Cerebral Cortex Advance Access originally published online on November 23, 2005
Cerebral Cortex 2006 16(10):1389-1417; doi:10.1093/cercor/bhj076
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Cortical Connections of the Inferior Parietal Cortical Convexity of the Macaque Monkey
Dipartimento di Neuroscienze, Sezione di Fisiologia, Università di Parma, I-43100 Parma, Italy, 1 Current address: Department of Pharmacology and Physiology, School of Medicine and Dentistry, University of Rochester, Rochester, NY, USA, 2 Current address: Laboratory of Neuropsychology, NIMH, NIH, Bethesda, MD, USA
Address correspondence to Prof. Giuseppe Luppino, Dipartimento di Neuroscienze, Sezione di Fisiologia, Università di Parma, Via Volturno 39, I-43100 Parma, Italy. Email: luppino{at}unipr.it.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
We traced the cortical connections of the 4 cytoarchitectonic fields—Opt, PG, PFG, PF—forming the cortical convexity of the macaque inferior parietal lobule (IPL). Each of these fields displayed markedly distinct sets of connections. Although Opt and PG are both targets of dorsal visual stream and temporal visual areas, PG is also target of somatosensory and auditory areas. Primary parietal and frontal connections of Opt include area PGm and eye-related areas. In contrast, major parietal and frontal connections of PG include IPL, caudal superior parietal lobule (SPL), and agranular frontal arm-related areas. PFG is target of somatosensory areas and also of the medial superior temporal area (MST) and temporal visual areas and is connected with IPL, rostral SPL, and ventral premotor arm- and face-related areas. Finally, PF is primarily connected with somatosensory areas and with parietal and frontal face- and arm-related areas. The present data challenge the bipartite subdivision of the IPL convexity into a caudal and a rostral area (7a and 7b, respectively) and provide a new anatomical frame of reference of the macaque IPL convexity that advances our present knowledge on the functional organization of this cortical sector, giving new insight into its possible role in space perception and motor control.
Key Words: area 7a • area 7b • dorsal visual stream • space coding • visuomotor transformations
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
The posterior parietal cortex of the macaque contains a multiplicity of areas involved in the analysis of visual information necessary for motor planning and execution of eye, limb, and body movements (see, e.g., Rizzolatti and others 1997
; Colby 1998).
The rich parietofrontal connections of these areas mediate the transformation of visual information into action, and a series of parietofrontal circuits has been so far identified, linking visually related areas of the caudal superior parietal lobule (SPL) and of the intraparietal sulcus (IPS) with different sectors of the agranular frontal cortex or with the frontal eye fields. These circuits are involved in the visual guidance of reaching, grasping, or eye movements (Colby 1998; Rizzolatti and others 1998).
Within this general framework, there are still several aspects of the anatomical organization of the cortical convexity of the inferior parietal lobule (IPL) and its possible role in visuomotor transformations and/or space coding that need to be elucidated.
This cortical sector is usually subdivided according to the architectonic studies of Vogt O and Vogt C (1919) into a caudal and a rostral area, 7a and 7b, respectively, considered as functional and hodological different entities (see, e.g., Andersen and others 1997; Siegel and Read 1997a). According to this view, 7a is a visually responsive area, located at the vertex of the occipitoparietal visual information flow (dorsal visual stream, Ungerleider and Mishkin 1982), linked with oculomotor area lateral intraparietal area (LIP) and the rostral prearcuate cortex and where retinal and extraretinal signals are combined to construct a representation of space. In contrast, 7b is mostly related to the analysis of somatosensory information, connected with the ventral premotor cortex, and involved in the control of arm and face movements.
Area 7a, however, is also involved in the control of arm-reaching movements (Mountcastle and others 1975; Blum 1985; MacKay 1992; Battaglia-Mayer and others 2005), and according to Hyvärinen (1981) there is a functional segregation in this area between a more rostral, visually and somatosensory responsive, arm field and a more caudal field, in which eye movement signals predominate. Furthermore, in the rostral IPL convexity (area 7b) there is a visual and somatosensory responsive arm/hand and face field (Hyvärinen 1981; Ferrari and others 2003), where visual neurons appear to be involved in higher order visuomotor processings (Gallese and others 2002; Yokochi and others 2003; Fogassi and others 2005). These data, therefore, suggest, first, that area 7a is not homogeneous and, second, that 7b is not exclusively involved in somatomotor functions.
In their architectonic study, Pandya and Seltzer (1982) indeed suggested that the IPL convexity contains at least 3 distinct areas: a rostral, an intermediate, and a caudal one, defined as PF, PG and Opt, respectively, plus a transitional area located between areas PF and PG and named PFG. Accordingly, areas 7a and 7b are both cytoarchitectonically not homogeneous and, in particular, area 7a would consist of at least 2 areas, PG and Opt. This subdivision, however, was never validated by clear connectional and/or functional data, and it is common practice in the literature to refer to areas 7a and PG as synonyms (Siegel and Read 1997a).
In the present study we used cytoarchitectonic data to guide the location of neural tracer injections to study the cortical connections of the IPL convexity. Specific aims were 1) to examine whether patterns of connections validate the subdivision of this sector into more than 2 distinct areas, 2) to identify all the possible sources of sensory information to each of these areas, and 3) to trace their projections to the frontal lobe, where there are multiple representations of different effectors (see, e.g., Rizzolatti and others 1998; Rizzolatti and Luppino 2001) and identify all the several possible parietofrontal circuits involving the IPL convexity and their possible role in space representation and motor control. The results provide strong support for a subdivision of the IPL convexity into 4 distinct areas, referred, in agreement with Pandya and Seltzer (1982), to as PF, PFG, PG, and Opt.
Preliminary data have been presented in abstract form (Luppino, Belmalih, and others 2004).
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
The experiments were carried out on 6 macaque monkeys (3 Macaca nemestrina and 3 Macaca fascicularis) in which neural tracers were injected in cytoarchitectonic fields PF, PFG, PG, and Opt. Additional data from 2 M. nemestrina, in which retrograde tracers were injected in the lateral funiculus of the spinal cord, were used for the definition of the corticospinal projections from the IPL. The brains of 5 additional monkeys (4 M. nemestrina and 1 M. fascicularis, 8 hemispheres), 2 of them used in tracing experiments not related to the present one, were used for preliminary cytoarchitectonic analysis of the IPL convexity.
All experimental procedures were approved by the Veterinarian Animal Care and Use Committee of the University of Parma and complied with the European law on the care and use of laboratory animals.
Surgical Procedures and Tracers Injections
Each animal was anasthetized with ketamine hydrochloride (15 mg/kg intramuscularly) and placed in a stereotaxic apparatus.
In all animals in which tracers were injected in the IPL areas, under aseptic conditions, an incision was made in the scalp, the skull was trephined over the target region, and the dura was opened to expose the IPL convexity. Injection sites were chosen by using cytoarchitectonic data as frame of reference, referred in terms of stereotaxic coordinates and location of anatomical landmarks such as the IPS, the lateral fissure (LF), and the superior temporal sulcus (STS).
Once the appropriate site was chosen, fluorescent tracers (Fast Blue [FB] 3% in distilled water, Diamidino Yellow [DY] 2% in 0.2 M phosphate buffer at pH 7.2, True Blue [TB] 5% in distilled water, EMS-POLYLOY GmbH, Gross-Umstadt, Germany), wheat germ agglutinin–horseradish peroxidase conjugated (WGA-HRP, 4% in distilled water, SIGMA, St. Louis, Missouri), biotinilated dextran amine (BDA, 10% phosphate buffer 0.1 M, pH 7.4; Molecular Probes, Eugene, Oregon), and cholera toxin B subunit, gold conjugated (CTB-g, 0.5% in distilled water, LIST, Campbell, California) or conjugated with Alexa 488, Alexa 555, or Alexa 594 (CTB-A, 1% in phosphate-buffered saline, Molecular Probes) were slowly pressure injected at about 1.2–1.5 mm below the cortical surface as described in detail in previous studies (e.g., Luppino and others 2003). Table 1 summarizes the locations of injections, the injected tracers, and their amounts. After the injection, the dural flap was sutured, the bone replaced, and the superficial tissues sutured in layers.
|
In the 2 animals in which tracers were injected in the spinal cord, under aseptic conditions, following a laminectomy, the dura was opened and the segment of the spinal cord selected for the injection exposed. Retrograde tracers were, then, pressure injected with a 5 µL Hamilton microsyringe in the left lateral funiculus. In 1 animal (Case 10), DY (2%, 8 injections, total amount 12 µL) was injected at the T6 spinal level and 26 days later (HRP, 30% in 2% lysolecithin, SIGMA, 6 injections, total amount 10 µL) at the C4–C5 spinal level. In the second animal (Case 21), HRP was injected at the C3–C5 level. Upon the completion of the injections, the spinal cord was covered with Gelfoam and wounds were closed in layers.
During surgeries, hydration was maintained with saline (about 10 cc/h, intravenously) and temperature with a heating pad. Heart rate, blood pressure, respiratory depth, and body temperature were continuously monitored. Upon recovery from anesthesia, the animals were returned to their home cage and closely monitored.
Histological Procedures
After appropriate survival periods following cortical (28 days for BDA, 12–14 days for fluorescent tracers and CTB-A, 7 days for CTB-g and 2 days for WGA-HRP) or spinal cord (29 days for DY and 3 days for HRP) injections, each animal was anesthetized with ketamine hydrochloride (15 mg/kg intramuscularly) followed by an intravenous lethal injection of sodium thiopental and perfused through the left cardiac ventricle with saline, 3.5–4% paraformaldehyde, and 5% glycerol in this order. All solutions were prepared in phosphate buffer 0.1 M, 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 days and 20% buffered glycerol for 4 days. Finally, it was cut frozen in coronal sections 60 µm thick. In Cases 10 and 21 (spinal cord injections) the spinal cord was removed and, after cryoprotection, cut transversally at 60 µm.
In Cases 27 and 29, 1 section of 5 was mounted, air-dried, and quickly coverslipped for fluorescence microscopy. In Cases 13, 20, and 23, 1 section of 5 was processed for WGA-HRP histochemistry with tetramethylbenzidine as chromogen (Mesulam 1982). In Case 13, in 1 section of 5, CTB-g was revealed by the silver-intensification protocol described by Kritzer and Goldman-Rakic (1995). In Cases 14 and 29, 1 series of each fifth section was processed for the visualization of BDA, using a Vectastain ABC kit (Vector Laboratories, Burlingame, California) and 3,3'-diaminobenzidine (DAB) as a chromogen. The reaction product was intensified with cobalt chloride and nickel ammonium sulfate. In all cases, 1 series of each fifth section was stained with the Nissl method (thionin, 0.1% in 0.1 M acetate buffer pH 3.7), and in Cases 23, 27, and 29 a further series was stained for myelin (Gallyas 1979).
All the other brains, but Case 1, used for cytoarchitectonic analysis were processed as described above and cut frozen in coronal (5 hemispheres) or parallel to the direction of the IPS (2 hemispheres) sections, 60 µm thick. The 2 hemispheres of Case 1, embedded in celloidin, were cut, one in a plane perpendicular to the direction of the IPS, the other in a plane parallel to the direction of the IPS, both at 40 µm. In all cases, 1 series of each fifth section was stained with the Nissl method.
Data Analysis
Injection Sites and Distribution of Retrogradely Labeled Neurons
Injection sites were defined according to criteria previously described in detail (Luppino and others 2001, 2003) and attributed to the architectonic areas of the IPL convexity with analysis of adjacent Nissl-stained sections. The injection sites presented in this study (listed in Table 1) were all restricted within the limits of a single cytoarchitectonic area. One WGA-HRP injection in Case 27 involved both PG and PFG and was not considered for this study.
FB, DY, TB, WGA-HRP, and CTB-g labeling was identified as described in detail in Luppino and others (2001, 2003). CTB-A labeling was analyzed by using standard fluorescein (for CTB-A 488) or rhodamine (for CTB-A 555 and CTB-A 594) sets of filters. CTB-A 488–labeled neurons were identified for a green granular fluorescence in the cytoplasm and CTB-A 555– and CTB-A 594–labeled neurons for a red–orange and a red granular fluorescence in the cytoplasm, respectively. These 2 last tracers were never used in the same animal.
The distribution of retrograde and anterograde (for WGA-HRP and BDA injections) labeling was analyzed in each section every 300 µm and plotted in each section every 600 µm, together with the outer and inner cortical borders, by using a computer-based charting system. Data from individual sections were then imported into a three-dimensional (3D) reconstruction software (Bettio and others 2001), creating volumetric reconstructions of the hemispheres from individual histological sections containing connectional and/or architectonic data. The results of this processing allowed us to obtain realistic visualizations of the labeling distribution for a more precise comparison of data from different hemispheres. Distribution of labeling on exposed cortical surfaces was visualized in standard mesial, dorsolateral, or bottom views of the hemispheres. Distribution of labeling within sulci was visualized in nonstandard views of the hemispheres in which sulcal banks were exposed with appropriate dissections of the 3D reconstructions (Fig. 1).
|
Areal Attribution of the Labeling
Retrograde and anterograde labeling was found in several areas of the parietal, temporal, cingulate, agranular frontal, and prefrontal cortices.
In the parietal cortex, outside the IPL convexity, connections were attributed, when possible, to functional areas that, although in many cases still lack a clear architectonic definition, have been well established in electrophysiological studies. Accordingly, the lateral bank of the IPS was subdivided into a caudal (LIP), a rostral (anterior intraparietal, AIP), and a ventral (ventral intraparietal, VIP) area, according to Blatt and others (1990), Murata and others (2000), and Colby and others (1993), respectively. The SPL and the posterior cingulate cortex were subdivided as in Matelli and others (1998) (see also Marconi and others 2001) where functional areas V6A (Galletti and others 1999) and medial intraparietal (MIP) (Colby and others 1988; Colby and Duhamel 1991) were included in the map of Pandya and Seltzer (1982). Area V6A was subdivided into a dorsal (V6Ad) and a ventral (V6Av) sector according to Luppino and others (2005). For the parietal operculum the functional maps of the SII region and neighboring areas of Robinson and Burton (1980a, 1980b) and Krubitzer and others (1995) were considered, although these areas could not be precisely distinguished one from another. In cases of uncertain functional correspondence, labeling was attributed according to the architectonic maps of Pandya and Seltzer (1982) and Lewis and Van Essen (2000a). Temporal areas of the STS and inferior temporal gyrus were defined according to Boussaoud and others (1990) and Saleem and Tanaka (1996). In the frontal lobe, agranular frontal and cingulate areas were cytoarchitectonically defined according to Matelli and others (1985, 1991) and Geyer and others (2000). The prefrontal cortex was subdivided according to the cytoarchitectonic map of Walker (1940) and the prearcuate cortex also according to Stanton and others (1989) and Petrides and Pandya (2002).
Quantitative Analysis and Laminar Distribution of the Labeling
To obtain more objective information on the relative strength of the connections observed within the same case or across different cases, for each cortical injection, but those of BDA (because of the paucity of retrograde labeling observed with this tracer), we counted the number of labeled neurons plotted in the ipsilateral hemisphere in one section every 600 µm and located beyond the limits of the injected field. Because the absolute number of labeled neurons was largely variable across cases, mainly because of differences in amount, spread, and sensitivity of injected tracers, afferents to the injected field were expressed in terms of percent of labeled neurons found in a given cortical area or sector, with respect to the total number of labeled neurons. The percent distribution of the retrograde labeling observed for each area was then used for guiding the qualitative description of its connections. In this analysis, some sectors (e.g., parietal operculum) in which labeling extended across adjacent areas, which could not be precisely defined, were considered as a whole.
To obtain information on possible hierarchical relationships of the observed cortical connections, labeling attributed to a given area and reliably observed across different sections and cases, was analyzed in each section every 300 µm, in terms of laminar distribution of the anterogradely labeled terminals and in terms of percent of labeled neurons located in the superficial (I–III) versus deep (V–VI) layers. These data were then analyzed according to the criteria reviewed by Felleman and Van Essen (1991) (see also Andersen and others 1990). Based on the laminar distribution of labeled terminals, projections were classified as "feedforward" when mostly concentrated in layer IV and lower III, "feedback" when distributed in superficial and deep layers, but avoiding layer IV, "lateral" when fairly even distributed in all cortical layers, and "mixed" when patches of "feedforward" projections were found together with patches of "feedback" projections. Based on the laminar distribution of labeled neurons, connections were classified as "feedforward" or "feedback" when labeled neurons in the superficial layers were >70% or <30%, respectively. More equal distributions were classified as "bilaminar." This last pattern has been generally used to infer that 2 given areas are located at the same hierarchical level. However, as noted by Felleman and Van Essen (1991) (see also Andersen and others 1990; Boussaoud and others 1990; Lewis and Van Essen 2000b), bilaminar connections, at least at the level of parietal and temporal areas, can be compatible with different types of hierarchical relationship, according to the disposition of the anterograde labeling. Because most of the connections observed in the present study, as already noticed by Andersen and others (1990) and Lewis and Van Essen (2000b) in their connectional studies of different parietal areas, showed a bilaminar projection pattern, where not otherwise specified, possible hierarchical relationships as suggested by Felleman and Van Essen (1991) and following Andersen and others (1990) were established on the basis of the laminar distribution of labeled terminals (data available for Opt, PG, and PFG).
In the agranular frontal and cingulate areas, the lack of layer IV forced us to modify these criteria and connections characterized by retrograde labeling mostly in layers III and VI, and anterograde terminals mostly in layers III and V were considered as "feedforward" connections. In some cases, connections were characterized by retrograde labeling mostly in layers III and VI, and terminal labeling was densest in layers I and II and very weak in layer VI. This pattern only partially fits with the criteria used for the definition of "feedback" connections and was left undefined.
Photographic Presentation
Photomicrographs shown in the present study were obtained by capturing images directly from the sections with a digital camera attached to the macroscope or to the microscope. Individual images were then imported in Adobe Photoshop in which they could be processed, eventually assembled into digital montages, and reduced to the final enlargement. In most of the cases, image processing required lighting, contrast, brightness, and contrast adjustments.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
Cytoarchitecture of the IPL Convexity and Location of Injection Sites
The cytoarchitectonic analysis of the IPL convexity showed, in substantial agreement with Pandya and Seltzer (1982), that in this cortical sector 4 different fields can be defined and located at different rostrocaudal levels. Following the nomenclature of Pandya and Seltzer (1982) these fields will be here referred, from rostral to caudal, to as PF, PFG, PG and Opt. The major cytoarchitectonic criteria used in defining these fields are illustrated in Figure 2, in low-power photomicrographs of 4 Nissl-stained coronal sections taken at different rostrocaudal levels from Case 10 and in Figure 3 (upper part) in higher magnification views of representative fields from the same sections.
|
|
In PF, a radial pattern is recognizable in lower layer III as well as in layers V and VI. Cells in layer III display a size gradient with medium-sized pyramids spread in its lower half. Layer IV is homogeneous and lacks a sharp upper border with layer III. Layer V is relatively poor and thin, with rather small pyramids, and layer VI is broad and subdivided into 2 sublayers.
In PFG a columnar organization is evident only in layer III. In this layer, medium-sized pyramids are mainly concentrated in its lowest part. A well-developed layer V is evident, even in low-power views. This layer is populated mainly not only by medium-sized pyramids but also by scattered larger pyramids, which represent a major distinctive feature of PFG, compared with PF and PG. Layer VI is rather uniform.
In PG, the overall cellular density in layer III appears higher, compared with the more rostral areas. This layer is mainly formed by small pyramids, and the almost complete absence of larger cells gives it a rather uniform appearance. Layer V is well developed and populated by densely packed small pyramids. Layer VI is relatively homogeneous.
Opt displays a clear, broad columnar pattern particularly evident in layer III. A size gradient is present in layer III with many medium-sized pyramids occupying its lowest part. Layer IV is sharply defined, and it is denser than in PG. Cell size is also increased in layer V, compared with that of PG, with many medium-sized pyramids. Layer VI has a clear border with layer V and can be subdivided into 2 sublayers.
All these fields enter medially in the IPS for about 2 mm, whereas laterally PF, PFG, and PG border with opercular areas extending in the dorsal bank of the LF (PFop and PGop of Pandya and Seltzer 1982). In general, architectonic features were found to change gradually from one field to another, in the range of less than 1 mm. For this reason, cytoarchitectonic borders presented in this study represent the intermediate point of the transitions and were found to run roughly in the coronal stereotaxical plane, slightly obliquely in caudoventral direction.
The average location along the IPL convexity of the identified cytoarchitectonic fields was quantitatively estimated in 13 hemispheres of M. nemestrina and 8 hemispheres of M. fascicularis. Cytoarchitectonic borders, set as the intermediate points of the transitional zones, were measured in terms of antero-posterior stereotaxic coordinates (AP), according to the M. nemestrina atlas of Winters and others (1969) (AP values referred to the interaural line), and to the M. fascicularis atlas of BrainInfo (2000) (AP values referred to the anterior commissural line, AC). Average cytoarchitectonic maps, shown in Figure 3 (lower part) were, then, generated separately for the 2 species by plotting the average AP values on a dorsolateral view of a standardized hemisphere. To provide an estimate of the interindividual variability across hemispheres of the same species, standard deviation values of the mean AP position of the borders at the level of the lateral crown of the IPS and at the level of the border with the opercular areas are also shown in the maps. The result of this analysis showed that the location of these fields was quite constant across different cases and similar to that shown by Pandya and Seltzer (1982).
In selecting the location of the injection sites our aim was not only to involve different parts of each IPL field but also the more peripheral transitional zones to avoid spread of tracers across different fields. All the injection sites considered for this study (most of them shown in Fig. 4, in drawings of representative sections through the core and the rostral and caudal part of the halo) were restricted to a single architectonic field, only in few cases marginally involving transitional zones. For this reason, the patterns of connections described in this study mostly concern much more the core of each field rather than the transitional zones. Given the relatively low interindividual variability in the location of cytoarchitectonic borders, only in 1 case a WGA-HRP injection in Case 27 was found to be not restricted to a single field, involving both PG and PFG. The labeling distribution observed of in this case was fully compatible with an almost equal involvement of both these fields and, therefore, not considered in this study.
|
|
The general distribution of the retrograde labeling observed in Cases 23 and 27 DY and TB and drawings of representative coronal sections from Case 23 are presented in Figures 6 and 7, respectively. The percent distribution of the labeled neurons observed in Cases 23 and 27 DY, as well as the mean values of all the 3 Opt injections, are shown in Table 2. Representative patterns of the laminar distribution of retrograde and anterograde labeling observed in Case 23 are illustrated in Figure 8.
|
|
|
|
Parietal and Posterior Cingulate Cortices
In the IPL very strong "lateral" connections of Opt were observed with PG (Figs. 6, 7, section f, and 8, PG), whereas PFG was only very marginally labeled. Weak, "lateral" connections were observed in the parietal operculum, only with its outermost and caudalmost part, corresponding to area PGop of Pandya and Seltzer (1982). Caudal to Opt, moderate "feedback" connections were observed with the dorsal aspect of the prelunate gyrys (area DP, Andersen and others 1990). In the lateral bank of the IPS, numerous and dense patches of marked cells were observed in its caudal part in both the dorsal (LIPd) and the ventral (LIPv) subdivisions of area LIP (Blatt and others 1990). In this area, the anterograde labeling showed a "feedback" pattern, and retrograde labeling in layers I–III was >70% (Fig. 8, LIP). In the SPL, connections were limited to the mesial surface of the hemisphere and to the anterior wall of the parietooccipital sulcus (Fig. 6). In particular, these very strong connections extensively involved, with some variability in the relative distribution across cases, area PGm (Fig. 7, sections b–e), extending caudally in V6Av (Luppino and others 2005; Fig. 7, section a) and rostrally, in the caudal part of the cingulate gyrus (posterior cingulate cortex, CGp; Olson and others 1996). In all these areas the anterograde labeling showed a "feedback" pattern, and in PGm the labeled neurons in layers I–III were >70% (Fig. 8, PGm). Some labeling was also found more rostrally, in areas 23a and 23b.
Temporal Cortex, Including Area MST and Insula
Opt was connected with different STS and inferotemporal areas. In the caudal part of the STS (Fig. 6) very strong "feedback" connections (retrograde labeling in layers I–III >70%) were found in area MST (Figs. 7, section e, and 8, MST), mostly in its dorsal and caudal part (presumably dorsal MST, MSTd; Komatsu and Wurtz 1988). Weak labeling was observed in the middle temporal area (MT) (Fig. 7, sections d and e) and very sparse labeled cells in the fundal superior temporal area (FST). In the upper bank of the STS robust, "lateral," or "mixed" connections (Fig. 8, superior temporal polysensory area, STP) were observed in restricted sectors lateral and rostral to MST, attributable to both the posterior (STPp; Fig. 7, section e) and anterior (STPa; Fig. 7, sections h, i, and n) subdivisions of the superior temporal polysensory area, respectively. Ventral to STPa, "feedforward" connections were observed with the fundal region of the sulcus (area IPa; Fig. 7, section l; Fig. 8, IPa), extending also in the ventral bank, in the medial part of area TE (TEm) (Fig. 7, sections m and n). Additional labeling in the inferotemporal cortex, showing a "lateral" pattern, was observed in the postero-ventral part of area TE (TEpv) (Figs. 6 and 7, sections g and h), on the lateral lip and the fundus of the occipitotemporal sulcus. With the only exception of a small cluster of marked neurons observed in the postero-dorsal part of area TE (TEpd) in Case 27 (Fig. 6), in both Cases 23 and 27, labeling in TEm and TEpv was observed in very similar locations, suggesting that Opt is target of specific subsectors of these inferotemporal areas. Spots of labeling were also observed at different rostrocaudal levels in the parahippocampal area TF, and few scattered marked neurons were located in the perirhinal cortex (Figs. 6 and 7, sections g, i, l, and m). In Case 23, some purely anterograde labeling was found in the caudal part of the presubiculum. Very poor labeling was inconstantly located in the granular insula (Fig. 7, sections l and n).
Agranular Frontal and Cingulate Cortices
Two agranular frontal sectors, located in the dorsal premotor cortex (PMd) and ventral premotor cortex (PMv), respectively, showed relatively weak connections with Opt (Fig. 6). In PMd (Fig. 7, section o), labeling was consistently observed in the lateral part of the rostral PMd area F7, not including the supplementary eye field (F7 non-SEF [Luppino and others 2003]). In this premotor sector, anterograde labeling was very weak in deep layers, and much denser in layers I and II (Fig. 8, F7). In PMv (Fig. 7, section p), labeling, with some variability across cases, was found in the rostral area F5, in a relatively rostral part of the posterior bank of the arcuate sulcus, the anterograde labeling being mostly focused in layer III ("feedforward" pattern). In the agranular cingulate cortex, sparse labeling was observed in area 24b.
Prefrontal Cortex
Several relatively weakly labeled sectors were observed in the prefrontal cortex (Figs. 6 and 7, sections p–r). In both Cases 23 and 27, some labeling was located relatively caudally in the principal sulcus, mostly in the ventral bank and much weaker labeling was found on the mesial surface of the hemisphere, in medial area 8B. Some labeling was also found in the ventral prearcuate cortex, in area 45 ventral to the frontal eye field (FEF), as defined cytoarchitectonically in adjacent Nissl-stained sections (area 45b of Petrides and Pandya 2002). Part of the labeling was also observed on the dorsal lip of the principal sulcus (dorsal area 46) and in the dorsalmost part of area 8A. Finally, a very weak connection was also observed with the orbitofrontal area 12o (Fig. 6). All prefrontal connections of area Opt showed a "feedforward" pattern (Fig. 8, 45).
Connections of Area PG
Five tracer injections in 3 animals (Case 20, WGA-HRP; Case 27, CTB-A 594; Case 29, TB and CTB-A 488; Case 29, BDA) were placed in different parts of area PG. Figure 5B shows the location of the WGA-HRP injection in Case 20, placed in the middle of the inferior parietal gyrus (see also Fig. 4), in a sector where cytoarchitectonic features typical of area PG, for example, a layer III quite homogeneous in cell size and density, a layer V densely populated by relatively small pyramids, could be observed in the adjacent Nissl-stained section (Fig. 5B1).
The general distribution of retrograde labeling observed in Cases 20 (WGA-HRP) and 27 (CTB-A 594) and drawings of representative coronal sections from Case 20 are presented in Figures 9 and 10, respectively. The percent distribution of the labeled neurons observed in these 2 cases, as well the mean values of all the PG injections, but the BDA one, is shown in Table 2. Representative patterns of laminar distribution of retrograde and anterograde labeling observed in Cases 29, BDA, and 20 are illustrated in Figure 11.
|
|
|
Parietal and Posterior Cingulate Cortices
In the IPL, strong "lateral" connections were observed with areas PFG, PGop, and the rostral part of area Opt (Figs. 9, 10, sections d–g, and 11, Opt). A few marked neurons were also found in area PF and in area DP. In the parietal operculum, in addition to PGop, very strong connections showing a "feedback" pattern were observed with the retroinsular cortex, but also more rostrally and deeply, with area SII (Figs. 9, 10, sections f–h, and 11, SII). A more rostral, minor labeling can be attributed to area PV. In the lateral bank of the IPS, moderate "lateral" connections were observed with the mid-rostral part of it, mostly involving area AIP and, at a very minor extent, area VIP (Figs. 9 and 10, sections f–h). Area LIP, which was heavily connected with Opt, was virtually devoid of marked neurons. Moderate to rich connections showing a "lateral" pattern (Fig. 11, PEci) were observed in different areas of the caudal part of the SPL, with some variability in their relative distribution across cases. These connections involved the caudal part of the ventral bank of the cingulate sulcus (area PEci), area V6Ad (Figs. 9 and 10, sections a and b) and area PEc. Area PGm and the CGp were virtually devoid of labeling. Dense patches of "lateral" connections (Fig. 11, MIP) were also observed in the medial bank of the IPS, where, although with some variability across cases, most of them were observed in the mid-caudal part of it, attributable mostly to area MIP. In the cingulate area 23, rich "lateral" connections (Fig. 11, 23c) were observed with the ventral bank of the cingulate sulcus (area 23c) and the dorsal part of the cingulate gyrus (area 23b).
Temporal Cortex, Including Area MST and Insula
Very rich labeling was observed in MST (Figs. 9 and 10, sections d and e), which extended also more rostrally (and deeply) with respect to Opt injections, possibly involving also the lateral part of MST (MSTl, Komatsu and Wurtz 1988). Dense, but restricted labeling was also found at different rostrocaudal levels of area STP attributable to both STPp and STPa (Figs. 9 and 10, sections f, i, and o). In particular, the labeling in STPa shown in Figure 10, section i, appears to occupy a location similar to the labeling observed following injections in Opt. In both MST and STP the anterograde labeling showed a "feedback" pattern (Fig. 11, MST and STP). Very weak "feedforward" connections (more evident following BDA injection) were also observed in area IPa (Figs. 10, section m, and 11, IPa). Very sparse labeling was observed in area MT (Fig. 10, section c) and in perirhinal and parahippocampal cortices (Fig. 10, section l). In 2 cases (Case 29 BDA and 29 TB), some labeling was observed in area TEav, close to the anterior medial temporal sulcus. In Case 29 BDA, labeled terminals were also observed in the caudal part of the presubiculum. One additional and distinctive relatively strong connection of PG, showing a "feedback" pattern (retrograde labeling in layers I–III >70%), was located caudally, in the ventral bank of the LF, extending also on the adjacent part of the superior temporal gyrus (Figs. 9, 10, sections f and g, and 11, C), involving areas C of Morel and others (1993) and Tpt of Pandya and Sanides (1973; see also Lewis and Van Essen 2000a). Rich "feedforward" connections (retrograde labeling in layers I–III <70%) were also observed in insular area Ig (Figs. 9 and 10, section l).
Agranular Frontal and Cingulate Cortices
Minor connections were observed with several agranular frontal areas. In PMv, clusters of marked cells, with labeled terminals densest in layer III ("feedforward" pattern; Fig. 11, F5), were located along the entire extent of the posterior bank of the inferior arcuate sulcus (area F5; Figs. 9 and 10, sections n and o). In PMd, labeling was consistently observed in the ventrorostral part of area F2 (F2vr; Luppino and others 2003), close to the spur of the arcuate sulcus (Figs. 9 and 10, sections m–o), and additional, sparse labeling was located in area F6/pre-supplementary motor area (pre-SMA) (Fig. 9). Finally, clusters of labeled cells were also found in the cingulate motor area 24d (Fig. 10, section n), mainly in the ventral bank of the cingulate sulcus and in area 24b (Fig. 9). In F2vr, F6 and area 24, labeled terminals were mostly confined to layers I and II (Fig. 11, F2vr).
Prefrontal Cortex
The only prefrontal sector consistently labeled following PG injections was the ventral part of area 46 (Figs. 9 and 10, sections p and q). In particular, "lateral" connections (Fig. 11, 46v) were found mostly in the ventral bank of the principal sulcus, partially overlapping with the sector connected with Opt, but also extending more dorsally to include the ventral lip of the sulcus and the adjacent cortical convexity.
Connections of Area PFG
Five tracer injections were placed in 4 animals in different parts of area PFG (Case 13, WGA-HRP; Case 14, BDA; Case 27, CTB-A 488; Case 29, FB and CTB-A555). Figure 5C shows the location of the WGA-HRP injection site in Case 13, placed in the mid-ventral part of the inferior parietal gyrus (see also Fig. 4), where cytoarchitectonic features typical of area PFG, for example, medium-sized pyramids in the lowest part of layer III and a well-developed layer V, with occasional relatively large pyramids, could be observed in the adjacent Nissl-stained section (Fig. 5C1).
The distribution of the retrograde labeling observed in Cases 13, WGA-HRP, and 29, FB (shown in light blue), is illustrated in Figure 12. The percent distribution of the labeled neurons observed in these 2 cases, as well as the mean values of all the PFG injections, but the BDA one, is shown in Table 2. Three-dimensional reconstructions of Case 29 in Figure 12 also show in red the distribution of the neurons marked following TB injection in PG to provide direct comparison between the differential patterns of cortical connectivity of PFG and PG found in the present study. Drawings of representative coronal sections from Case 13 are presented in Figure 13, and representative patterns of the laminar distribution of retrograde and anterograde labeling observed in Cases 14, BDA, and 13 are illustrated in Figure 14.
|
|
|
Parietal and Posterior Cingulate Cortices
Following PFG injections, the labeled territory extended to the adjacent caudal and rostral areas PG and PF, respectively (Figs. 12 and 13, sections b, c, and f). The strong, "lateral" connections between PFG and PG (Fig. 14, PG) were very evident in Case 29, where retrograde tracers were injected in both these areas (Fig. 12). In the parietal operculum, similarly to PG, very dense labeling ("lateral" connections) was found more caudally in correspondence of PGop (Fig. 13, sections c and d) and retroinsular cortex, and more rostrally ("feedback" connection; Fig. 14, SII) in correspondence of SII (Fig. 12, sections e–h) and PV. In the lateral bank of IPS, strong "lateral" connections were observed in areas AIP and VIP (Figs. 12, 13, sections d–g, and 14, VIP). The dense labeling in VIP extended also in the rostral part of the medial bank of the IPS, in the rostral part of area PEa, rostral to MIP, corresponding to the medial IPS sector source of corticospinal projections (area PEip of Matelli and others 1998; Figs. 12, 13, sections d, e, and h, and 14, PEa). In contrast, in areas MIP and V6Ad, source of rich afferents to PG (Fig. 12, Case 29, TB, red labeling) marked neurons were poor. Weak connections ("lateral" pattern) were observed with areas PEci and 23.
Temporal Cortex, Including Area MST and Insula
"Feedback" connections (retrograde labeling in layers I–III >70%), weaker than those observed following PG injections, were observed in area MST in all cases of injections in PFG (Figs. 12, 13, section b, and 14, MST). In STP restricted but relatively robust labeling showing a "mixed" type of connections (Fig. 14, STP) was found in both STPp and STPa (Figs. 12 and 13, sections c, g, and h). In particular, the labeled STPa sector in Figure 13, section g, appeared to largely overlap with the STPa sector connected with PG and Opt. In the ventral bank of the STS, in addition to few marked neurons found in area MT (Fig. 13, section a), some labeling was also located in area FST (Fig. 13, section d). More rostrally, some clusters of marked neurons were consistently observed, in all cases, in areas IPa and TEm, where labeled terminals showed a "feedforward" pattern (Figs. 12, 13, sections e and f, and 14, TEm). Robust "feedforward" connections (retrograde labeling in layers I–III <70%) were observed in the insular area Ig.
Agranular Frontal and Cingulate Cortices
Connections of PFG with the agranular frontal cortex were rich and by far densest in PMv. (Figs. 12 and 13, sections i–m). In F4, which is virtually not connected with PG, relatively weak labeling was almost completely confined to the dorsal part of it, whereas in F5 a rich labeling involved the whole posterior bank of the arcuate sulcus, extending also on the cortical convexity. Labeled terminals in both these areas tended to be relatively evenly distributed across all layers ("lateral" connections, Fig. 14, F5). Weak labeling (with high degree of variability among cases) was observed, especially in Case 14, BDA, in F2vr, where terminals were mostly concentrated in the superficial layers (Fig. 14, F2vr) and even sparser labeling could also be observed in F6/pre-SMA, F3 and F1. In area 24 labeling showing a "lateral" pattern was observed in areas 24d and 24b.
Prefrontal Cortex
Substantial connections, with some variability across cases, were observed with ventral area 46 in a location similar to that labeled following injections in PG (Figs. 12 and 13, sections o and p). Additional and distinctive connections of area PFG, with respect to PG (although with some variability in their strength among cases), were observed in the caudal and lateral part of the orbitofrontal cortex (area 12o of Carmichael and Price 1994) and in the disgranular, precentral opercular area (PrCO) ventral to F5. All these connections could be classified as "lateral" (Fig. 14, 46v).
Corticospinal Projections from the IPL Convexity
Figure 15 shows the distribution of the labeled corticospinal neurons observed in the hemispheres of Cases 10 and 21 contralateral to large HRP injections in the lateral funiculus of the spinal cord at upper cervical levels. Considering the size and the level of the injections, the distribution of labeled neurons can be considered, in both cases, quite representative of the origin of projections to all spinal levels caudal to C4–C5. In the IPL convexity a cortical sector, very well corresponding to area PFG, was found to be a source of corticospinal projections. These projections represent a distinctive connectional feature of this area with respect to the other IPL convexity areas. In Case 10, DY was also injected in the same lateral funiculus at the upper thoracic level to identify the origin of projections to the thoracic and lumbar spinal levels. The results (not shown in the figure) showed that DY-labeled neurons were virtually absent in PFG, thus suggesting that this area is a source of corticospinal projections mostly directed to the cervical levels of the spinal cord.
|
Connections of Area PF
Three tracer injections were placed in area PF in 3 different monkeys (Case 13 CTB-g, Case 27 FB, and Case 29 DY), which produced remarkably similar distributions of retrograde labeling. Figure 5D shows the location of the DY injection site in Case 29, placed in the middle of the inferior parietal gyrus (see also Fig. 4), in a cortical sector showing a dense layer III with medium-sized pyramids in the lower half, a well-developed layer IV, and a relatively poor layer V (area PF, Fig. 5D1). Figure 16 shows the results observed in Case 29 DY in 3D reconstructions of the injected hemisphere and in drawings of representative coronal sections. The percent distribution of the labeled neurons observed in Case 29 DY and 27 FB, as well as the mean values of all the PF injections, is shown in Table 2. In general, PF displayed connections with a much more limited set of cortical areas, with respect to the other IPL convexity fields, virtually all confined to the parietal and frontal cortices.
|
Parietal Cortex
The major feature of the cortical connections of PF was the very strong connection with the postcentral gyrys (Fig. 16, dorsolateral view and sections d and e). In this sector, retrograde labeling was very dense and almost completely confined to area 2, in its ventral part. In the IPL convexity, the connections with area PFG were very strong, extending also more caudally to involve, at a minor extent, area PG (Fig. 16, dorsolateral view and sections a and b). In the parietal operculum (Fig. 16, upper bank of the LF and sections d–f), very dense labeling was found in the SII region, with substantial labeling involving, presumably, also area PV. In the IPS, connections were quite strong with areas AIP and VIP (Fig. 16, sections a–c) but weak with rostral PEa. Weak labeling was observed in the insular cortex (Fig. 16, section f).
Frontal Cortex
In the agranular frontal cortex the labeling was quite dense in both PMv areas F4 and F5. In F4, labeling was densest in the dorsal part of this area, close to the spur of the arcuate sulcus (Fig. 16, dorsolateral view and sections f and g). In F5 (Fig. 16, postarcuate cortex and sections f–i), it included the whole extent of the posterior bank of the inferior arcuate sulcus and the cortical convexity, where labeled neurons appeared to extend also more ventrally with respect to those observed following injections in PFG. Relatively weak labeling was also found ventral to F5 in area PrCO and in ventral area 46 (Fig. 16, dorsolateral view and section m). In all cortical areas, but the granular insula, the distribution of the retrograde labeling was bilaminar. In the granular insula, marked cells showed a "feedback" pattern.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionsReferences |
|---|
In this study we traced the cortical connections of the 4 architectonic fields—PF, PFG, PG, and Opt—forming the macaque IPL cortical convexity by placing tracer injections aimed to involve different parts of each field, but the more peripheral, transitional ones. The results showed that each of these fields is robustly connected with the neighboring ones and displays markedly different patterns of connections with visual-, somatosensory-, auditory-, and limbic-related areas and with parietal and frontal areas related to the control of different effectors. Figure 17 summarizes the main results of this study.
|
It is largely accepted in neuroscience that the cerebral cortex contains many functionally distinct domains, usually referred to as "areas," although there is no consensus on what precisely constitutes a cortical area and what the best criteria are for their definition (see, e.g., Van Essen 1985
). In general, 3 main experimental approaches, the architectural, the connectional, and the functional, are considered most useful for the definition of a cortical area. Converging evidence from these 3 approaches is generally considered a strong argument for a reliable identification and delineation of a cortical area (see, e.g., Van Essen 1985; Felleman and Van Essen 1991; Lewis and Van Essen 2000a). In this respect, connectivity patterns undoubtedly provide an important basis for identifying cortical areas and determining whether neighboring regions belong to the same or different areas (see, e.g. Lewis and Van Essen 2000b