[1] 18w Reevaluation of Ipsilateral Corticocortical Inputs to the Orofacial Region of the Primary Motor Cortex in the Macaque Monkey

[1] 238w An anatomical approach to possible areas in the cerebral cortex involved in somatic motor behavior is to analyze the cortical areas containing neurons that connect directly to the primary motor cortex (MI). To define the cortical areas related to orofacial movements, we examined the distribution of cortical neurons that send their axons to the orofacial region of the MI in the macaque monkey. Injections of retrograde tracers into the electrophysiologically identified orofacial region of the MI revealed that labeled neurons were distributed in the following cortical areas: the orbital cortex (area 12), insular cortex, frontoparietal operculum (including the deep part of the cortical masticatory area and the secondary somatosensory cortex), ventral division of the premotor cortex (especially in its lateral part), orofacial region of the supplementary motor area, rostral division of the cingulate motor area (CMA), and CMA on the ventral bank. A number of labeled neurons were also seen in the MI around the injection sites and in the parietal cortex (including the primary somatosensory cortex and area 7b). No labeled neurons were found in the dorsal division of the premotor cortex. Fluorescent retrograde double labeling further revealed virtually no overlap of distribution between cortical neurons projecting to the orofacial and forelimb regions of the MI. Based on the present results, we discuss the functional diversity of the cortical areas related to orofacial motor behavior and the somatotopical organization in the premotor areas of the frontal cortex.

[1] 201w ICMS of the lateral part of the MI on the precentral gyrus elicited orofacial movements in each of the six monkeys. Movements of the distal part of the forelimb were evoked in the adjacent, more medial aspect of the MI (Fig. 1). In each animal, a tracer was injected into the center of the region that produced low-threshold responses in the lip, jaw, and tongue. In two of these animals, a second fluorescent tracer was injected into the center of the forelimb representation, where digit and wrist movements could be evoked (Table 1). In subsequent histological examination of coronal sections, we defined the regions of tracer uptake as those areas where the neuropil was homogeneously stained by the tracer. Histological examination of the extent of the sites of tracer injection confirmed that the injection sites of WGA-HRP (in cases 4E6F, 4172, and 596F; see Figs. 1b, 2a), FB (in cases 546B and 4869; see Fig. 1c), and DY (in case 536172; see Fig. 1d) were restricted to the orofacial region of the MI. The injection sites of DY (in case 4869; see Fig. 1c) and FB (in case 536172; see Fig. 1d) were localized to the forelimb region of the MI.

[2] 62w Numerous retrogradely labeled neurons were seen in multiple areas of the cerebral cortex ipsilateral to the injection sites in each monkey. These labeled neurons were located mainly in layers III and V and, additionally, in layer IV. They were small or medium-sized pyramidal neurons (Fig. 2b-d). In the present paper, we illustrate data from three representative cases (cases 595F, 4869, and 536172).

[3] 179w After tracer injection into the orofacial region of the MI, a large number of retrogradely labeled neurons were found in the orbital cortex, which corresponds to area 12 (Fig. 3, sections 30 and 66). Some labeled neurons were extended medially into part of area 13. Dorsal to area 12, clusters of the labeled neurons continued throughout the lateral portion of the PMv. Considerable numbers of labeled neurons were seen in the insular cortex and frontoparietal operculum (Figs. 3,4,sections 78,102,114,138,150,and 162; see also Fig. 5). This labeling probably included the deep part of the cortical masticatory area (CMaAd; Fig. 2d; see also Godschalk et al., 1984;Huang et al., 1989) and the secondary somatosensory cortex. No labeled neurons were observed in the orbital cortex after tracer injection into the forelimb region of the MI. In the frontoparietal operculum, we found a cluster of neurons projecting to the forelimb region of the MI. These labeled neurons were located in an area that probably corresponded to the secondary somatosensory cortex (Fig. 4, section 162; Figs. 5, 8; see also Tokuno and Tanji, 1993).

[4] 115w It is generally accepted that the PM, which is situated in the lateral aspect of Brodmann's area 6, is classified into two subdivisions. These are termed the PMd and PMv (for reviews, see Kurata, 1989;Wise et al., 1991). The PMd corresponds to cytoarchitectonically defined area 6D of Barbas and Pandya (1987) and the caudal part of histochemically defined area F2 of Matelli et al. (1985Matelli et al. ( , 1989)). On the other hand, the PMv corresponds to areas 4C and 6V of Barbas and Pandya and areas F4 and F5 of Matelli et al. The boundary between the PMd and the PMv is considered to be located around the spur of the arcuate sulcus.

[5] 47w In the PMd, a number of retrogradely labeled neurons were found after tracer injection into the forelimb region of the MI (Fig. 7, section 128; see also Fig. 8). Virtually no neurons in the PMd were labeled after tracer injection into the orofacial region of the MI.

[6] 137w After tracer injection into the orofacial region of the MI, labeled neurons in the PMv (Fig. 2c) were distributed extensively in its lateral part (Fig. 3, sections 66 and 78; Fig 7, sections 92, 110, and 128; see also Figs. 5, 8). On the other hand, most PMv neurons that were labeled from the forelimb region of the MI were observed more medially in the posterior bank of the arcuate sulcus, particularly in an area around the genu of the arcuate sulcus (Fig. 7, sections 92, 110, and 128; see also Fig. 8). Cases 4869 and 536172 showed that there was no overlap of distribution between PMv neurons labeled from the orofacial region of the MI and those labeled from the forelimb region of the MI (Fig. 7, sections 92, 110, and 128; see also Fig. 8).

[7] 116w In each of the six monkeys that received tracer injection into the orofacial region of the MI, labeled neurons were consistently found in the medial wall of the hemisphere, Fig. 4. Distribution of retrogradely labeled cortical neurons after WGA-HRP injection into the orofacial region of the MI in case 596F (see also Figs. 1b,2a,3; continued from Fig. 3). Each number indicates the position of the section in a series of serial frontal 60-µm-thick sections that are arranged rostrocaudally. The approximate positions of the sections are also specified by dashed lines in the lateral aspect of the hemisphere in Figure 3. CMAv, cingulate motor area on the ventral bank; SI, primary somatosensory cortex; SII, secondary somatosensory cortex.

[8] 173w which corresponds to the SMA (medial aspect of Brodmann's area 6; Fig. 2b). These labeled neurons were localized to the dorsal portion of the SMA and were distributed in small clusters (Fig. 3, sections 66 and 78; Fig. 6, sections 92 and 110). In cases 596F and 546B, ICMS mapping of the medial wall of the hemisphere revealed that somatotopical representations in the SMA were arranged rostrocaudally in the order of the orofacial, forelimb, trunk, and hind limb (Fig. 6a; Mitz and Wise, 1987;Luppino et al., 1991). In these monkeys, the pattern of distribution of SMA neurons labeled from the orofacial region of the MI was superimposed on the somatotopical map determined by ICMS (Fig. 6). The labeled neurons were distributed exclusively in areas representing orofacial movements, especially lip movements, which were elicited at sites in the dorsal portion of the SMA (Fig. 6d,e). Forelimb movements were often elicited at sites in the more ventral portion of the SMA as well as those in the more caudal portion of the SMA (Fig. 6d-f).

[9] 281w In cases in which two fluorescent tracers were injected (cases 4869 and 536172), neurons labeled from the orofacial region of the MI were restricted to the dorsal portion of the SMA. Neurons labeled from the forelimb region of the MI were located more ventrally and caudally in the medial wall of the hemisphere (Fig. 7, sections 92, 110, and 128; see also Fig. 8). Virtually no overlap of distribution was Fig. 6. a: Results of ICMS mapping of the SMA in case 596F. The dorsal view of the monkey brain shows the area in the medial wall of the hemisphere (rectangle) where ICMS mapping was performed. Each circled number represents the site of electrode penetration that is specified by numbered lines in b-f. The body parts in which movements were evoked are indicated as follows: Ak, ankle; H, hip; S, shoulder; Tr, trunk. b-f: Distribution of retrogradely labeled neurons in the SMA after WGA-HRP injection into the orofacial region of the MI in case 596F. The locations of labeled neurons were superimposed on the somatotopical map. Each solid circle represents one labeled neuron. Each section number indicates the position of the section in a series of serial frontal 60-µm-thick sections that are arranged rostrocaudally. The approximate positions of the sections are also indicated in a. Open circles represent the sites of ICMS, and the body parts in which movements were elicited are indicated as follows: Ar, arm; B, bucca; D, digit; E, elbow; Ey, eye; LL, lower lip; S, shoulder; UL, upper lip; W, wrist; Ϫ, no movement elicited. seen between SMA neurons labeled from the orofacial region of the MI and those labeled from the forelimb region of the MI.

[10] 173w Around the cingulate sulcus, small patches of labeled neurons were found after tracer injection into the orofacial region of the MI 8). The patches were located in an area corresponding probably to the CMAr (area 24c) and the CMAv (area 23c). The pattern of distribution of patches differed from case to case (see Figs. 5,8), especially in the caudal part. In some cases, labeled neurons were distributed in the CMAv (area 23c). Our retrograde doublelabeling experiments revealed that, in the CMAr, the neurons labeled from the orofacial region of the MI were located rostral to an area containing the neurons labeled from the forelimb region of the MI (Fig. 8). More caudally, however, neurons labeled from the orofacial region of the MI were located caudal to an area containing the neurons labeled from the forelimb region of the MI. There was no overlap of distribution between the two neuronal populations in the CMA. No neurons that projected to the orofacial region of the MI were seen in area 6c, which corresponds the CMAd.

[11] 115w In each of the six monkeys, large numbers of retrogradely labeled neurons were seen in the precentral gyrus around the sites of tracer injection. After tracer injection into the orofacial region of the MI, the labeled neurons were distributed predominantly in an area lateral to the injection site (Figs. 3,4,sections 102,114,138,and 150; see also Figs. 5,8). After tracer injection into the forelimb region of the MI, the labeled neurons were distributed mainly in an area medial to the injection site (Fig. 7, section 254; see also Fig. 8). In retrograde double-labeling experiments, almost no intermingling of distribution was observed between MI neurons labeled from the orofacial region and those from the forelimb region (Fig. 8).

[12] 156w After tracer injection into the orofacial region of the MI, labeled neurons were found in the postcentral gyrus, which corresponds to the primary somatosensory cortex (Fig. 4, section 162; see also Figs. 5,8), especially to areas 3a, 1, and 2, and in the anterior part of the lateral bank and surface of the intraparietal sulcus, which corresponding to area 7b (Fig. 4, section 222; see also Figs. 5, 8) or area PF of Pandya and Seltzer (1982). After tracer injection into the forelimb region of the MI, labeled neurons were seen in the postcentral gyrus, which corresponds to the primary somatosensory cortex, and in the medial bank of the intraparietal sulcus, which corresponds to area 5 or area PEa (Pandya and Seltzer, 1982;Fig. 8). Small clusters of labeled neurons were seen in the lateral bank of the intraparietal sulcus, which corresponds to area POa (Pandya and Seltzer, 1982;Fig. 8; see also Leichnetz, 1986;Ghosh et al., 1987).

[13] 155w The present results indicate that CMA neurons sending their axons to the orofacial region of the MI are distributed in several separate areas, which probably correspond to the CMAr and the CMAv. No such neurons are apparent in an area corresponding to the CMAd. Previous studies have shown the rostrocaudal arrangement of CMAr neurons projecting to the forelimb and hind limb regions of the MI and cervical enlargement of the spinal cord (Hutchins et al., 1988;He et al., 1995). In addition, the present data indicate that the orofacial representation in the CMAr is located just rostral to the forelimb representation. An extensive ICMS mapping study (Luppino et al., 1991) has also shown that orofacial movements are elicited in the CMAr, which corresponds to area 24c defined by Matelli et al. (1991). Thus, the CMAr has a complete set of body part representations: The orofacial, forelimb, and hind limb parts are arranged rostrocaudally in the CMAr.

[14] 164w According to previous anatomical work on the distribution of corticospinal neurons that send their axons directly to the cervical and lumbar segments of the spinal cord (He et al., 1995), forelimb and hind limb representations exist in the CMAv. By injecting tracers into orofacial and forelimb regions of the SMA, Morecraft et al. (1996) have recently suggested that the orofacial and forelimb regions of the CMAv are arranged rostrocaudally. Thus, the CMAv can be considered to represent a complete set of body parts, including the orofacial, forelimb, and hind limb parts. The present results have also confirmed the orofacial representation in the CMAv. However, the neurons in the CMAv that project to the orofacial region of the MI are located more caudally than those that project to the forelimb region of the MI (Fig. 8). These data are contradictory to those of Morecraft et al. (1996). Further anatomical as well as electrophysiological investigations are needed to fully elucidate the somatotopical organization of the CMAv.

[1] 104w To define possible cortical areas, including the premotor areas, that are involved in various aspects of orofacial motor behavior, we have examined the distribution of cortical neurons that send their axons to the orofacial region of the MI. We have further demonstrated the patterns of distribution of cortical neurons that project to the orofacial and forelimb regions of the MI by using fluorescent retrograde double labeling. The present results suggest the somatotopical organization of premotor areas and cortical areas related to only orofacial movements but not to forelimb movements. All of them may play an important role in cortical motor control of orofacial movements.

[2] 205w The present data have shown that the anterior part of the lateral bank of the intraparietal sulcus, which corresponds to the anterior part of area 7 (area 7b or area PF of Pandya and Seltzer, 1982), contains some neurons that project directly to the orofacial region of the MI. On the other hand, neurons that project to the forelimb region of the MI are not evident in this area but, instead, are seen in the medial bank of the intraparietal sulcus, which corresponds to area 5 or area PEa of Pandya and Seltzer (1982), as previously reported (Jones et al., 1978;Strick and Kim, 1978;Godschalk et al., 1984;Leichnetz, 1986;Tokuno and Tanji, 1993). By using a single-unit recording technique in alert monkeys, the anteriormost part of area 7 has been revealed to contain neurons that respond to visual stimuli approaching the face, passive hand movements toward the mouth, and reaching and mouthing movements of the lips (Leinonen and Nyman, 1979). These data strongly suggest that the anterior part of area 7 may play significant roles in orofacial motor behavior. Thus, the present results indicate that the anterior part of area 7 is likely to exert some influence on the orofacial region of the MI through direct projection.

[3] 88w In addition, an anatomical study has shown that the anterior part of area 7 receives projection fibers directly from the posterior bank of the inferior limb of the arcuate sulcus, whereas the more posterior part of area 7 does not (Neal et al., 1990). It should be noted here that the posterior bank of the inferior limb of the arcuate sulcus is equivalent to the lateral part of the PMv, which contains neurons that project to the orofacial region of the MI, as demonstrated in the present study.

[1] 492w MI mapping and tracer injections. Experiments were performed on six female Japanese monkeys (Macaca fuscata) weighing 4.2-6.0 kg. Each monkey was anesthetized with ketamine hydrochloride (10 mg/kg body weight, i.m.) and sodium pentobarbital (30 mg/kg body weight, i.p.) and received surgery to gain easy access to the brain for electrophysiological mapping. The use of the animals in the present study followed the principles approved by the animal care committee at the National Institute for Physiological Sciences. Under aseptic conditions, the skull was widely exposed, and small screws were attached to the skull for anchors. The exposed skull and screws were covered completely with transparent acrylic resin. Two stainless-steel pipes were mounted in parallel with each other over the frontal and occipital lobes for head fixation. A few days after the surgery, the monkeys were anesthetized with ketamine hydrochloride (10 mg/kg body weight, i.m.) and xylazine hydrochloride (1-2 mg/kg body weight, i.m.), and they sat quietly in a primate chair with their heads fixed in a stereotaxic frame that was attached to the chair. Under aseptic conditions, a skull portion over the lateral part of the central sulcus was removed. Following recovery from the anesthesia, each monkey underwent ICMS to determine the boundary between the orofacial and the forelimb representation in the MI. Glass-insulated Elgiloy-alloy microelectrodes, the impedance of which measured 0.9-1.4 M⍀ at 500 Hz, were used for ICMS and for recording of extracellular unit activity. The cortex was stimulated through the electrode by currents of 5-50 µA (12 cathodal pulses of 200 µsec duration at 333 Hz through a constant-current stimulator), and evoked movements were observed. After ICMS mapping, electrolytic microlesions (anodal direct currents of 10-15 µA, 20-30 seconds) were placed at selected sites. The electrode was then removed, and two or three sites were selected for injection of wheat germ agglutinin-conjugated horseradish peroxidase (WGA-HRP) into the orofacial region of the MI in cases 4E6F, 4172, and 596F (Table 1). A total volume of 0.2-0.3 µl of a 4% solution of WGA-HRP (Toyobo, Tsuruga, Japan) dissolved in 0.1 M Tris-HCl buffer, pH 7.0, was injected by pressure through a 1-µl Hamilton microsyringe (Reno, NV) that was attached to the same manipulator as the electrode. The tip of the injection needle was set carefully at an insertion point of the electrode at the dural surface under a surgical microscope and was then moved to the intended injection point with the manipulator. In case 546B (Table 1), Fast blue (FB; Illing, Groβ-Umstadt, 1), the medial wall of the hemisphere was also mapped by using ICMS. After removal of a skull portion over the midline, the cortex was stimulated through the electrode by currents of 10-50 µA (22 cathodal pulses of 200 µsec duration at 333 Hz), and evoked movements were observed. Particular attention was paid to precise examination of the anterior part of the SMA that represents the orofacial part. Other technical details were similar to those for ICMS mapping of the MI.

[2] 152w After a survival period of 3-4 days for WGA-HRP (cases 4E6F, 4172, and 596F) and of 3 weeks for fluorescent dyes (cases 546B, 4869, and 536172), the monkeys were anesthetized deeply with an overdose of sodium pentobarbital (60 mg/kg body weight, i.p.) and perfused transcardially with 2 liters of phosphate-buffered saline, pH 7.3, followed by 5 liters of 8% formalin in 0.1 M phosphate buffer, pH 7.3. The monkeys were then perfused with 3 liters of 0.1 M phosphate buffer, pH 7.3, containing 10% sucrose and, finally, with 2 liters of the same buffer containing 30% sucrose. The brains were removed immediately and immersed in the same buffer containing 30% sucrose at 4°C until they sank. The brains were cut into blocks in the stereotaxic plane on a dissecting mold for the monkey brain. Hemispheres ipsilateral to the injection sites were then cut serially into frontal 60-µm-thick sections on a freezing microtome.

[3] 73w WGA-HRP. For visualizing the injected and transported WGA-HRP, every third or sixth section was reacted with tetramethylbenzidine (Mesulam, 1978) and then incubated in a 3% aqueous solution of ammonium molybdate for 15 minutes to stabilize the reaction product (Fujii and Kusama, 1984). The sections were mounted onto gelatin-coated glass slides, counterstained with 1% neutral red, and observed with a light microscope (Nikon, Tokyo, Japan) under brightfield and darkfield illumination and a profile projector.

[4] 83w Fluorescent dyes. Every third or sixth section was mounted, air dried, and observed with an epifluorescence microscope (Zeiss, Oberkochen, Germany) under an ultraviolet filter that provided excitation light of approximately 360 nm wavelength. These sections or adjacent sections were stained with 1% cresyl violet to identify the cortical areas that contained labeled neurons. In all cases, histological reconstruction of the positions of the tracks of the stimulation electrode and injection needle was carried out with the aid of electrolytic microlesions as reference points.

[5] 165w Data analysis. Every sixth section was printed onto photographic paper (20.3 ϫ 25.4 cm) by using an enlarger. Labeled neurons were plotted on the photographic paper. Representative sections were then selected to show the distribution of labeled neurons in the coronal plane. In addition, to demonstrate the distribution of labeled neurons and the amount of labeling in cortical areas, we made surface-view reconstructions in two cases (596F and 536172). Lines of layer V of each section were drawn, and the cortex was divided into about 1-mm bins approximately perpendicular to the lines of layer V with a pair of bow compasses. The number of labeled neurons was counted in each bin. The medial wall and lateral surface of the hemisphere were unfolded upward and downward, respectively. Unfolded lines were aligned on the junction of the medial wall and the lateral surface. The distribution of labeled neurons was depicted on the prepared unfolded map of the cortex with solid or open circles of four different sizes.

[1] 201w It is well known that multiple somatic motor areas exist in the frontal lobe of the cerebral cortex in primates (Wiesendanger, 1981;Tanji and Kurata, 1989;Wise et al., 1991;Tanji, 1994). Accumulated evidence indicates that these cortical motor areas are composed of the primary motor cortex (MI), supplementary motor area (SMA), premotor cortex (PM), and cingulate motor area (CMA), according to the nomenclature of Wise et al. (1991) and Tanji (1994). All of these areas contain representations of multiple body parts, such as the orofacial forelimb and hind limb. Some of the somatic motor areas may be subdivided. The PM is comprised of the dorsal and ventral divisions (PMd and PMv). The CMA is divided into three areas: the rostral CMA (CMAr), CMA on the dorsal bank (CMAd), and CMA on the ventral bank (CMAv; Dum and Strick, 1991b). In addition, recent anatomical and physi-ological studies have identified the presupplementary motor area (pre-SMA), which is also related to the control of forelimb movements (Luppino et al., 1991(Luppino et al., , 1993;;Matsuzaka et al., 1992). The pre-SMA is located just anterior to the SMA in the medial wall of the hemisphere and corresponds to cytoarchitectonically and chemoarchitectonically defined area F6 (Matelli et al., 1991).

[2] 206w According to Strick and his colleagues, the somatic motor areas of the frontal cortex that project monosynaptically to the MI can be regarded as the nonprimary motor or so-called ''premotor'' areas (see Dum and Strick, 1991a). Given such an operational definition, the premotor areas are composed of the SMA, PM, and CMA (Muakkassa and Strick, 1979). The remaining somatic motor area, the pre-SMA, is not ''premotor,'' because it does not project directly to the MI (Matsuzaka et al., 1992;Luppino et al., 1993). In this context, the input organization of the MI has been investigated extensively from an anatomical viewpoint. In particular, fiber connections to the forelimb region of the MI have continuously attracted much attention, because the control of orchestrated forelimb movements is likely to require complex neural mechanisms. In fact, many anatomical studies have been performed in attempts to clarify the cortical (Leichnetz, 1986;Dum and Strick, 1991a;Tokuno and Tanji, 1993) as well as the thalamic (Leichnetz, 1986;Holsapple et al., 1991;Hoover and Strick, 1993;Rouiller et al., 1994;Inase and Tanji, 1995;Shindo et al., 1995) origin of inputs to the forelimb region of the MI. Likewise, physiological studies in behaving monkeys have focused mostly on the control mechanisms underlying complex forelimb movements (for reviews, see Hepp-Reymond, 1988;Tanji, 1994).

[3] 203w Despite the increasing amount of literature on the connections and functions of the somatic motor areas related to forelimb movements, only a limited amount of information has been available so far on the somatic motor areas involved in various aspects of orofacial motor behavior. Previous electrical stimulation (Gru ¨nbaum and Sherrington, 1902;McGuinness et al., 1980;Huang et al., 1988) and single-unit recording studies (Hoffman and Luschei, 1980;Murray and Sessle, 1992a-c) have revealed that a wide region in the lateral part of the MI participates in orofacial movements, such as movements of the face, lip, jaw, and tongue. In several anatomical studies, portions of the cortical input to the orofacial region of the MI have been examined in the macaque monkey (Muakkassa and Strick, 1979;Godschalk et al., 1984;Luppino et al., 1993;Morecraft andVan Hoesen, 1992, 1993) and in the owl monkey (Stepniewska et al., 1993). However, comprehensive maps of input to the orofacial region of the MI, particularly following physiological verification of the injection sites, have not been published for the macaque. Furthermore, some questions remain concerning the somatotopical relationship between the orofacial and forelimb regions within the various premotor areas (Morecraft andVan Hoesen, 1992, 1993;He et al., 1993He et al., , 1995;;Morecraft et al., 1996).

[4] 109w In the present study, we have addressed these issues by injecting different retrograde tracers into the orofacial and forelimb regions of the MI of the macaque monkey under the guidance of intracortical microstimulation (ICMS). The location and density of retrogradely labeled neurons were reconstructed on unfolded maps of the frontal and parietal lobes. Then, within each cortical area, we compared the cortical distribution of retrogradely labeled neurons that projected directly to each representation in the MI. By using this anatomical approach, we have identified the cortical areas that are potentially involved in orofacial motor behavior and have further defined the somatotopy within the premotor areas of the frontal lobe.

[5] 247w It has been revealed in the present study that the orbital cortex (area 12) projects to the orofacial region of the MI but not to its forelimb region (see also Morecraft and Van Hoesen, 1992). An anatomical study has reported the existence of a direct projection of orbital cortical neurons to the forelimb region of the MI (Leichnetz, 1986). In this study, however, the injection site of a retrograde tracer in the forelimb region of the MI may have infringed upon its orofacial region, because the extent of the site of tracer injection was not assessed electrophysiologically. Although the functional significance of the projection from the orbital cortex to the orofacial region of the MI is still unknown, it is conceivable that such a direct projection may play some role in limbic aspects of orofacial motor behavior, for example, feeding behavior and emotionally guided facial expression. Huang et al. (1989) have shown that, in the macaque monkey, rhythmical jaw movements can be induced by ICMS in the cortical area, termed the CMaAd, which is located in the inner face of the frontal operculum. In the present study, we have found that substantial numbers of cortical neurons that project to the orofacial region of the MI are distributed in the area corresponding to the CMaAd, as previously suggested by Godschalk et al. (1984). Thus, such a direct projection from the CMaAd to the orofacial region of the MI may be involved in the control of rhythmical jaw movements.

[6] 110w In our study, we found that virtually no neurons in the PMd, which is well known to have certain areas representing the forelimb and hind limb parts (Godschalk et al., 1984;Leichnetz, 1986;Kurata, 1989;He et al., 1993;Tokuno and Tanji, 1993), project to the orofacial region of the MI. By contrast, Stepniewska et al. (1993) reported the existence of an orofacial representation in the PMd of the owl monkey by using ICMS with high currents. However, such a physiological finding has not been confirmed anatomically by using a retrograde tracing technique: They detected no labeled neurons in the PMd after injection of a retrograde tracer into the orofacial region of the MI.

[7] 268w The present results have demonstrated that many neurons in the PMv, especially along the inferior limb of the arcuate sulcus, project to the orofacial region of the MI. A recent study on the somatotopical organization of the PMv has shown that ICMS in this area elicits orofacial movements (Godschalk et al., 1996). In previous studies, the PMv has been shown to contain a large area related to forelimb movements, particularly to movements of the distal part of the forelimb, such as digit and wrist movements (Kurata, 1989;Luppino et al., 1991;Tokuno and Tanji, 1993). In addition, by means of single-unit recordings in monkeys trained to perform hind limb tasks (Kurata, 1989), it has been revealed that a small restricted area in the PMv is related to hind limb movements. A retrograde tracing study on the distribution of corticospinal neurons has shown that the neurons in the PMv project to the lumbar segments of the spinal cord (He et al., 1993). Moreover, the existence of corticocortical neurons projecting to the hind limb region of the SMA (Tokuno and Inase, 1994) also suggests the hind limb representation in the PMv. Thus, the PMv can be considered to have a complete set of body part representations (Fig. 9). In the PMv, the orofacial, forelimb, and hind limb parts are represented in this order, from lateral to medial, as in the MI. However, the relative expansion of each body part is quite uneven in the PMv compared with the MI: The PMv contains large areas related to orofacial and forelimb movements and a much smaller but discrete area related to hind limb movements.

[8] 89w A previous study (Mushiake et al., 1991; for review, see Kurata, 1994) suggests the functional difference between the PMd and the PMv: The PMd is involved in the preparation for forthcoming movements and the execution of intended action and conditional motor behavior, whereas the PMv is specialized for visually guided movements. This notion is based mostly on experimental data from monkeys trained to perform forelimb tasks. It is not clear yet whether the orofacial region of the PMv is related specifically to visually guided movements of the orofacial part.

[9] 121w The present study has shown that neurons projecting to the orofacial region of the MI exist consistently in the medial wall of the hemisphere. Such neurons are localized to the electrophysiologically identified orofacial region of the SMA. Our retrograde double-labeling experiments further indicate that SMA neurons projecting to the orofacial region of the MI are distributed in an area different from that containing SMA neurons projecting to the forelimb region of the MI. In agreement with the present data, it has been reported previously that a number of neurons in the SMA are labeled after injecting a retrograde tracer into the orofacial region of the MI of the macaque monkey (Godschalk et al., 1984;Morecraft andVan Hoesen, 1992, 1993;Luppino et al., 1993).

[10] 161w In contrast to these results with the macaque monkey, Stepniewska et al. (1993) have shown the paucity of SMA neurons projecting to the orofacial region of the MI in the owl monkey. In their study on the distribution of corticocortical neurons projecting to the orofacial region of the MI, they have observed that only a few or no neurons are labeled in the SMA in three cases examined after injection of a retrograde tracer into the orofacial region of the MI. With respect to New World monkeys, Welker et al. (1957) have reported that, by using electrical surface stimulation, no orofacial representation is found in the SMA of the squirrel monkey. By using ICMS, Gould et al. (1986) reported that, in the SMA of the owl monkey, there may exist only a small area representing the orofacial part. Thus, the orofacial representation of the SMA in New World monkeys differs considerably in its relative size from that in Old World monkeys.

[11] 174w In the present study, numbers of retrogradely labeled neurons were observed in the precentral gyrus around the sites of tracer injection in the orofacial and forelimb regions of the MI. In a previous study (Huntley and Jones, 1991), the intrinsic connections of MI were examined by injecting WGA-HRP iontophoretically into the forelimb sites that were identified with ICMS mapping. These intrinsic connections were focused within the forelimb region and avoided the orofacial representation of the MI. The present results have confirmed these findings. We observed that the labeled neurons seen after tracer injection into the orofacial region of the MI were located primarily in areas lateral to the injection site, whereas those seen after tracer injection into the forelimb region of the MI were located mostly in areas medial to the injection site (see also Godschalk et al., 1984;Dum and Strick, 1991a). These observations indicate the existence of strong intrinsic connections within each of the orofacial and forelimb regions of the MI and deny the presence of interconnections between these regions of the MI.