Abstract
In nonhuman primates, major input to the striatum originates from ipsilateral cortex and thalamus. The striatum is a target also of crossed corticostriatal (CSt) projections from the contralateral hemisphere, which have been so far somewhat neglected. In the present study, based on neural tracer injections in different parts of the striatum in macaques of either sex, we analyzed and compared qualitatively and quantitatively the distribution of labeled CSt cells in the two hemispheres. The results showed that crossed CSt projections to the caudate and the putamen can be relatively robust (up to 30% of total labeled cells). The origin of the direct and the crossed CSt projections was not symmetrical as the crossed ones originated almost exclusively from motor, prefrontal, and cingulate areas and not from parietal and temporal areas. Furthermore, there were several cases in which the contribution of contralateral areas tended to equal that of the ipsilateral ones. The present study is the first detailed description of this anatomic pathway of the macaque brain and provides the substrate for bilateral distribution of motor, motivational, and cognitive signals for reinforcement learning and selection of actions or action sequences, and for learning compensatory motor strategies after cortical stroke.
SIGNIFICANCE STATEMENT In nonhuman primates the striatum is a target of projections originating from the contralateral hemisphere (crossed CSt projections), which have been so far poorly investigated. The present study analyzed qualitatively and quantitatively in the macaque brain the origin of the crossed CSt projections compared with those originating from the ipsilateral hemisphere. The results showed that crossed CSt projections originate mostly from frontal and rostral cingulate areas and in some cases their contribution tended to equal that from ipsilateral areas. These projections could provide the substrate for bilateral distribution of motor, motivational, and cognitive signals for reinforcement learning and action selection, and for learning compensatory motor strategies after cortical stroke.
Introduction
The identification of all the inputs to the striatum and the way in which they distribute in the various parts of the striatum is an essential aspect for understanding the mode of information processing in the basal ganglia for different motor and nonmotor functions.
The ipsilateral (ipsi) cerebral cortex is certainly the major source of afferents to the striatum. Early studies of these corticostriatal (CSt) projections in nonhuman primates have favored a parallel processing model in which different striatal territories are a target of specific cortical regions and, in turn, are at the origin of largely segregated basal ganglia-thalamo-cortical loops (Alexander et al., 1986).
Subsequent studies have suggested a parallel processing and information convergence model (Nambu, 2011) in which each main basal ganglia-thalamo-cortical loop consists of several largely segregated closed subloops. Each subloop originates from and projects to limited sets of closely related areas and involves distinct, relatively restricted striatal zones, which have been referred to as input channels (Strick et al., 1995; Middleton and Strick, 2000).
More recent studies showed an even more complex pattern of information convergence in the primate striatum. First, there is evidence for overlap in restricted striatal zones of projection fields of distant, interconnected areas jointly involved in specific large-scale functionally specialized cortical networks (Gerbella et al., 2016; Choi et al., 2017). Second, the laminar origin of the CSt projections from a given area can largely vary according to the target striatal zone so that different striatal zones are targets of characteristically weighted laminar projections from the various input areas (Borra et al., 2021).
The striatum is also a target of crossed CSt projections originating from the contralateral (contra) hemisphere. Although noted in the macaque brain since at least 50 years ago (Kemp and Powell, 1970), these projections have been so far somewhat neglected. Indeed, crossed CSt projections in the macaque brain have been reported in several studies after anterograde tracer injections in different cortical areas (Künzle, 1975, 1978; Liles and Updyke, 1985; Huerta and Kaas, 1990; Cavada and Goldman-Rakic, 1991; McGuire et al., 1991; Parthasarathy et al., 1992). However, these studies could not give a comprehensive qualitative and quantitative view of the regional origin of these projections to the various parts of the striatum. To our knowledge, Jones et al. (1977) have been the only ones to describe the origin of the crossed CSt projections based on retrograde tracer injections in the putamen in squirrel monkeys. However, this description was only qualitative and shown for only one subject.
Accordingly, it is still largely unknown which is the effective size of the crossed CSt projections compared with the direct ones, whether crossed CSt projections originate from all the cortical areas that project to the ipsilateral striatum, and whether the origin of crossed and direct CSt projections to a given striatal zone is symmetrical. Thus, it is still an open question whether crossed CSt projections could represent a potentially important variable to consider in defining the pattern of information convergence in the striatum.
To address these issues, in the present study we have placed retrograde tracer injections in different parts of the striatum and compared qualitatively and quantitatively the regional distribution of the labeled CSt cells in the contralateral versus the ipsilateral hemisphere.
Materials and Methods
Subjects, surgical procedures, and selection of the injection sites
The experiments were conducted in four Macaca mulatta (Cases 71 female; 75, 76, and 77 male) in which retrograde neural tracers were injected in different parts of the caudate and putamen. Animal handling as well as surgical and experimental procedures complied with the European law on the humane care and use of laboratory animals (Directives 86/609/EEC, 2003/65/CE, and 2010/63/EU), Italian laws in force regarding the care and use of laboratory animals (DL 116/92 and 26/2014), and were periodically approved by the Veterinarian Animal Care and Use Committee of the University of Parma and authorized by the Italian Ministry of Health.
Before the injection of neural tracers, we obtained scans of each brain using magnetic resonance imaging (MRI; General Electric 7T) to calculate the stereotaxic coordinates of the striatal target regions and the best trajectory of the needle to reach them.
Under general anesthesia (Case 71, Zoletil, initial dose 20 mg/kg, i.m., supplemental 5–7 mg/kg/h, i.m.; Cases 75, 76, and 77, induction with Ketamine 10 mg/kg, i.m., followed by intubation, isoflurane 1.5–2%) and aseptic conditions, each animal was placed in a stereotaxic apparatus, and an incision was made in the scalp. The skull was trephined to remove the bone and the dura was opened to expose a small cortical region. After tracer injections, the dural flap was sutured, the bone was replaced, and the superficial tissues were sutured in layers. During surgery, hydration was maintained with saline, and heart rate, blood pressure, respiratory depth, and body temperature were continuously monitored. On recovery from anesthesia, the animals were returned to their home cages and closely observed. Dexamethasone (0.5 mg/kg, i.m.) and prophylactic broad-spectrum antibiotics (e.g., Ceftriaxone 80 mg/kg, i.m.) were administered preoperatively and postoperatively, as were analgesics (e.g., Ketoprofen 5 mg/kg, i.m.).
Tracer injections and histologic procedures
Based on stereotaxic coordinates, the neural tracers Fast Blue (FB; 3% in distilled water, Dr. Illing Plastics), Diamidino Yellow (DY; 2% in 0.2 m phosphate buffer, pH 7.2, Dr. Illing Plastics), Wheat Germ Agglutinin (WGA; 4% in distilled water, Vector Laboratories), Dextran conjugated with Lucifer yellow (LYD; 10,000 MW, 10% 0.1 m phosphate buffer, pH 7.4, Invitrogen, Thermo Fisher Scientific), and Cholera Toxin B (CTB) subunit, conjugated with Alexa Fluor 488 [CTB green (CTBg); 1% in 0.01 m phosphate-buffered saline, pH 7.4, Invitrogen, Thermo Fisher Scientific] were slowly pressure injected through a stainless steel 31 gauge beveled needle attached through a polyethylene tube to a Hamilton syringe. For all tracer injections, the needle was lowered to the striatum within a guiding tube to avoid tracer spillover in the white matter. Table 1 summarizes the locations of the injections, the injected tracers, and the amounts injected.
Animals used, location of injection sites, and type and amount of injected tracers
After appropriate survival periods following the injections (48 h for WGA, 21–28 d for the other tracers; Table 1), each animal was deeply anesthetized with an overdose of sodium thiopental and perfused through the left cardiac ventricle consecutively with saline (∼2 L in 10 min), 3.5% formaldehyde (5 L in 30 min), and 5% glycerol (3 L in 20 min), all prepared in 0.1 m phosphate buffer, pH 7.4. Each brain was then blocked coronally on a stereotaxic apparatus, removed from the skull, photographed, and placed in 10% buffered glycerol for 3 d and 20% buffered glycerol for 4 d. In Case 75, the right inferotemporal cortex was removed for other experimental purposes. Finally, each brain was cut frozen into coronal sections of 60 µm (Case 75) or 50 µm (Cases 71, 76, and 77) thickness.
In all cases, sections spaced 300 µm apart, that is, one section in each repeating series of six in Cases 71, 76, and 77 and one in series of five in Case 75, were mounted, air dried, and quickly coverslipped for fluorescence microscopy. Other series of sections spaced 300 µm apart were processed for visualizing CTBg (Cases 75, 76, and 77), LYD (Case 75), or WGA (Cases 76 and 77) with immunohistochemistry. Specifically, in all sections endogenous peroxidase activity was eliminated by incubation in a solution of 0.6% hydrogen peroxide and 80% methanol for 15 min at room temperature. For the visualization of CTBg, the sections were then incubated for 72 h at 4°C in a primary antibody solution of rabbit anti-Alexa Fluor 488 (1:15,000; Thermo Fisher Scientific) in 0.5% Triton and 5% normal goat serum in PBS and for 1 h in biotinylated secondary antibody (1:200; Vector Laboratories) in 0.3% Triton and 5% normal goat serum in PBS. For the visualization of LYD, the sections were then incubated for 96 h at 4°C in a primary antibody solution of or rabbit anti-LY (1:3000; Life Technologies, Thermo Fisher Scientific) in 0.5% Triton X-100 and 5% normal goat serum in phosphate buffer (PB) 0.1 m, and for 1 h in biotinylated secondary antibody (1:200; Vector Laboratories) in 0.3% Triton and 5% normal goat serum in PB. For the visualization of the WGA, the sections were incubated overnight at room temperature in a primary antibody solution of goat anti-WGA (1:2000; Vector Laboratories) in 0.3% Triton and 5% normal rabbit serum in PBS and for 1 h in biotinylated secondary antibody (1:200; Vector Laboratories) in 0.3% Triton and 5% normal rabbit serum in PBS. Finally, in all sections the labeling was visualized using the Vectastain ABC Kit (Vector Laboratories) and then a solution of DAB (50 mg/100 ml; Sigma Millipore), 0.01% hydrogen peroxide, 0.02% cobalt chloride, and 0.03% nickel ammonium sulfate in 0.1 m PB.
In all cases, one series of each sixth section (fifth section in Case 75) was stained using the Nissl method (0.1% thionin in 0.1 m acetate buffer, pH 3.7).
Data analysis
Injection sites, distribution of retrogradely labeled neurons, and areal attribution of the labeling
All the injection sites used in this study shown in Figure 1 were completely restricted to the target striatal nucleus (caudate or putamen). The cortical distribution of retrograde labeling in the ipsilateral and contralateral hemispheres was plotted in sections every 600 µm (every 1200 µm in Case 75l LYD) together with the outer and inner cortical borders, using a computer-based charting system. Data from individual sections were then imported into three-dimensional (3D) reconstruction software (Demelio et al., 2001) providing volumetric reconstructions of the monkey brain, including connectional and architectonic data.
Location of the injection sites shown in drawings of coronal sections and in brightfield (for WGA and LYD injections) and epifluorescence (for CTBg, DY, and FB injections) photomicrographs. In the drawings, all injection sites except for WGA are depicted as a black zone corresponding to the core, surrounded by a gray zone, corresponding to the halo. WGA injection sites are depicted as a gray zone because of the poor definition of the core versus the halo. Calibration bars shown for the section drawing and the photomicrograph of Case 75l LYD apply to all section drawings and photomicrographs, respectively. C, Central sulcus; Cd, caudate nucleus; Cg, cingulate sulcus; Cla, claustrum; GP, globus pallidus; IA, inferior arcuate sulcus; L, lateral sulcus; Put, putamen; S, spur of the arcuate sulcus; SA, superior arcuate sulcus; ST, superior temporal sulcus.
The criteria and maps adopted for the areal attribution of the labeling were similar to those adopted in previous studies (Borra et al., 2017; Caminiti et al., 2017). Specifically, prefrontal, frontal, and cingulate motor and opercular frontal areas, where most of the labeling was located, were defined according to cytoarchitectonic and/or chemoarchitectonic criteria described in Matelli et al. (1985, 1991), Carmichael and Price (1994), Gerbella et al. (2007), Belmalih et al. (2009), and Saleem et al. (2014). Some prefrontal and cingulate areas have been considered together and are referred to as 9/8B, 24/32, 24a/b, 24c/d, 23a/b, 23/31, and 29/30. In the inferior parietal lobule, the gyral convexity areas were defined according to cytoarchitectonic and chemoarchitectonic criteria described in Gregoriou et al. (2006) and those of the lateral bank of the intraparietal sulcus based on connectional criteria described in Borra et al. (2008). The superior and medial parietal cortex were defined according to architectonic criteria described in Pandya and Seltzer (1982) and Luppino et al. (2005). For the caudal cingulate cortex, we adopted the cytoachitectonic map proposed by Morecraft et al. (2004). Finally, the temporal labeling in the superior temporal sulcus was attributed based on the electrophysiological and architectonic map proposed by Boussaoud et al. (1990) and guided by the atlas of Saleem and Logothetis (2012), and the labeling in the superior temporal gyrus and auditory belt cortex was based on the architectonic, functional, and connectional map described by Kaas and Hackett (2000; Saleem et al. 2008). For the quantitative analysis, the temporal lobe was subdivided into four regions, the temporal pole (Tp), medial temporal (Tm), superior temporal (Ts), and inferior temporal (Ti).
Quantitative analysis and laminar distribution of the labeling
In all cases, the number of labeled neurons plotted in the ipsilateral and contralateral hemispheres was counted, and the cortical input to the injected striatal zone was then expressed in terms of the percentage of labeled neurons found in each cortical subdivision, with respect to the overall cortical labeling found for each tracer injection in the ipsilateral hemisphere or in both ipsilateral and contralateral hemispheres.
For the areas where the percentage of ipsi plus contra labeling was >1%, the number of labeled cells observed in the contralateral area was subdivided by the total number (ipsi plus contra) of labeled cells to obtain a contralaterality index, which could range from one (all cells in the contralateral area) to zero (all cells in the ipsilateral area).
Results
Crossed CSt projections have been observed in all the cases of the present study. The proportion of labeled cells in the contralateral versus ipsilateral hemisphere largely varied according to the location of the injection site, showing a gradient in which crossed CSt projections were strongest to the caudate head and body, less strong to the rostral putamen and dorsal motor putamen, and lowest to the middle or midventral motor putamen, where the hand is represented (Alexander and DeLong, 1985). In general, the distribution of the labeled CSt cells in the contralateral hemisphere differed for several aspects from that observed in the ipsilateral hemisphere (Fig. 2). Indeed, contralateral areas with CSt projections were always a subset of the ipsilateral areas with CSt projections. Specifically, in all the cases, the areas in the posterior parietal, temporal, and insular cortex that exhibited even relatively robust labeling in the ipsilateral hemisphere had virtually no labeling in the contralateral hemisphere (Fig. 2). Nevertheless, there were areas in the contralateral hemisphere whose relative contribution to the crossed CSt projections was much higher than that of the homolog ipsilateral areas to the direct CSt projections. The frequency distributions of the labeled CSt cells per area in the ipsilateral and the contralateral hemisphere were compared by the Pearson's chi-square test and likelihood ratio chi-square test (Fig. 2). Both test showed a statistically significant difference (p < 0.001) among the distributions in all the cases.
Percentage of areal distribution of the total retrograde labeling observed in the ipsilateral hemisphere (black) compared with that of the total retrograde labeling observed in the contralateral hemisphere (gray) in all the cases of the present study. In each graph, areas are ordered based on the amount of ipsilateral labeling (only areas with ipsilateral labeling >1%). Superior parietal (SPL) areas and inferior parietal (IPL) areas are grouped. The results of the statistical analysis are reported in which the frequency distributions of the labeled CSt cells per area in the ipsilateral and the contralateral hemisphere were compared (Pearson chi-square test for independence). cCg, Caudal cingulate cortex (areas 23, 31, 29, 30); DF, degree of freedom; Ia, agranular insula; Idg, disgranular insula.
CSt projections to the caudate nucleus
Two tracer injections were placed in the caudate head. In Case 76 right (76r) the WGA injection site was slightly more rostral and medial than the LYD injection site of Case 75 left (75l; Fig. 1). As largely expected, in the ipsilateral hemisphere the highest proportion of labeled cells was located in the prefrontal cortex (Figs. 3, 4, Table 2), mostly involving areas 9/8B, followed by the rostral cingulate cortex, mostly involving the rostral area 24/32 and the rostral cingulate gyrus (area 24a/b). Relatively robust labeling was also observed in the temporal cortex involving in both cases belt/parabelt auditory areas, the superior temporal polysensory area, and the medial temporal cortex and in Case 75l LYD, the rostral inferotemporal cortex. Much weaker was the labeling observed in rostral premotor, insular, and caudal cingulate cortex and in Case 75l LYD, in the parietal cortex. In both cases, the percentage of labeled CSt cells observed in the contralateral hemisphere was robust, i.e., ∼30% of the total number of ipsilateral plus contralateral CSt cells (Figs. 3, 4, Table 3).
Regional distribution (%) and total number (n) of labeled cortical neurons observed in the ipsilateral hemisphere, following tracer injections in the caudate and in the putamen
Regional distribution (%) and total number (n) of labeled cortical neurons observed in both ipsilateral and contralateral hemisphere
Distribution of the retrograde labeling observed after injection of WGA in the head of the caudate in Case 76r. The labeling is shown in dorsolateral and medial views of the 3D reconstructions of the injected (ipsilateral) and the contralateral hemisphere and in drawings of coronal sections. For the sake of comparison, in this and in the subsequent figures, all the 3D reconstructions are shown as a right hemisphere with the injected hemisphere on the left and all drawings with the injected hemisphere on the right. a–f, Sections are shown in a rostral to caudal order. The levels at which the sections were taken are shown on the 3D reconstructions of both hemispheres. Each dot corresponds to one labeled neuron. C, Central sulcus; Ca, calcarine fissure; Cd, caudate nucleus; Cg, cingulate sulcus; GP, globus pallidus; IA, inferior arcuate sulcus; Ins, insula; IP, intraparietal sulcus; L, lateral sulcus; LO, lateral orbital sulcus; Lu, lunate sulcus; MO, medial orbital sulcus; Opt, occipito-temporo-parietal area; P, principal sulcus; Put, putamen; SA, superior arcuate sulcus; ST, superior temporal sulcus; Th, thalamus.
Distribution of the retrograde labeling observed after injection of LYD in the head of the caudate in Case 75l. The labeling is shown in dorsolateral and medial views of the 3D reconstructions of the injected (ipsilateral) and contralateral hemispheres and in drawings of coronal sections. C, Central sulcus; Ca, calcarine fissure; Cd, caudate nucleus; Cg, cingulate sulcus; GP, globus pallidus; IA, inferior arcuate sulcus; Ins, insula; IP, intraparietal sulcus; L, lateral sulcus; LO, lateral orbital sulcus; Lu, lunate sulcus; MO, medial orbital sulcus; P, principal sulcus; ParOp, parietal operculum; Put, putamen; SA, superior arcuate sulcus; ST, superior temporal sulcus; Th, thalamus. Conventions are defined in the legend to Figure 3.
In the contralateral hemisphere virtually all the labeling was located in frontal and cingulate areas, whereas the parietal, temporal, and insular cortex were virtually devoid of labeling (Table 3). In both cases the labeling in the contralateral hemisphere was mostly concentrated in the rostral cingulate areas 24/32 and in prefrontal areas 9/8B, which were the most densely labeled areas in the ipsilateral hemisphere (Fig. 5). In both cases these two contralateral areas were among the five most labeled ones. However, in contralateral area 10 in Case 76r WGA and in the orbital (12o) and lateral (12l) part of area 12 in Case 75l LYD the labeling was relatively much weaker compared with the ipsilateral hemisphere. Weaker labeling also involved rostral premotor areas and the caudal cingulate cortex.
Percentage distribution of the total (ipsi plus contra) retrograde labeling in ipsilateral (black) and contralateral (gray) areas observed after the tracer injections in the caudate. The asterisks indicate the five most labeled areas. cCg, Caudal cingulate cortex (areas 23, 31, 29, 30); Ia, agranular insula; Idg, disgranular insula.
In Case 75r DY, the injection site was located in the caudate body at about the level of the anterior commissure (AC; Fig. 1). The distribution of the labeling in the ipsilateral hemisphere (Fig. 6, Table 2) was markedly different from that observed in the two cases previously described, involving mostly dorsal and medial premotor areas, primarily areas F3, F2, and F6 and to a lesser extent area F7. In area F3, the labeling mostly involved its rostral half. Less dense labeling was located in the ventral premotor cortex, motor cingulate (24c/d), inferior parietal (anterior intraparietal, PFG, PG) and caudal superior parietal (caudal PE (PEc), cingulate PE (PEci), medial PG (PGm), V6A) areas. This labeling distribution suggests that the injection site involved a striatal sector related to visuomotor and somatomotor control of arm movements. Also in this case, the labeling in the contralateral hemisphere was quite robust (∼30% of the ipsilateral plus contralateral CSt labeled cells) and was virtually all located in frontal and cingulate areas (Table 3). Specifically, the areal distribution of the labeled CSt cells (Fig. 5) in all the various premotor and cingulate areas was quite similar in the two hemispheres, and contralateral areas F2 and F3 were among the five most labeled areas, considering both the ipsilateral and contralateral hemispheres.
Distribution of the retrograde labeling observed after injection of DY in the caudate body in Case 75r. The labeling is shown in dorsolateral and medial views of the 3D reconstructions of the injected (ipsilateral) and contralateral hemisphere and in drawings of coronal sections. AIP, anterior intraparietal area; C, Central sulcus; Ca, calcarine fissure; Cd, caudate nucleus; Cg, cingulate sulcus; IA, inferior arcuate sulcus; IP, intraparietal sulcus; L, lateral sulcus; LO, lateral orbital sulcus; Lu, lunate sulcus; MST, medial superior temporal area; MO, medial orbital sulcus; P, principal sulcus; PO, parieto occipital sulcus; Put, putamen; S, spur of the arcuate sulcus; SA, superior arcuate sulcus; ST, superior temporal sulcus. Conventions as in Figure 3.
CSt projections to the putamen
Five tracer injections involved the putamen, one more rostrally (Case 77r WGA) and the others at the same level of or caudal to the AC, at different dorsoventral levels (Fig. 1).
The injection site in Case 77r WGA was located relatively ventrally in the precommissural putamen. In the ipsilateral hemisphere (Fig. 7), the labeling was densest in the premotor cortex, especially area F5, frontal operculum, and rostral cingulate cortex, especially area 24c/d. Less dense labeling involved prefrontal, parietal, and insular cortex and, more weakly, temporal and caudal cingulate cortex. The very weak labeling in area F1 suggests that the injection site involved a putaminal sector rostral to the motor putamen as usually defined. The distribution of the labeling involving the ventral premotor area F5, frontal operculum, rostral F3, ventrolateral prefrontal sectors including area 12 and ventral area 46, and inferior parietal and opercular parietal sectors suggests that the injection site involved a striatal sector related to hand and mouth motor control, at least in part corresponding to the rostral striatal target of the lateral grasping network projections (Gerbella et al., 2016). In the contralateral hemisphere, the amount of CSt labeled neurons was relatively robust, although lower than that observed for the caudate tracer injections (Table 3). Specifically, labeled CSt cells were mostly equally distributed between the rostral cingulate and premotor cortex and only marginally involved prefrontal and caudal cingulate cortex. Again, parietal and temporal cortex, but also the insula, were virtually devoid of labeling. The areal distribution of the labeling (Fig. 8) shows that the ratio of labeled CSt cells in the contralateral versus ipsilateral cortical areas was quite low for F5 and frontal operculum and relatively high for the rostral cingulate areas and premotor areas F6 and F7. Contralateral area 24c/d was among the five most labeled areas, considering ipsilateral and contralateral areas.
Distribution of the retrograde labeling observed after injection of WGA in the precommissural putamen in Case 77r. The labeling is shown in dorsolateral and medial views of the 3D reconstructions of the injected (ipsilateral) and contralateral hemisphere and in drawings of coronal sections. AIP, anterior intraparietal area; C, central sulcus; Ca, calcarine fissure; Cd, caudate nucleus; Cg, cingulate sulcus; GP, globus pallidus; IA, inferior arcuate sulcus; Ins, insula; IP, intraparietal sulcus; L, lateral sulcus; LO, lateral orbital sulcus; Lu, lunate sulcus; MO, medial orbital sulcus; P, principal sulcus; ParOp, parietal operculum; Put, putamen; SA, superior arcuate sulcus; S, spur of the arcuate sulcus; ST, superior temporal sulcus; Th, thalamus. Conventions as in Figure 3.
Percentage distribution of the total (ipsi plus contra) retrograde labeling in the ipsilateral (black) and contralateral (gray) areas observed after the tracer injections in the putamen. The asterisks indicate the five most labeled areas. AIP, anterior intraparietal area; cCg, caudal cingulate cortex (areas 23, 31, 29, 30); Ia, agranular insula; Idg, disgranular insula; ParOp, parietal operculum.
In Case 75r CTBg, the injection site was placed in the dorsal part of the putamen at about the level of the AC (Fig. 1). In the ipsilateral hemisphere (Fig. 9), labeled CSt cells were mostly concentrated in the frontal and cingulate motor cortex. As expected from the somatotopy of the motor putamen (Alexander and DeLong, 1985), in the frontal motor cortex labeled cells were densest in the medial part of F1 and in the caudal half of F3 suggesting that the injection site involved the leg/trunk representation. Additional weaker labeling was observed in dorsomedial primary somatosensory area (SI), in the superior parietal area PE, and in the caudal cingulate cortex. In the contralateral hemisphere, labeled CSt cells were ∼22% of the total ipsilateral plus contralateral labeling and were observed almost completely in the frontal motor cortex, mainly in F1, F3, and F2, and in area 24c/d, except for some weak labeling in the parietal cortex (mostly in SI) and in the caudal cingulate cortex. As observed in other cases, of the five most labeled cortical areas, two were in the contralateral hemisphere (Fig. 8). Furthermore, based on the number of labeled cells, about one-third of the overall input to the injected striatal zone from areas 24c/d, F3, and F2 originated from the contralateral hemisphere (nearly one-fifth from contralateral F1).
Distribution of the retrograde labeling observed after injection of CTBg in the dorsal part of the motor putamen in Case 75r. The labeling is shown in dorsolateral and medial views of the 3D reconstructions of the injected (ipsilateral) and contralateral hemisphere and in drawings of coronal sections. C, central sulcus; Ca, calcarine fissure; Cd, caudate nucleus; Cg, cingulate sulcus; GP, globus pallidus; IA, inferior arcuate sulcus; Ins, insula; IP, intraparietal sulcus; L, lateral sulcus; LO, lateral orbital sulcus; Lu, lunate sulcus; MO, medial orbital sulcus; MST, medial superior temporal area; P, principal sulcus; ParOp, parietal operculum; PO, parieto occipital sulcus; Put, putamen; S, spur of the arcuate sulcus; SA, superior arcuate sulcus; ST, superior temporal sulcus; Th, thalamus. Conventions as in Figure 3.
Three tracer injections were placed more ventrally in the motor putamen at different rostrocaudal levels (Fig. 1). In Case 71r DY the injection site was located in the middle of the motor putamen and was much smaller than expected, likely because the volume of tracer injected was smaller than planned. In Cases 77l CTBg and 71l FB, the injections were located slightly more caudal and ventral. In all these cases, the labeling in the ipsilateral hemisphere was mostly located in the frontal and cingulate motor cortex, but ∼15–20% of the labeled cells were in the parietal cortex (Table 2). As expected from the somatotopy of the motor putamen (Alexander and DeLong, 1985), the labeling involved in all cases the middle part of F1 and F3 and ventral premotor areas F5 and F4, suggesting involvement by the injection sites of the arm/hand representation (Figs. 8, 10, 11). In Cases 77l CTBg and 71l FB the labeling extended also more ventrally in F1 and in F5, more rostrally in F3 and in the frontal operculum, suggesting an involvement also of the face/mouth representation. In the parietal cortex the most labeled areas were SI, secondary somatosensory area (SII), and PE and rostral inferior parietal areas. In the contralateral hemisphere, the number of labeled cells (5–12%) was considerably lower than that observed after the other putaminal and the caudate injections (Table 3). However, this labeling involved a limited number of contralateral areas (Fig. 2), so the relative contribution of the crossed projections to the whole input from areas 24c/d, F3, and F1 was ∼10–20%. Interestingly, in Cases 71l FB and 77l CTBg the contralateral labeling tended to be densest in rostral F3 and in Case 71l FB in the lateral part of the ventral premotor cortex, likely involving preferentially face/mouth fields. Furthermore, except for some sparse labeling located in SI (also SII in Case 71r DY), the contralateral parietal labeling was negligible, even if in these three cases the ipsilateral parietal cortex hosted 13–19% of the overall labeled cells.
Distribution of the retrograde labeling observed after injection of CTBg in Case 77l, DY in Case 71r, and FB in 71l in the midventral motor putamen, shown in dorsolateral and medial views of the 3D reconstructions of the injected (ipsilateral) and contralateral hemispheres. On the 3D reconstructions of both hemispheres of all cases are indicated the levels at which the sections, shown in Figure 11, were taken. C, Central sulcus; Cg, cingulate sulcus; IA, inferior arcuate sulcus; IP, intraparietal sulcus; L, lateral sulcus; Lu, lunate sulcus; P, principal sulcus; SA, superior arcuate sulcus; ST, superior temporal sulcus. Conventions as in Figure 3.
Distribution of the retrograde labeling observed after injection of CTBg in Case 77l, DY in Case 71r, and FB in 71l in the midventral motor putamen, shown in drawings of coronal sections taken at levels indicated in Figure 10. AIP, anterior intraparietal area; C, central sulcus; Cd, caudate nucleus; Cg, cingulate sulcus; GP, globus pallidus; IA, inferior arcuate sulcus; Ins, insula; IP, intraparietal sulcus; L, lateral sulcus; ParOp, parietal operculum; Put, putamen; S, spur of the arcuate sulcus; SA, superior arcuate sulcus; ST, superior temporal sulcus; Th, thalamus. Conventions as in Figure 3.
Laterality of CSt projections
Our data showed that the relative balance of ipsilateral and contralateral projections from frontal and cingulate areas varied widely. Accordingly, we examined whether the proportion of contralateral CSt projections (i.e., contralaterality index equals contralateral cells in area X/ipsi plus contra cells in area X) in each cortical area was related to the total amount of ipsi plus contralateral labeling in that area (Fig. 12). This analysis did not show a correlation between the contralaterality index and total labeling across cortical areas (r = 0.27). Rather, the cortical areas with the largest projections to a striatal injection site could have either high (>0.3, area 24/32 after caudate head injections) or low contralaterality indexes (F5 and F1 after midventral motor putamen injections). Similarly, other areas [F6, F7, and frontal operculum (FrOp)] with moderate projections (<5%) also exhibited a relatively high contralaterality index in some cases and a low contralaterality index in other cases. Furthermore, in several cases, the contralaterality index of the CSt projections appears to vary according to the target striatal zone. For example, area 24c/d exhibited a relatively high contralaterality index for the projections to the caudate and a lower one for the projections to the putamen, independently from the strength of its projections. Among the frontal motor areas, F3 showed a relatively high contralaterality index for the projections to the caudate body and dorsal motor putamen and a lower one for the projections to the midventral motor putamen, again independently from the strength of its projections. Finally, F1 had a higher contralaterality index for its projections to the leg/trunk-related motor putamen and a quite low one for the projections to the hand-related motor putamen.
Contralaterality [contralateral cells in area X/(ipsi + contra cells in area X)] of CSt projections from rostral cingulate and prefrontal (A), rostral premotor (B), and caudal premotor and primary motor (C) areas shown in relation to the richness (percentage of ipsilateral plus contralateral CSt cells) of the labeling. Only areas in which the total (ipsi plus contra) labeling was >1% are considered in the graphs. Values from injections in the caudate are shown with dots, those from injections in rostral and dorsal motor putamen with diamonds, and those from injections in midventral motor putamen with stars. Different colors identify different areas.
Discussion
The present study provides a detailed description of the origin and strength of striatal input from the ipsilateral and contralateral (crossed CSt projections) hemispheres in the nonhuman primate. As has been noted over the last 50 years, striatal input is dominated by the ipsilateral cortex, but our results now demonstrate that a substantial amount originates from the contralateral cortex. Whereas ipsilateral projections originate from cingulate, frontal, parietal, insular, and temporal cortex, crossed CSt projections originate almost exclusively from cingulate, prefrontal, and frontal motor areas. In some cases, the contribution of contralateral cingulate and frontal areas is quite high and even equal to that from the same area of the ipsilateral hemisphere. The crossed CSt projections from the primary motor cortex tend to be relatively robust for the leg/trunk representation and weak for the hand representation. Overall, the distribution of crossed CSt projections suggests that they may provide a substrate for bilateral integration of motor, motivational, and cognitive signals during behavior.
Accordingly, the present study highlights that crossed CSt projections could be an important variable to consider in defining the pattern of information convergence in the striatum.
Crossed CSt projections
Crossed CSt projections have been first noted in the rabbit, cat, and rat (Carman et al., 1965) and then in the macaque (Kemp and Powell, 1970; Fallon and Ziegler, 1979) based on degenerative changes in the striatum after cortical lesions. Crossed CSt projections have been then described in the macaque after neural tracer injections in the primary motor and premotor (Künzle, 1975, 1978; Liles and Updyke, 1985; Huerta and Kaas, 1990; Parthasarathy et al., 1992), prefrontal (McGuire et al., 1991), and parietal (Cavada and Goldman-Rakic, 1991) cortex. These studies, based on anterograde tracer injections at the cortical level, could not give a comprehensive picture of the origin of the crossed CSt projections to a given striatal sector and an estimate of the weight of these projections compared with the direct ones. To our knowledge, only Jones et al. (1977) provided a qualitative description of crossed CSt projections based on injections of a retrograde tracer in the motor putamen in the squirrel monkey. As in Jones et al. (1977), after injections in the motor putamen we found an almost symmetrical distribution of labeled CSt cells in the ipsilateral versus contralateral frontal cortex and a virtual absence of labeled CSt cells in the contralateral parietal cortex. The very poor or even absent labeling observed in the contralateral parietal cortex in all the cases of the present study, including those in which labeling in the ipsilateral parietal cortex was quite robust, is in apparent discrepancy with the observations of Cavada and Goldman-Rakic (1991) after very large tracer injections in different parietal areas. It is possible that CSt parietal cells are relatively few and sparsely project to the striatum, so they are very difficult to be labeled after restricted injections of retrograde tracers in the striatum. On the other hand, as in the present study, Griggs et al. (2017) showed contralateral labeling in the frontal and cingulate, but not in the parietal, temporal, and insular cortex after tracer injections in the caudate head or tail.
The present observations that the crossed CSt projections in the macaque originate predominantly from frontal and cingulate areas are in good agreement with the results of McGuire et al. (1991) and Innocenti et al. (2017), based on tracer injections at the cortical level. Specifically, McGuire et al. (1991) noted that some areas (e.g., area 46) display relatively weak crossed projections, whereas other areas (e.g., F3-supplementary motor area proper) appear to project almost equally to both ipsilateral and contralateral striatum. In this context, it could be worth noting that in the present data there were some areas that were always sources of relatively weak additional projections, in most cases showing a relatively high contralaterality index. It is possible that in our cases the target striatal zones of these areas have been only marginally involved. However, it could be also that some areas tend to have sparse and diffuse bilateral CSt projections.
Crossed CSt projections may also be present in the human brain. Based on tractographic data, Innocenti et al. (2017) found that as in macaques, crossed CSt projections in humans originate predominantly from frontal areas but also from parietal regions of the superior and inferior parietal lobule and from the superior temporal gyrus. The presence in humans of crossed CSt projections from specific posterior parietal regions is supported by functional connectivity data (Jarbo and Verstynen, 2015). Neurodevelopmental studies showed that in mice soon after birth, crossed CSt projections originate from almost the entire hemisphere, whereas after 2 weeks they originate mainly from frontal and cingulate areas possibly because of either retraction of initial collaterals to the contralateral striatum or death of an early developmental population of CSt neurons (Sohur et al., 2014). Accordingly, the presence of crossed CSt projections from parietal and temporal areas in humans, but not in macaques, could be accounted for by a differential maturation of these projections across different species.
Functional considerations
Early models of CSt projections in nonhuman primates have favored a modular organization of the striatum in which different striatal zones are targets of specific cortical regions (Alexander et al., 1986) and, in turn, are at the origin of largely segregated basal ganglia-thalamo-cortical loops (Middleton and Strick, 2000; Kelly and Strick, 2004). Subsequent studies have revealed a higher level of complexity of the corticostriatal projections topography. First, the cortical input to a specific striatal zone originates not only from a limited set of closely related neighbor areas, as initially described (Takada et al., 1998; Nambu, 2011; Averbeck et al., 2014) but also from distant, interconnected areas jointly involved in large-scale functionally specialized cortical networks (Gerbella et al., 2016; Choi et al., 2017). Second, the different striatal zones are targets of characteristically weighted laminar projections from multiple input areas (Griggs et al., 2017; Borra et al., 2021). Finally, the present data suggest that information processing in the striatum can also rely on substantial input from the contralateral hemisphere.
As crossed CSt projections in the macaque brain have been so far poorly considered, their possible contribution to information processing in the striatum still remains to be defined. These projections could indeed provide a substrate for interhemispheric transfer of signals in parallel to the callosal connectivity. Based on conduction delay estimations, it has been suggested that crossed CSt projections mediate a transfer of information considerably faster than through the callosal connectivity (Innocenti et al., 2017). Furthermore, whereas callosal projections connect almost entirely the two hemispheres (Innocenti et al., 2022), crossed CSt projections originates mostly from frontal and cingulate areas, suggesting a role more focused on motor control and/or executive functions. In this context, crossed CSt projections could provide the substrate for bilateral diffusion of motor, motivational, and cognitive signals necessary for reinforcement learning and selection of those actions or action sequences that are most appropriate for achieving a behavioral goal (Averbeck and O'Doherty, 2022) and for inhibitory control of impulsive motor behavior (Oguchi et al., 2021). Furthermore, in the motor putamen, crossed CSt projections could have a role in controlling actions or action sequences involving both sides of the body.
Indeed, brain imaging evidence in humans showed bilateral activation in the motor putamen during the execution of unilateral simple foot, hand, and mouth movements (Gerardin et al., 2003). Interestingly, in the left putamen the activation foci for the right- or left-hand movements appeared largely segregated (Gerardin et al., 2003), resembling the segregation between the terminal fields of direct and crossed CSt projections to the striatal hand representation in the squirrel monkey (Flaherty and Graybiel, 1993). Furthermore, bilateral activation of the striatum has been observed during the execution of right-hand finger-tapping tasks with increasing degrees of complexity (Lehéricy et al., 2006; Bednark et al., 2015). The observed striatal activation foci tended to shift progressively more rostrally with the increase in complexity or frequency of the task execution (Lehéricy et al., 2006). Finally, bilateral striatal activation has been observed during early phases of visuomotor adaptation of arm movements (Seidler et al., 2006) and especially in the anterior striatum during the encoding of novel working memory items (Geiger et al., 2018).
According to Wymbs et al. (2012) bilateral putamen activity is necessary for the strengthening of motor–motor associations at the basis of action chunking processes. Chunking in motor sequencing is considered a key function of the basal ganglia and allows groups of individual movements to be prepared and executed as a single motor program facilitating learning and performance of complex sequences (Halford et al., 1998).
Crossed CSt projections could play a potentially important role in behavioral compensation after brain lesions. Specifically, these projections could have a role in compensatory relearning of motor strategies in the context of the reorganization mechanisms of the motor system occurring after cortical stroke (Balbinot and Schuch, 2019). In support of this proposal, there is evidence in rodents for axonal sprouting of crossed CSt projections and neurochemical signs of increased cell activity in the denervated striatum after sensorimotor cortex lesions (Napieralski et al., 1996; Cheng et al., 1998; Uryu et al., 2001). The presence in macaques of crossed CSt projections that appear even stronger than in rodents is an incentive for experimental studies in nonhuman primates and clinical studies in neurologic patients focused on the role of these projections in compensatory mechanisms after stroke.
Footnotes
This work was supported by the University of Parma and cosponsored by Fondazione Cariparma Grant Fondo di Interesse Locale (FIL) 2019–Quota Incentivante (G.L.), Ministero dell'Istruzione, dell'Università e della Ricerca Grant Progetti di Rilevante Interesse Nazionale (PRIN) 2017: 2017KZNZLN_002 (E.B.). The “Centro di Ricerca, Sviluppo e Studi Superiori in Sardegna” (CRS4) Pula, Cagliari, Italy, developed the 3D reconstruction software. The authors thank Giuseppe Pedrazzi for advice on statistical analysis.
The authors declare no competing financial interests.
- Correspondence should be addressed to Elena Borra at elena.borra{at}unipr.it
