the cat. Exp. Brain Res. 46:403412. Wiesendanger, M. (1973) Input from muscle and cutaneous nerves of the hand and forearm to neurones of the precentral gyrus of baboons and monkeys. J. Physiol. 228:203-219. Wise, S.P. (1985) The primate premotor cortex: Past, present, and preparatory. Annu. Rev. Neurosci. 8:l-19. Yumiya, H., and C. Ghez (1984) Specialized subregions in the cat motor cortex: Anatomical demonstration of differential projections to rostra1 and caudal sectors. Exp. Brain Res. 53.259-276. Zarzecki, P. (1989) Influence of somatosensory cortex on different classes of cat motor cortex output neuron. J. Neurophysiol. 62487-494. Zarzecki, P., P.L. Strick, and H. Asanuma (1978) Input to primate motor cortex from posterior parietal cortex (area 5). 11. Identification by antidromic activation. Brain Res. 157:331-335.