Three-Dimensional Topography of Corticopontine

THE JOURNAL OF COMPARATIVE NEUROLOGY 478:306–322 (2004)

Three-Dimensional Topography of Corticopontine Projections from Rat Sensorimotor Cortex: Comparisons with Corticostriatal Projections Reveal Diverse Integrative Organization

TRYGVE B. LEERGAARD,1 KEVIN D. ALLOWAY,2 TUYET A.T. PHAM,1 INGEBORG BOLSTAD,1 ZACHARY S. HOFFER,2 CHRISTIAN PETTERSEN,1 AND JAN G. BJAALIE1* 1Neural Systems and Graphics Computing Laboratory, Centre for Molecular Biology and Neuroscience and Department of Anatomy, University of Oslo, N-0317 Oslo, Norway 2Department of Neural and Behavioral Sciences, Penn State University, Hershey, Pennsylvania 17033

ABSTRACT The major cortical–subcortical re-entrant pathways through the basal ganglia and cerebellum are considered to represent anatomically segregated channels for information originating in different cortical areas. A capacity for integrating unique combinations of cortical inputs has been well documented in the basal ganglia circuits but is largely undefined in the precerebellar circuits. To compare and quantify the amount of overlap that occurs in the first link of the cortico–ponto–cerebellar pathway, a dual tracing approach was used to map the spatial relationship between projections originating from the primary somatosensory cortex (SI), the secondary somatosensory cortex (SII), and the primary motor cortex (MI). The anterograde tracers biotinylated dextran amine and Fluoro-Ruby were injected into homologous whisker representations of either SI and SII, or SI and MI. The ensuing pontine labeling patterns were analyzed using a computerized three-dimensional reconstruction approach. The results demonstrate that whisker-related projections from SI and MI are largely segregated. At some locations, the two projections are adjoining and partly overlapping. Furthermore, SI contributes significantly more corticopontine projections than MI. By comparison, projections from corresponding representations in SI and SII terminate in similar parts of the pontine nuclei and display considerable amounts of spatial overlap. Finally, comparison of corticopontine and corticostriatal projections in the same experimental animals reveals that SI–SII overlap is significantly larger in the pontine nuclei than in the neostriatum. These structural differences indicate a larger capacity for integration of information within the same sensory modality in the pontocerebellar system compared to the basal ganglia. J. Comp. Neurol. 478:306–322, 2004. © 2004 Wiley-Liss, Inc.

Indexing terms: somatosensory cortex; motor cortex; cerebellum; pontine nuclei; basal ganglia; 3-D reconstruction

other investigators. For other related papers, see www.nesys.uio.no/ Grant sponsor: The Research Council of Norway; Grant sponsor: Euro- Database. pean Community Grant; Grant number: QLRT-2000-02256; Grant spon- *Correspondence to: Jan G. Bjaalie, NeSys, CMBN & Anatomy, Univer- sor: National Institutes of Health Grant; Grant number: NS-37532. sity of Oslo, P.O. Box 1105, Blindern, N-0317 Oslo, Norway. This article is linked to a neuroinformatics Web archive available at E-mail: [email protected] http://www.nesys.uio.no/. The archive contains (1) superimposed multila- Received 16 April 2004; Revised 1 July 2004; Accepted 5 July 2004 beling microscope images, (2) atlas-style illustrations for all experimental animals, (3) videos of three-dimensional distribution patterns and overlap DOI 10.1002/cne.20289 analysis, and (4) original coordinate ﬁles for downloading and free reuse by Published online in Wiley InterScience (www.interscience.wiley.com).

The basal ganglia and cerebellum are both part of projec- pared the relative patterns of corticostriatal and cortico- tion systems that originate in the cerebral cortex and return pontine projections from the same pairs of cortical sites. to the cortex via the thalamus (Mercier et al., 1990; Brodal Compared with the neostriatum, the pontine nuclei pro- and Bjaalie, 1992; Heimer et al., 1995; Parent and Hazrati, vide more opportunities for integrating information from 1995; Schmahmann and Pandya, 1997). The cortical regions related sensory areas. devoted to motor, somatosensory, or other distinct modali- ties, project to separate territories in the neostriatum and pontine nuclei (Ku¨ nzle, 1975, 1977; Brodal, 1978; MATERIALS AND METHODS Hartmann-von Monakow et al., 1981; Wiesendanger and Experiments were conducted on male Sprague-Dawley Wiesendanger, 1982b; Mihailoff et al., 1985; Selemon and rats, ranging in weight from 300 to 600 g. The animals Goldman-Rakic, 1990; Flaherty and Graybiel, 1991; previously were used for analysis of corticostriatal and Schmahmann and Pandya, 1997). The presence of topo- corticothalamic projections (Alloway et al., 2000; Hoffer graphic organization and limited convergence in corticostria- and Alloway, 2001). Among 32 animals used in these tal and corticopontine projections, as well as in the remain- previous studies, 19 were examined for labeling in the ing divisions of the basal ganglia and cerebellar projection pontine nuclei. All animal procedures were reviewed by an systems, has been interpreted to represent functionally seg- institutional animal welfare committee and were in com- regated cortico-subcortical re-entrant channels, or “parallel pliance with NIH guidelines for the use and care of labo- circuits” (Alexander et al., 1986, 1990; Alexander and ratory animals. Crutcher, 1990; Goldman-Rakic and Selemon, 1990; Hoover and Strick, 1993; Middleton and Strick, 2000). Surgical procedures Although segregation is prominent, some evidence indicates that functional integration occurs within cortico- Each rat was initially anesthetized with an intramus- striatal and corticopontine projections. In the neostria- cular injection of ketamine (20 mg/kg) and xylazine (6 tum, structural overlap has been reported for projections mg/kg). Additional anesthesia was administered approxi- arising from adjacent parts of the primary somatosensory mately every 45–60 minutes or when withdrawal reflexes (SI) cortex (Flaherty and Graybiel, 1991; Alloway et al., were evoked. Atropine sulfate (0.05 mg/kg) was adminis- 1998; Hoover et al., 2003), from homologous representa- tered to reduce bronchial secretions. The animals were tions in the SI and primary motor (MI) cortical areas intubated through the oral cavity, placed in a stereotaxic (Flaherty and Graybiel, 1993, 1995; Graybiel and Kimura, frame, and artificially ventilated with 100% oxygen. Body 1995; Hoffer and Alloway, 2001; Ramanathan et al., 2002), temperature, end-tidal CO2, and heart rate were moni- and from different cortical areas involved with the same tored continuously. After exposure of the skull, wound modality (Parthasarathy et al., 1992; Inase et al., 1996; margins were infiltrated with 2% lidocaine. Separate Alloway et al., 2000; Nambu et al., 2002). By comparison, craniotomies were made at coordinates consistent with the few data available for the pontine nuclei are incongru- published maps of the SI, SII, and MI whisker represen- ous (for discussion, see Brodal and Bjaalie, 1997; Schwarz tations (Hall and Lindholm, 1974; Chapin and Lin, 1984; and Thier, 1999; Bjaalie and Leergaard, 2000; Leergaard, Neafsey et al., 1986; Fabri and Burton, 1991). 2003). Partial overlap has been reported for projections originating in different visual areas (Bjaalie and Brodal, Electrophysiological mapping 1989) and for projections from neighboring sites within SI SI–MI experiments. High impedance Elgiloy micro- barrel cortex (Leergaard et al., 2000a). Other combina- electrodes (1 M⍀) were placed in MI to evoke twitches tions of cortical sites, however, project to segregated fields from the contralateral whiskers. Stimulating microelec- in the pontine nuclei (Wiesendanger and Wiesendanger, trodes were lowered 1,750 ␮m below the pial surface to 1982b; Bjaalie and Brodal, 1989; Schwarz and Thier, activate layer V pyramidal cells at coordinates that 1995; Schwarz and Mo¨ck, 2001). matched the vibrissal representations in previous reports In this report, a dual tracing approach was used to (Hall and Lindholm, 1974; Donoghue and Wise, 1982; assess the spatial relationships among rat corticopontine Neafsey et al., 1986). Electrical stimulation, which con- projections originating in areas of different or related mo- sisted of cathodal pulse trains lasting for 50 msec (0.7- dalities. Because overlap is most likely to occur among msec pulses and 2.3-msec interpulse intervals), was de- functionally related projections (Bjaalie and Brodal, 1989; livered periodically to elicit contralateral whisker see also, Flaherty and Graybiel, 1993; Alloway et al., 2000; movements. Several penetrations were often made to de- Hoffer and Alloway, 2001), we injected small amounts of termine the threshold current (25–90 ␮A) that barely tracers into electrophysiologically identified, homologous evoked visible movements of a discrete group of mystacial whisker representations. While one tracer was placed in whiskers. In some instances, however, even the lowest the SI barrel cortex, a second tracer was placed in the microstimulation currents evoked movements from sev- corresponding whisker representation in either MI or in eral rows of whiskers. After locating an appropriate whis- the secondary somatosensory (SII) cortex. Computerized ker representation in MI, one tracer was deposited at that three-dimensional reconstruction techniques were used to cortical site (see below). characterize the distribution and overlap of the ensuing A tungsten microelectrode was used to record SI neuro- anterograde labeling patterns. Our findings indicate that nal discharges, evoked by moving individual whiskers homologous sites in SI and MI project to partly adjacent with a small wooden probe. A map of the SI body repre- but primarily nonoverlapping regions in the pontine nu- sentation (Chapin and Lin, 1984) was used to guide the clei. Similar projections from SI and SII, however, show recording locations until we found sites in SI barrel cortex considerable overlap. Because the experimental animals that appeared to match the whisker representations in- previously were used to study corticostriatal projections jected in MI cortex. These SI barrel sites were then in- (Alloway et al., 2000; Hoffer and Alloway, 2001), we com- jected with small quantities of the other tracer. 308 T.B. LEERGAARD ET AL.

TABLE 1. Location, Size, Separation, and Histology of Tracer Injections in SI, SII, and MI Whisker Representations1 SI injections SII injections Separation Cortical Case Whiskers Tracer (n) Area (mm2) Whiskers Tracer (n) Area (mm2) (␮m) sections

SS40 D2 BDA (2) 0.148 D2-3 FR (1) 0.047 1,470 coronal SS42 D2 BDA (1) 0.070 D3-5 FR (1) 0.054 1,760 tangential SS43 D2-4 BDA (3) 0.270 D3-4 FR (3) 0.267 1,496 tangential SS45 D2 BDA (2) 0.103 rows C-E FR (2) 0.071 1,025 coronal SS46 C3 BDA (1) 0.404 C3-4 FR (2) 0.233 1,090 coronal SS47 C2-3 BDA (1) 0.281 C1-3 FR (2) 0.062 1,100 coronal SS48 D3-4 BDA (1) 0.379 D2-3 FR (2) 0.118 1,020 coronal SS49 D2 BDA (1) 0.084 D2-5 FR (2) 0.097 1,225 coronal SI injections MI injections M18 D2,E3 FR (2) 0.112 row E BDA (2) 0.255 5,897 tangential M22 C2 BDA (2) 0.411 rows C-D FR (2) 0.117 5,500 tangential M25 D3-4 FR (3) 0.182 D2-3 BDA (3) 0.655 4,030 tangential M26 D3-4 BDA (3) 0.302 D2-3 FR (3) 0.183 4,040 tangential M27 D2-4 BDA (3) 0.868 rows D-E FR (3) 0.217 4,388 tangential M28 E2 FR (2) 0.226 D2-4 BDA (4) 0.478 3,678 tangential

1Values in parentheses indicate number of tracer penetrations. SI, primary somatosensory cortex; SII, secondary somatosensory cortex; BDA, biotinylated dextran amine; FR, Fluoro-Ruby.

SI–SII experiments. A second group of animals re- Published maps of the MI, SI, and SII cortical areas (Hall ceived separate anterograde tracer injections in the whis- and Lindholm, 1974; Chapin and Lin, 1984; Fabri and ker representations of the SI and SII cortical areas. The Burton, 1991) were adapted according to our recent ana- procedures for mapping SI and SII whisker representa- tomical tracing study that investigated the interconnec- tions were identical to those outlined for SI above. After tions of all three cortical regions (Hoffer et al., 2003). injecting the SII cortex with one tracer, the other antero- While the outlines of the SI body map (Fig. 1B, solid lines) grade tracer was then injected into the corresponding SI are well defined in the literature (Welker, 1971, 1976; whisker barrel representation. Chapin and Lin, 1984), the outlines of the MI and SII body maps (Fig. 1B, dashed lines) are less certain (Hall and Tracer injections Lindholm, 1974; Donoghue and Wise, 1982; Fabri and The tracers used were rhodamine-conjugated dextran Burton, 1991; Hoffer et al., 2003). Nonetheless, several amine (Fluoro-Ruby, FR; D-1817, Molecular Probes, Eu- anatomical tracing studies suggest that the MI and SII gene, OR) and biotinylated dextran amine (BDA; D-1956, whisker representations have a topographical organiza- Molecular Probes). tion (Alloway et al., 2000; Hoffer and Alloway, 2001, Hof- BDA injections. A 10% solution of BDA was loaded fer et al., 2003). into a glass pipette with a tip diameter of 45–75 ␮m. The pipette was lowered into layer 5 of the cortex (at a depth of Perfusion and histochemistry 1.7 mm in MI and 1.4 mm in SI and SII) and a small Each rat was deeply anesthetized with sodium pento- quantity of BDA was iontophoretically injected using pos- barbital (50 mg/kg i.p.) and then placed in a stereotaxic itive current (4–6 ␮A) pulses with a duty cycle of 7 sec- frame. Fiduciary marks were made in the brain with a onds. Iontophoresis continued for 9 minutes, and then the tungsten wire coated with India ink. The animal was then pipette was raised 200 ␮m for another iontophoretic injec- transcardially perfused with 500 ml of heparinized saline, tion that lasted for 9 minutes. This process was repeated followed by 500 ml of cold 4% paraformaldehyde, pH 7.4, several times until the pipette was 600–800 ␮m below the and then 500 ml of 4% paraformaldehyde with 10% su- pia. crose. The brain was placed overnight in cold fixative with FR injections. A 10% solution of the tracer was pres- 30% sucrose. Sections through the brainstem and cortex sure injected into the brain by using a Hamilton syringe were cut at 50 ␮m on a freezing microtome. The angle of that had a 50- to 75-␮m glass tip cemented to the needle. sectioning was transversal in the brainstem and coronal The syringe was placed in a pressure injector (model 5000; or tangential in the cortex (Table 1). To visualize BDA Kopf Instruments, Tujunga, CA) and lowered into layer 5 labeling, sections were processed according to steps 1–7 in of the cortex. A volume of 25 nl of FR was injected at each Lanciego and Wouterlood (Lanciego and Wouterlood, of three different cortical depths (at 200-␮m increments). 1994) using a streptavidin–biotinylated horseradish per- Injection penetrations for all injections were made peroxidase complex (Amersham, Buckinghamshire, UK) as a pendicular to the cortical surface (vertically for MI, tilted substitute for the avidin–biotin solution used in the orig- at 25 degrees for SI, and tilted at 34 degrees or 50 degrees inal protocols. Sections were mounted on gelatin-coated for SII). In most cases the pipette was reinserted into a glass and cover-slipped with Eukit. Alternate cortical sec- nearby cortical site for another set of injections. Table 1 tions were processed for cytochrome oxidase according to shows the tracers used in the individual cases and the whis- the procedure of Wong-Riley (Wong-Riley, 1979). ker representations that were injected. After completion of the injections, the wound margins were sutured and the Data acquisition animal was returned to its home cage for 7–10 days. The distribution of labeled axons within the pontine Composite brain maps. The location and size of the nuclei and several anatomical landmarks were recorded injection sites were drawn on a composite brain map of using an image-combining computerized microscope sys- MI, SI, and SII (Fig. 1) based on the electrophysiology tem, based on a Zeiss Axioskop II light and fluorescence protocol notes and the histological data (Table 1; Fig. 2). microscope. Details concerning the software and technical SENSORIMOTOR CORTICOPONTINE PROJECTIONS 309

Fig. 1. Mystacial vibrissae representations in primary motor cor- (positive values) and posterior (negative values) to Bregma. The ver- tex (MI), primary somatosensory cortex (SI), and secondary somato- tical axis shows distances from the midline. The outlines of the MI, SI, sensory cortex (SII). A: The cartoon representation indicates the po- and SII maps were adapted from several published maps of these sition of MI, SI, and SII in the right cerebral hemisphere (modiﬁed areas (Hall and Lindholm, 1974; Chapin and Lin, 1984; Fabri and from Welker, 1971). B: This composite maps shows the topographical Burton, 1991; Hoffer et al., 2003). A–E, whisker rows; 1–8, whisker organization of whisker representations in MI (blue), SI (red), and SII arcs; El, eyelid; FBP, furry buccal pad; FL, forelimb; HL, hindlimb; (green), in relation to a two-dimensional skull-based stereotaxic coor- LJ, lower jaw; N, nose; PO, perioral; TR, trunk; UZ, unresponsive dinate system with a 1-mm grid using the intersection of Bregma and zone; V, vibrissae. the midline as origin. The horizontal axis shows numbers anterior

solutions have been reported previously (Leergaard and plexuses of BDA-labeled axons were viewed with translu- Bjaalie, 1995; Lillehaug et al., 2002). Complete series of cent light and the FR-labeled axons with excitation light sections through the pontine nuclei were digitized by us- of 534–558 nm (Zeiss ﬁlter set no. 15). ing a Zeiss Plan-APOCHROMAT 10ϫ/0.45 or Plan- The labeled plexuses within the pontine gray were APOCHROMAT 20ϫ/0.75 lens. The ventral surface of the coded semiquantitatively as points (see also Leergaard brainstem, the outlines of the pontine gray, the contours of and Bjaalie, 1995; Leergaard et al., 1995, 2000a,b). In the corticobulbar and corticospinal ﬁber tracts (in the areas with a low density of labeling, point coordinates following referred to as the peduncle), and the midline and were placed at regular intervals, corresponding to the the fourth ventricle were recorded as references for the spacing of varicosities, along the length of single axons. In three-dimensional (3-D) reconstruction (see below). The areas with dense labeling, it was impossible to assign

Fig. 2. Location of dual tracer injections of biotinylated dextran section through the cerebral cortex showing BDA (black labeling) and amine (BDA) and Fluoro-Ruby (FR) in case M25, SS49, and SS43 FR injection sites in secondary somatosensory cortex (SII) and SI, (Table 1). A: Image of a tangential section through cortical layer V respectively, from case SS49. D: Composite image of a tangential showing the location of the BDA injection site in primary motor section through cortical layer V, showing BDA and FR injections in SI cortex, case M25. B: Image of a tangential section through cortical and SII, respectively, from case SS43. A, anterior; D, dorsal; L, lateral; layer V showing the location of the FR injection site in primary M, medial. Scale bars ϭ 500 ␮m. somatosensory cortex (SI), case M25. C: Composite image of a coronal 310 T.B. LEERGAARD ET AL. coordinates to each labeled fiber and varicosity. Nonethe- squares of ϳ20 ␮m2 by using a grid. Each square was less, we sought to make digital reconstructions in which assigned a color corresponding to the density of dots the density of plotted points corresponded roughly to the within a radius of 100 ␮m centered on the square. density of labeling that was visualized in the tissue (Fig. Illustrations were assembled with Adobe Illustrator 10 3). Labeling from the two tracers was digitized indepen- and Adobe Photoshop 7.0. Digital photomicrographs were dently, and the correspondence of labeling and digitized obtained through a CoolSnap camera (Photometrics, Tuc- image for both tracers was studied carefully using alter- son, AZ). The grayscale levels and contrast of the images nating light- and fluorescence microscopy with the 10ϫ were optimized by using Adobe Photoshop. and 20ϫ lenses. Overlap analysis 3-D reconstruction and visualization Analysis of overlapping corticopontine projections were For the 3-D reconstruction, visualization, and analysis performed in all cases and compared with overlap data for of labeling patterns, we used the program Micro3D for corticostriatal and corticothalamic projections from the Silicon Graphics workstations (Neural Systems and same experimental animals (from Alloway et al., 2000; Graphics Computing Laboratory, University of Oslo, Oslo, Hoffer and Alloway, 2001). Each digitized section was Norway; http://www.nesys.uio.no/). Earlier versions of subdivided into a grid, and the amount of labeled overlap this software were used in several recent investigations (i.e., co-location of fibers labeled with FR and BDA, respec- (Leergaard et al., 1995, 2000a,b; Bjaalie et al., 1997a,b; tively, in the same bin) was determined by using three Malmierca et al., 1998; Berg et al., 1998; Bajo et al., 1999; different bin sizes: 35, 50, and 75 ␮m2. For each of these Vassbø et al., 1999). The serially ordered digitized sections bin sizes, bins containing at least two data points, repre- were aligned interactively on the screen using the above senting FR or BDA labeling, were colored red or blue, mentioned anatomical boundaries and landmarks. To fa- respectively. Those bins that contained at least two data cilitate the alignment, the 3-D reconstructions were in- points of both types of labeling were colored yellow. The spected from different angles of view using real-time ro- number of red, blue, and yellow bins was separately tation. Each section was assigned a z-value defined by its counted in the pontine nuclei, neostriatum, and the me- thickness and serial number. The 3-D reconstruction from dial part of the posterior nucleus of the thalamus (POm). each case was scaled to an average-size pontine coordinate Total overlap was calculated as the sum of the yellow bins system (see below). The scaled reconstructions served as a divided by the sum of red, blue, and yellow bins. Each bin basis for the further visualization and analyses of the size (35, 50, or 75 ␮m2) was used to calculate the amount densities and distribution of the point clusters represent- of overlap in each brain region because the overlap esti- ing the anterograde labeling (Figs. 4–6). For this purpose, mate is highly influenced by bin size. If very small bins tools available in Micro3D allowed re-slicing of data sets (e.g., 5 ␮m2) were used, the amount of overlap appeared in multiple user-defined section planes. A novel applica- minimal, whereas large bins produced a substantial in- tion (J.O. Nygaard, S. Gaure, C. Pettersen, H. Avlesen, crease in overlap. Statistical analysis of variance and J.G. Bjaalie, see http://www.nesys.uio.no/) based on (ANOVA) was performed by using the statistical analysis Java 3D (Sun Microsystems, Inc., Santa Clara, CA) was package included with the Microsoft Excel spreadsheet used to create envelopes surrounding the point clusters program. (Fig. 7A–I). The boundaries of the clusters were defined with use of implicit representations of geometries. Thus, scalar fields were generated by binning point coordinates RESULTS and estimating the relative density for each bin. For visu- Of the 19 animals examined, 5 contained very limited alization, isosurfaces were extracted by marching cube labeling in the pontine nuclei of one or both tracers and like algorithms. were discarded from further analysis. The exact reason for To compare data across animals, we applied the stan- the reduced labeling in those animals is uncertain, but the dard pontine coordinate system as described in our previ- cases involved animals that received MI injections in sites ous reports (Leergaard et al., 2000a,b; Brevik et al., 2001). that required high microstimulation currents to evoke A cuboid bounding box was oriented along the long axis of whisker movements and may have been slightly rostral to the brainstem at the level of the pons, and the boundaries the MI whisker representation. The remaining 14 cases were adjusted to fit the (histologically defined) rostral, were submitted to a detailed 3-D analysis of corticopon- caudal, lateral, medial, and ventral limits of the pontine tine labeling patterns. The animals were injected in SI nuclei. We also used the ventral surface of the peduncles and MI (n ϭ 6) or in SI and SII (n ϭ 8) as listed in Table as an intermediate landmark (for details, see Brevik et al., 1. For this report, results from three representative ani- 2001). The origin of the pontine coordinate system was mals are illustrated and described in detail. Complete defined as the intersection of the three planes formed by documentation of all cases, including original data sets the medial and rostral sides of the bounding box, and the with 3-D spatial coordinates representing labeling re- plane halfway between the ventral and dorsal planes. corded within the pontine nuclei, is available by means of Relative coordinates, from 0 to 100%, were defined from the NeSys database at http://www.nesys.uio.no/. rostral to caudal, medial to lateral, and from “central” (the plane halfway from ventral to dorsal) to ventral (100%) Terminology and dorsal (Ϫ100%) within the pontine nuclei. Affine transformations were used to combine data from all ani- In the term pontine nuclei, we here include the grey mals in the same pontine coordinate system. matter that partly embraces the descending fiber tract Density maps of corticopontine labeling were produced (corticobulbar and corticospinal fibers, referred to as the by dividing a particular projection (view from ventral, Fig. peduncle) and the medial lemniscus. The dorsally located 8D–F) of different combinations of 3-D point-data sets into nucleus reticularis tegmenti pontis, which is differently Fig. 3. A–I: Images of characteristic anterograde labeling patterns same sections. The dot map is subdivided into 35-␮m2 bins, and the within the ipsilateral pontine nuclei after dual injection of Fluoro- number of FR and BDA dots is counted in each bin. Bins containing at Ruby (FR) and biotinylated dextran amine (BDA) into primary so- least one or more FR or BDA dots are colored red or black, respec- matosensory cortex (SI) and primary motor cortex (MI), or SI and tively; those containing at least two of each type are colored yellow (cf. secondary somatosensory cortex (SII), from rats M25 (A–C), SS49 Materials and Methods section, see also Fig. 7J–O). A–C show adja- (D–F), and SS43 (G–I). A,D,G: The left column shows FR labeling in cent, largely segregated axonal clusters labeled after injection of FR transverse sections through the pontine nuclei (from approximately into SI and BDA into MI. D–F show smaller, overlapping axonal midpontine rostrocaudal levels, cf. Fig. 9B,G) viewed by fluorescence clusters labeled after injection of FR into SII and BDA into SI. At this microscopy. B,E,H: The middle column shows BDA labeling in the level, the patch of BDA labeling arising in SI is located entirely within same sections viewed by light microscopy. C,F,I: The right column the boundaries of the FR patch arising in SII. G–H shows partially shows an overlay of the two images. Corresponding positions in the overlapping axonal clusters after injection of FR into SII, and BDA image rows are indicated by red triangles. K–M: Raw data plots of the into SI. D, dorsal; M, medial. Scale bars ϭ 100 ␮m in I (applies to A–I); same sections as shown in A–I. (K corresponds to A–C; L corresponds 500 ␮minMЈ. to D–F, and M corresponds to G–I). K؅–M؅: Overlap analysis of the 312 T.B. LEERGAARD ET AL.

Fig. 4. Computer-generated three-dimensional reconstruction fied from Fig. 1). A: Reconstruction shown in views from ventral, from case M25, showing the topography of corticopontine projections rostral, or medial. B,C: Consecutive series of transverse (B) and after injections of biotinylated dextran amine (BDA) and Fluoro-Ruby sagittal (C) slices of 200 ␮m thickness through the reconstruction. The (FR) into homologous whisker representations in primary motor cor- numbers assigned to each slice refer to the coordinate system and tex (MI)and primary somatosensory cortex (SI). Dots represent the indicate the rostrocaudal or mediolateral level of each slice. On the distribution of corticopontine fibers within the pontine nuclei, labeled right side (ipsilateral to the tracer injections), the red dots are dis- with BDA (black) or FR (red). A frame of reference, defined by planes tributed in a large lateral and a smaller medial cluster. The black dots tangential to the boundaries of the pontine nuclei, is superimposed are distributed in two clusters in the rostral half, and one smaller onto the reconstruction (for details, see Materials and Methods sec- cluster in the caudal half. Together, the clusters form a ring-like tion). A coordinate system of relative values from 0 to 100% is intro- volume. Although red and black dots have a close spatial relationship duced. The halfway (50%) reference lines are shown as dotted lines in at several locations, it is evident from the slices in C and D that the the ventral, rostral, and medial views. The position of the injection labeled clusters are largely segregated. Scale bars ϭ 500 ␮minA,B sites is indicated in the outline drawing of the cortical maps (simpli- (applies to B,C).

organized with respect to cytoarchitecture and connectiv- SI, by comparisons with CO histochemistry. Whisker rep- ity (Mihailoff et al., 1981; Torigoe et al., 1986), is not resentations for the MI injections were determined by included. Subdivisions of the pontine nuclei, as previously using microstimulation to evoke twitches among a small used by Mihailoff et al. (1981) and others, are not referred group of whiskers. Because of these technical differences, to in the present study because there is no apparent cor- as well as differences in the topography and magniﬁcation relation between these subdivisions and the patterns of factors of different cortical body maps, the tracer injec- afferent and efferent connections (for review, see Brodal, tions in SI usually involved a small number of whiskers, 1987). whereas tracer injections in SII and MI generally involved a larger group of whiskers (Table 1). Injection sites The size of each tracer injection (Table 1) was estimated In all animals, BDA and FR injection sites were located by measuring the area of dense labeling in cortical layer V, within the cortical grey matter and were restricted to the the layer containing the cell bodies of the corticopontine whisker representations of SI, SII, or MI (Fig. 2). Whisker projection neurons. The boundary of the area of dense representations at the injection sites in SI and SII were labeling was placed at the location where the dense extra- determined by multiunit recordings and, additionally for cellular labeling was reduced to a level that made it pos- SENSORIMOTOR CORTICOPONTINE PROJECTIONS 313

Fig. 5. A–C: Computer-generated three-dimensional reconstruc- partially overlapping and located within a relatively narrow, arrow- tion from case SS49, showing the topography of corticopontine projec- shaped lamellar volume (see ventral view). Within this volume, the tions after injections of Fluoro-Ruby (FR) and biotinylated dextran two populations of dots are to some extent segregated with respect to amine (BDA) into homologous whisker representations in secondary the dorsal to ventral, and rostral to caudal distribution of dots. Scale somatosensory cortex (SII)and primary somatosensory cortex (SI). bars ϭ 500 ␮m in A,B (applies to B,C). Presentation as in Figure 4. The clusters of red and black dots are

sible to observe individual labeled fibers and cell bodies. mus of the same experimental animals (Alloway et al., These measurements were then used for making matched- 2000; Hoffer and Alloway, 2001). sample comparisons. This analysis showed no significant In all cases, BDA and FR gave rise to distinctly labeled differences between the sizes of the paired injections in fibers with beaded varicosities along the entire axonal the SI and MI groups of animals (MI ϭ 0.317 mm2;SIϭ trajectory, as well as plexuses of tortuous and branched 0.349 mm2,nϭ 6). By comparison, for the other group of fibers in the target regions. Because these tracers have animals the tracer injections in SII were approximately been reported to give rise to retrograde transport from the half the size of the injections in SI (SII ϭ 0.118 mm2;SIϭ injection site, followed by anterograde transport into the 0.217 mm2,nϭ 8). To place tracers in corresponding SI collaterals of the retrogradely labeled neurons (Merchań and SII whisker representations, the SII injections needed et al., 1994; Malmierca et al., 1998), we carefully searched to be smaller because the whisker representations in this for collateral labeling that could interfere with our anal- area are smaller than in SI (Fig. 1). ysis. This was particularly relevant for the comparison of the SI and SII projections, because these projections dis- General features of labeling played a high degree of overlap (as outlined below). As Tracer injections into the whisker representations of SI, reported in our parallel study of the neostriatal projec- SII, and MI, produced labeling in several intracortical and tions (Alloway et al., 2000), examination of our cortical subcortical targets. This report describes the pattern of injection sites revealed corticocortical connections be- corticopontine projections from these cortical areas and tween SI and SII, along with retrogradely labeled cortical compares the amount of labeled overlap in the pontine neurons. However, only corticocortical fibers proceeding nuclei with that observed in the neostriatum and thala- from these retrogradely labeled neurons to the other in- 314 T.B. LEERGAARD ET AL.

Fig. 6. A–C: Computer-generated three-dimensional reconstruc- somatosensory cortex (SII)and primary somatosensory cortex (SI). tion from case SS43, showing the topography of corticopontine projec- Presentation as in Figure 4. The clusters of red and black dots are tions after injections of Fluoro-Ruby (FR) and biotinylated dextran partly overlapping, with an external to internal shift in the prepon- amine (BDA) into homologous whisker representations in secondary derance of labeling. Scale bars ϭ 500 ␮m in A,B (applies to B,C).

jection site were seen. Descending fibers projecting to sub- was the similar sizes of the SI and MI injection sites (see cortical levels from the retrogradely labeled neurons in SI above). Matched sample analyses of the dot numbers indi- (or SII) were not observed. We thus conclude that labeled cated that the amount of labeling originating in SI was fibers in the pontine nuclei represent direct projections significantly higher than the amount originating in MI at from neurons at the injection site. P Ͻ 0.05 with a one-tailed paired t test (n ϭ 6, mean number In the pontine nuclei, most of the labeling was confined of recorded dots in SI ϭ 1,611 and in MI ϭ 731). The rela- to the ipsilateral side, where anterogradely labeled fibers tively stronger contribution of SI projections is also seen in branched extensively to form several dense axonal plex- Figure 4. In the case here illustrated, the more than three- uses, assumed to represent terminal fields (Fig. 3). Fewer fold larger MI injection site gave rise to only 20% more plexuses, and lower densities of fibers within each plexus, recorded data points in the pontine nuclei than the SI injec- were observed contralaterally. In single sections, the plex- tion site. Other examples are shown in our data repository uses of BDA- and FR-labeled axons typically formed (http://www.nesys.uio.no/). Similar analysis comparing the rounded patches with a diameter of 50–300 ␮m and elon- projections from SI and SII also showed that SI contributed gated bands that were ϳ50–100 ␮m wide (Fig. 3). more projections than SII but were deemed inconclusive because the paired injections were of different size (see Comparison of relative corticopontine above). projection strengths from different areas The relative anatomical strength of corticopontine projec- Spatial distribution of labeling tions from the paired injections in SI and MI was compared We analyzed the spatial relationship and amount of by using the number of points recorded for each category of overlap for corticopontine projections from SI and MI and labeling (Figs. 4–6) as an estimate of the spatial extent and from SI and SII. All data were first transferred to our density of terminal labeling. The basis for this comparison standard coordinate system for the pontine nuclei (Leer- SENSORIMOTOR CORTICOPONTINE PROJECTIONS 315

Fig. 7. Computer-generated stereo-pairs showing the three- overlap analysis (for details, see Fig. 3), with red (FR) and blue (BDA), dimensional (3-D) shape and relationship of Fluoro-Ruby–labeled and yellow (FR and BDA) bins. The shape and distribution of labeled (FR, red) and biotinylated dextran amine–labeled (BDA, blue) termi- clusters originating from primary somatosensory cortex (SI, A,E,F) nal ﬁelds within the right pontine nuclei (ipsilateral to the injection and secondary somatosensory cortex (SII, B,C) is similar (H,I), sites) for cases M25 (left column), SS49 (middle column), and SS43 whereas labeled clusters originating from primary motor cortex (MI, (right column). The position and extent of the injections is indicated in D) have a more rostral and ventral distribution (G). Compared to case the outline drawings of the cortical maps (simpliﬁed from Fig. 1). To SS49 (middle column), the larger SI and SII injections used in case see 3-D images (A–O), the viewer must cross the eye axis to let the SS43 (right column) produce thicker clusters located side-by-side. pair of images merge. A–I: Isodensity surfaces surrounding the clus- M–O: The dual injections in SI and SII produce more overlap (N,O) ters of FR and BDA labeling, shown individually (A–F) and combined than the dual injections in SI and MI (M). (G–I) in view from ventral. J–L: The 3-D distribution of the results of

gaard et al., 2000a,b; Brevik et al., 2001). This facilitated 4–6). In addition, to analyze spatial relationships at a comparisons across cases, as illustrated in Figures 4–9 more detailed level, we applied computerized tools for and documented more completely in our data repository at slicing the 3-D reconstructions in two orthogonal planes: http://www.nesys.uio.no/. To describe the spatial relation- transverse and sagittal (Figs. 4–6). Furthermore, to per- ships between the labeled clusters, including size, shape, ceive the shapes of clustered labeling, we used geometric and density of labeling, data were shown as dot maps in modeling with isodensity surfaces (Fig. 7). In our initial standard diagrams at three different angles of view (Figs. analysis of anatomical overlap, we inspected the spatial 316 T.B. LEERGAARD ET AL.

Fig. 8. A–F: Dot maps (A–C) and density maps (D–F) showing the sities Ͻ5% of the maximum value are not shown. Overall, the distri- distribution of whisker-related corticopontine projections from pri- bution of high density regions of labeling is quite similar for mary motor cortex (MI, A,D), primary somatosensory cortex (SI, B,E), projections from SI and SII, whereas MI labeling is differently dis- and secondary somatosensory cortex (SII, C,F). Data from all cases tributed. Together, the SI, SII, and MI high-density regions form (Table 1, Fig. A) have been accumulated in the same standard pontine complementary parts of a ring-like volume. Within the large lateral coordinate system and are shown in view from ventral. For further clusters of SI and SII labeling, the highest densities of SII labeling are details regarding presentation, see legend to Figure 4. Only the right located between two high-density spots of SI labeling (compare E and half of the pontine coordinate system is shown. D–F: The color gradi- F). Scale bar ϭ 500 ␮m in F (applies to A–F). ent shows the highest densities in red and the lowest in violet. Den- relationships among the two tracers in each individual observed across all animals in this group. Tracer injec- case (Figs. 4–6) and in diagrams showing accumulated tions into the MI whisker representations labeled multi- data for all experiments (Figs. 8, 9). To obtain a quanti- ple, relatively large, globular axonal clusters, typically in tative measure for the degree of overlap between projec- the rostromedial, rostrolateral, and caudomedial parts of tions from pairs of injection sites, we analyzed data, ani- the pontine nuclei (Figs. 3B, 4). In some cases, the rostral mal by animal, by applying a grid to each section and then clusters together occupied a large horizontally oriented, counting the number of FR- and BDA-labeled bins in the elongated and curved subspace (Fig. 7D). The 3-D stereo grid, before computing statistics from the data (He et al., images of isodensity surfaces surrounding the labeling 1993; Alloway et al., 1999; Leergaard et al., 2000a). (Fig. 7) clearly show the shapes and locations of the pon- SI corticopontine projections. Tracer injection into tine territories receiving projections from MI and SI. In the SI whisker representations gave rise to several, the contralateral pontine nuclei, labeling originating in sharply defined axonal clusters, most notably a large MI was usually sparse, appearing as a weak mirror image curved cluster laterally and a smaller round cluster me- of the ipsilateral labeling (Fig. 4B). Only very limited dially (Figs. 4–7). The obtained labeling patterns and overlap among the SI and MI projections was observed in overall topographic organization of all SI–corticopontine individual tissue sections (see Fig. 3A–C, showing an in- labeling observed in the present study were similar to dividual FR-labeled fiber originating in SI encroaching those previously described (Leergaard et al., 2000a; see upon a BDA-labeled plexus originating in MI) and in the also http://www.nesys.uio.no/ for full documentation of all 3-D reconstructions (Figs. 4, 7G). A density gradient anal- cases from both studies). Small axonal clusters originating ysis applied to data accumulated from all cases empha- in SI were frequently observed medially in the contralat- sized that the labeling originating in MI and SI is located eral pontine nuclei (Figs. 4–6). in nearby regions of the pontine nuclei, with regions of Comparison of SI and MI corticopontine projections. highest densities clearly segregated (Fig. 8). The overall Corticopontine projections from MI and SI terminated in spatial separation between SI– and MI–corticopontine specific, partly adjacent locations that were consistently projections is also shown to advantage in the slices SENSORIMOTOR CORTICOPONTINE PROJECTIONS 317

Fig. 9. Computer-generated dot maps showing accumulated data the large clusters of MI, SI, and SII labeling appear to be spatially from all cases in the same standard pontine coordinate system. Only associated in B, the transverse and sagittal slices show that the the right half of the pontine coordinate system is shown. Dots repre- labeling originating in primary motor cortex (MI, blue dots) is largely sent corticopontine labeling and are color coded according to site of segregated from labeling originating in primary somatosensory cortex origin. A: The outline drawing of the cortical maps shows the approx- (SI)and secondary somatosensory cortex (SII). Only minor topo- imated cortical areas covered by the respective multiple injections. graphic differences between SI and SII pontine projections are visible, B: View from ventral. C–E: The 50-␮m-thick sagittal slices, viewed especially in the sagittal slices (C,D). Scale bar ϭ 500 ␮m in H (applies from medial. F–H: The 50-␮m-thick slices, viewed from rostral. The to A–H). location and thickness of the slices in C–H is indicated in B. Although

through the accumulated data (see Fig. 9). Thus, corticopontine terminals from SI (red dots) and MI (blue dots) are mostly separated. Intermingling of red and blue dots, indicating a potential for synaptic convergence, is seen at some locations. Although determination of synaptic convergence would require ultrastructural analysis, which TABLE 2. Total Overlap in the Neostriatum, Pontine Nuclei, was not performed in this study, the amount of overlap and Nucleus POm that we observed using light microscopy can be quantified Pontine Nucleus and subjected to statistical analysis. Case nuclei Neostriatum POm For the quantitative analysis of overlap, we used data SI-SII cases from individual dual-tracing experiments and applied SS40 9.70 11.90 6.94 SS42 8.10 2.42 1.12 grids with different bin sizes to all of the reconstructed SS43 17.83 0.52 1.73 sections. The number of bins containing more than two SS45 1.78 1.88 1.42 SS46 15.17 1.67 4.56 data points of each tracer relative to the total number of SS47 15.22 4.20 6.90 bins with labeling was used as a measure of spatial over- SS48 11.46 9.21 2.30 ␮ 2 SS49 10.73 6.65 12.36 lap. When 35 m bins were used to calculate overlap, Mean (Ϯ SEM) 11.24 (Ϯ 1.77) 4.80 (Ϯ 1.43) 4.66 (Ϯ 1.38) total overlap for the SI–MI cases (n ϭ 6) ranged from 0% SI-MI cases to 3.67%, with a mean of 0.87 Ϯ 0.57% (mean Ϯ SEM, M18 0.00 0.37 0.83 M22 0.00 0.50 0.23 Table 2). With larger bin sizes, mean overlap increased to M25 0.81 0.69 2.83 2.17 Ϯ 1.30% for 50 ␮m2 bins and to 4.62 Ϯ 2.39% for 75 M26 0.18 1.53 2.09 ␮ 2 M27 3.67 2.20 3.75 m bins. These results reflect the overlap of the two M28 0.56 1.60 3.94 projections at a few specific locations in the pontine nuclei Mean (Ϯ SEM) 0.87 (Ϯ 0.57) 1.14 (Ϯ 0.29) 2.27 (Ϯ 0.62)

(Figs. 7J,M, 10, see also Figs. 4, 9). Overlap due to single 1Values expressed as percentages calculated from 35 ␮m2 bins. ﬁbers labeled with one tracer entering a plexus of ﬁbers POm, posterior nucleus of the thalamus. 318 T.B. LEERGAARD ET AL.

Fig. 10. Comparison of corticostriatal, corticopontine, and cortico- whisker representations, as listed in Table 1. For each bin size, thalamic overlap as a function of different pairs of cortical injection overlap was deﬁned as a minimum of two labeled varicosities for each sites and different bin sizes. Each bar represents the mean amount of tracer. Asterisks indicate instances in which corticostriatal or corti- labeled overlap (Ϯ SEM) for combinations of tracer deposits placed in copontine overlap were signiﬁcantly different from corticothalamic the primary somatosensory cortex (SI)and secondary somatosensory overlap for a given bin size (paired Student’s t tests, *P Ͻ 0.05, **P Ͻ cortex (SII, n ϭ 8) or in the SI and primary motor cortex (MI, n ϭ 6) 0.01).

labeled with the other tracer (as shown in Fig. 3A–C) did Fig. 9). Quantitative analysis on individual cases using 35 not usually exceed the threshold level for our operational ␮m2 bins showed that the total overlap index for the definition of overlap, in which two data points for each SI–SII cases (n ϭ 8) ranged from 1.78% to 17.83%, with tracer must occupy a bin (compare image and overlap mean overlap of 11.24 Ϯ 1.77% (see Table 2). With larger analysis of same sections in Fig. 3D–F,L,LЈ). bin sizes, mean overlap increased to 18.12 Ϯ 2.10% for 50 Comparison of projections from SI and SII. Tracer ␮m2 bins and to 26.18 Ϯ 2.42% for 75 ␮m2 bins. Collec- injections into the SII whisker representations produced la- tively, these data indicate that SI and SII corticopontine beling patterns that were highly similar to those produced projections displayed considerably more spatial overlap by tracer injections into homologous SI sites, both with re- than corresponding projections from the SI and MI corti- spect to shape, size, and location in the pontine nuclei (Figs. cal areas (Figs. 7K,L,N,O, 10, see also Figs. 4, 9). 5–7). In all cases, we observed a continuous large lateral Comparison of corticopontine and cluster of SI/SII labeling, which in some cases could be seen as two disjunctive clusters (Figs. 5B–medial view, and 7E). corticostriatal projections In nearly all cases, we also observed a smaller, medially The animals in this study previously were used to analyze located cluster of SI / SII labeling. The SI and SII cortico- the relative patterns of corticostriatal projections, either pontine clusters of labeling were typically partially overlap- from the SI and SII whisker representations or from the SI ping. At some locations, SI and SII axonal clusters were fully and MI whisker representations (Alloway et al., 2000; Hoffer overlapping (Fig. 3D–F; BDA patch located inside the bound- and Alloway, 2001). Therefore, in cases in which both the aries of the FR patch), but more frequently, the largely corticostriatal and corticopontine projections had been re- overlapping patches appeared to be spatially shifted with constructed, we compared tracer overlap to assess the rela- respect to each other (Fig. 3G–I,L–M). tive capacities of the neostriatum and pontine nuclei for Small tracer injections into homologous representations integrating sensorimotor information. As shown in Figure in SI and SII gave rise to labeled clusters that together 10, in both the pontine nuclei and the neostriatum, the formed a narrow lamella in the pontine nuclei (Figs. 5, amount of labeled overlap was highest when the tracers 7H). Larger tracer injections into SI and SII, presumably were placed in the SI and SII whisker representations. Thus, involving some nonhomologous cortical representations, when 35-␮m2 bins were used to calculate labeled overlap, the gave rise to pontine labeling in thicker lamellae that were mean amount of corticostriatal overlap was 4.80% when SI located side-by-side (Figs. 6, 7H). and SII were injected but only 1.14% when the tracers were Both qualitative and quantitative analysis of cortico- placed in SI and MI. Corresponding numbers for the corti- pontine projections from the SI and SII injection sites copontine overlap showed the same tendency (Table 2; Fig. revealed that the overall distributions of labeling were 10), with a mean overlap of 11,24% (SI–SII injections) and largely overlapping (Fig. 8). The spatial association be- 0.87% (SI–MI injections). tween the two data sets is evident from slices through the The amount of labeled overlap in the neostriatum and accumulated data shown in Figure 9. Red and green dots, pontine nuclei could be attributed to many parameters, in- which represent SI and SII labeling in Figure 9, are cluding the specific pair of injected cortical areas, their ter- clearly intermingled at many locations and indicate more minal projection strengths, and the proximity of the tracer overlap than observed for SI and MI (red and blue dots in injections. We have shown previously that the medial part of SENSORIMOTOR CORTICOPONTINE PROJECTIONS 319 the POm has precise topographically organized connections copontine projections than corresponding MI representa- with SI, SII, and MI (Alloway et al., 2003); therefore, we used tions. Finally, a comparison of corticopontine and cortico- labeled overlap in POm as a benchmark for normalizing the striatal projections revealed a higher degree of SI–SII comparisons of corticostriatal and corticopontine overlap in overlap in the pontine nuclei than in the neostriatum. The the two groups of animals (SI–SII vs. SI–MI). Although the opposite principle applied to the projections from SI and mean amount of labeled overlap in the nucleus POm was MI, with a lower degree of overlap in the pontine nuclei somewhat higher in the SI–SII cases than in the SI–MI compared with the neostriatum. cases (4.66% vs. 2.27%), this difference was statistically in- Due to the intricacy of the spatial patterns present significant (Student’s t test, P Ͼ 0.10). Given that the con- inside the pontine nuclei, the high resolution mapping and nections between POm and sensorimotor cortex are topo- 3-D reconstruction used in the present study was essential graphically organized, the amount of labeled overlap in the for analysis of the topography and overlap. The 3-D recon- neostriatum and pontine nuclei should be similar to that struction allowed distribution patterns to be studied in observed in the POm if corticostriatal and corticopontine totality and from different angles of view and, further- projections are similar to corticothalamic projection pat- more, was useful for demonstrating where the overlap terns. Therefore, we compared corticopontine, corticostria- occurred. The use of a standard representation of the tal, and corticothalamic overlap as a function of using 35, 50, pontine nuclei facilitated an overall comparison of data or 75 ␮m2 bins. from different animals. Finally, the dextran amine tracers By using a two-way ANOVA with replication model, in used in the present study provided specific results suitable which the same reconstructed points in each brain region as a basis for an accurate analysis of the topography and were analyzed repeatedly with different bin sizes, labeled overlap. These tracers label axonal arborizations in much overlap in the SI–SII cases varied significantly for differ- detail and offer opportunities for dual labeling in the same ent subcortical brain regions (F ϭ 19.6; P Ͻ 0.001) and sections (for review, see Reiner et al., 2000). ϭ Ͻ with changes in bin size (F 16.8; P 0.001). Labeled Relative contributions of SI and MI overlap in the SI–MI cases also varied significantly for these different subcortical brain regions (F ϭ 7.3; P Ͻ corticopontine afferents 0.01) and for changes in bin size (F ϭ 18.2; P Ͻ 0.001). The The pontine nuclei receive information from large parts results of this two-way ANOVA indicated that there were of the cerebral cortex, but there are clear regional differ- no interactions among these parameters (brain region and ences in the relative contribution of each cortical area. The bin size) for either group (SI–SII or SI–MI). most extensive corticopontine projections are known to The variation in labeled overlap attributed to brain arise from motor, somatosensory, and visual cortical areas region was mainly due to differences between the pontine (Brodal, 1978, 1981; Brodal and Bjaalie, 1992; Ruigrok nuclei and the other two brain regions. As depicted by and Cella, 1995). In general, the motor related areas of the Figure 10 for the SI–SII cases, matched-sample analyses cerebrocerebellar system have been emphasized (Allen indicated that labeled overlap was significantly higher in and Tsukahara, 1974), and strong corticopontine projec- the pontine nuclei than in the thalamus regardless of bin tions from motor cortex have been described in numerous size (paired t tests; P Ͻ 0.05 in all cases), whereas similar reports (see, e.g. Brodal, 1968, 1978, 1983; Mizuno et al., analyses revealed no difference in the amount of labeled 1973; Wisendanger et al., 1979; Hartmann-von Monakow overlap in the neostriatum and thalamus. When the trac- et al., 1981; Glickstein et al., 1985; Vassbø et al., 1999). ers were placed in SI and MI, however, labeled overlap In the rat, some retrograde tracing studies of cerebro- was lower in the pontine nuclei than in the thalamus. pontine pathways have indicated that the highest densi- Matched-sample analyses of the data in the SI–MI group ties of corticopontine neurons are found in motor areas, revealed that both corticopontine and corticostriatal over- closely followed by somatosensory and visual areas (Legg lap were significantly lower than corticothalamic overlap et al., 1989). An earlier study by Wiesendanger and regardless of bin size (paired Student’s t tests; P Ͻ 0.05 in Wiesendanger (1982a), however, reported that somatosen- all cases). Based on these findings, we conclude that cor- sory (granular cortex) areas provide a higher density of ticopontine projections from SI and SII have an organiza- corticopontine projections than motor areas (agranular tion that is functionally different from the organization of cortex). Our results support the view that SI contributions corticostriatal or corticothalamic projections from these to corticopontine projections are quantitatively stronger sensory cortical areas. than the corresponding MI contributions. A possible ex- planation for the discrepancy between data obtained with retrograde and anterograde tracing could be differences in DISCUSSION the amount of terminal branching of axons originating in This study demonstrates that whisker-related cortico- the two areas, suggesting that SI fibers are more profusely pontine projections from homologous sites of SI and MI branched within the pontine nuclei than fibers from MI. are largely segregated but occupy partly adjacent territo- This issue can presumably only be resolved by tracing of ries of the pontine nuclei. By comparison, projections from individual corticopontine fibers and their terminal fields corresponding representations in SI and SII have a closer (see also discussion in Vassbø et al., 1999). spatial relationship, with considerable spatial overlap of Furthermore, the SI pontine projections are highly or- the terminal fields of labeling. All of the corticopontine ganized, even at the level of projections from individual projections studied follow relatively simple principles of whisker barrels (Leergaard et al., 2000a, present study). 3-D topographical organization, in which the labeled ax- This strong relative contribution and precision of sensory ons aggregate in distinct clusters distributed within ring- cerebrocerebellar input is particularly interesting in light like or lamellar pontine volumes (see also, Leergaard, of the increasingly recognized role of the cerebellum in 2003). The projection strengths appeared to be different, nonmotor functions (see, e.g., Gao et al., 1996; Parsons with SI whisker representations contributing more corti- and Fox, 1997; Parsons et al., 1997; Schmahmann, 1997, 320 T.B. LEERGAARD ET AL.

1998; Middleton and Strick, 2000), including theories in 1985; Bjaalie and Brodal, 1989; Brodal et al., 1991; see which the cerebellum is thought to be involved in moni- also Brodal and Bjaalie, 1997). Electrophysiological inves- toring and adjusting the acquisition of sensory data, and tigations in rats, cats, and monkeys also support the no- thereby facilitating the efficiency with which other brain tion that pontine neurons are influenced by more than one structures perform various “motor,” “sensory,” or “cogni- cortical area (Oka et al., 1975; Ru¨ egg and Wiesendanger, tive” functions (Bower, 1997). In contrast to higher mam- 1975; Ru¨ egg et al., 1977; Potter et al., 1978). Taken to- mals, rodents rely heavily on somesthesis-guided whisk- gether, we surmise from these observations that different ing movements and related movements of the perioral populations of corticopontine neurons, which originate in structures and distal forelimbs during exploratory behav- localized but separate parts of the cortex, often terminate ior (Welker, 1964; Bower, 1997). The neostriatum, how- in overlapping pontine territories. ever, with MI afferents dominating those from SI (Hoffer Segregation and integration in the pontine and Alloway, 2001), may be involved primarily in initial- izing and regulating voluntary motor functions such as nuclei and neostriatum grooming and other behaviors characterized by repeated In both the pontine nuclei and the neostriatum, the patterns of automatic, programmed sequences of limb, presence of demarcated and topographically organized af- head, and facial movements (Cromwell and Berridge, ferent projections has been interpreted to represent mul- 1996; Aldridge and Berridge, 1998; Nambu, 2004; Takak- tiple, segregated channels for information originating in usaki et al., 2004), and secondarily with execution of other different cortical areas (Alexander et al., 1986, 1990; sensory and cognitive tasks (Brown et al., 1997). Bjaalie and Brodal, 1989; Alexander and Crutcher, 1990; Goldman-Rakic and Selemon, 1990; Hoover and Strick, Overlapping corticopontine projections 1993; Schwarz and Thier, 1999; Middleton and Strick, Corticopontine terminal fields show a clear tendency for 2000). But clustering of neuronal elements has also been partial overlap when the distance between two tracer in- taken as a substrate for interaction among multiple clus- jection sites in the cerebral cortex is shorter than 600 ␮m ters in a combinational map (Malach, 1994; Brown et al., (Schwarz and Mo¨ck, 2001, see also Schwarz and Thier, 1998). Indeed, a small cortical site may project to separate 1995, and Table 1 in Leergaard et al., 2000a). This sepa- territories within the pontine nuclei or neostriatum, and ration distance is approximately the same as the extent of the capacity for integrating unique combinations of corti- the dendritic arbors of cortical layer 5 neurons (the neu- cal inputs may occur differentially within these two brain rons of origin of the pontine projections), indicating that regions even though the cortical projection patterns to populations of cerebral cortical neurons with spatially both of these structures are organized topographically overlapping dendritic trees have overlapping terminal (see, e.g., Leergaard et al., 2000b). To further illustrate fields of labeling in the pontine nuclei (Schwarz and Mo¨ck, this point, the shape and extent of the dendritic trees of 2001). The assumption that nonoverlapping populations of the receiving neurons obviously will determine the possi- cortical neurons have nonoverlapping projections does not bilities for integration within such a framework. apply, however, because (1) some combinations of dual Specific combinations of overlapping projections from injections within SI cortex, even with injection separations different cortical areas have been identified in the neo- well above 600 ␮m, generate overlapping terminal fields striatum (Flaherty and Graybiel, 1991, 1993, 1995; (Leergaard et al., 2000a); and (2) injections in homeotopic Parthasarathy et al., 1992; Inase et al., 1996) and in the representations of SI and SII cortex, separated by as much less-studied pontine nuclei (as discussed above). Judging as 1,700 ␮m, generate considerable overlap between the two from the histological appearance of the overlapping termi- projections (Figs 5, 6, 7K–L and 7N–O; Table 1). nal fields, the profusely branched, intermingled, and often Prior work has emphasized segregation between corti- dense terminal fields of axons appear likely to contact the copontine projections originating in different cortical vi- same receiving neurons in the overlapping area (Fig. 3; sual areas of the cat (Bjaalie and Brodal, 1989). Neverthe- see also Alloway et al., 2000, their Figs. 3 and 5; Hoffer less, certain combinations of widely separated cortical and Alloway, 2001, their Figs. 2 and 6). The assumption sites indicate the presence of highly overlapping projec- that individual neostriatal neurons receive convergent tions to the pontine nuclei (Fig. 6, case 373 in Bjaalie and synaptic input from both motor and somatosensory corti- Brodal, 1989). After the present discovery of overlap re- ces was recently supported by a correlated light and elec- sulting from injections in homeotopic representations of tron microscopy analysis (Ramanathan et al., 2002). Thus, two cortical somatosensory areas, we reviewed the loca- labeled corticopontine overlap detected by light micros- tion of the cortical injection sites in Bjaalie and Brodal copy is also likely to reflect pontine convergence and inte- (1989) and compared them with functional maps of visual gration of information from the two (or more) cortical sites cortex (Rosenquist, 1985). We found that the sites in vi- that were injected. Functionally, it is likely that such sual cortex that gave rise to considerable overlap in the corticopontine convergence would be important for deter- pontine nuclei were located in homeotopic representations mining the firing properties of the receiving neurons and (both sites were located in the lower central visual field that the categories of inputs that converge would deter- representations of their respective areas). mine the modality and other activity patterns in the re- Many earlier studies have emphasized the possibility of spective regions. We surmise that specific combinations of overlapping projections from specific sites of the cerebral convergence could be suitable for transferring information cortex. Whereas single-tracer anterograde tracing experi- about selected behavioral states in the basal ganglia and ments have given the impression that the majority of the cerebellar circuits. The differences of organization be- pontine projection fields from different cortical areas are tween the basal ganglia and pontocerebellar systems, as separate or only slightly overlapping, the potential for pointed out by the present study, may reflect important overlap has been suggested repeatedly (Brodal, 1968; differences in the functional role of these systems, with Wiesendanger and Wiesendanger, 1982b; Mihailoff et al., the cerebellar system being more sensory oriented and SENSORIMOTOR CORTICOPONTINE PROJECTIONS 321 having a larger capacity for integrating information cat. In: King JS, editor. New concepts in cerebellar neurobiology. New within the same sensory modality. York: Alan R. Liss, Inc. p 151–182. Brodal P, Bjaalie JG. 1992. Organization of the pontine nuclei. Neurosci Res 13:83–118. ACKNOWLEDGMENTS Brodal P, Bjaalie JG. 1997. Salient anatomic features of the cortico-pontocerebellar pathway. Prog Brain Res 114:227–249. We thank Anna Torbjørg Bore for expert technical as- Brodal P, Bjaalie JG, Aas JE. 1991. Organization of cingulo-ponto- sistance with the preparation of the histological material. cerebellar connections in the cat. Anat Embryol (Berl) 184:245–254. Brown LL, Schneider JS, Lidsky TI. 1997. Sensory and cognitive functions of the basal ganglia. Curr Opin Neurobiol 7:157–163. LITERATURE CITED Brown LL, Smith DM, Goldbloom LM. 1998. Organizing principles of Aldridge JW, Berridge KC. 1998. Coding of serial order by neostriatal cortical integration in the rat neostriatum: corticostriate map of the neurons: a “natural action” approach to movement sequence. J Neuro- body surface is an ordered lattice of curved laminae and radial points. sci 18:2777–2787. J Comp Neurol 392:468–488. Alexander GE, Crutcher MD. 1990. Neural representations of the target Chapin JK, Lin CS. 1984. Mapping the body representation in the SI cortex (goal) of visually guided arm movements in three motor areas of the of anesthetized and awake rats. J Comp Neurol 229:199–213. monkey. J Neurophysiol 64:164–178. Cromwell HC, Berridge KC. 1996. Implementation of action sequences by Alexander GE, DeLong MR, Strick PL. 1986. Parallel organization of a neostriatal site: a lesion mapping study of grooming syntax. J Neu- functionally segregated circuits linking basal ganglia and cortex. Annu rosci 16:3444–3458. Rev Neurosci 9:357–381. Donoghue JP, Wise SP. 1982. The motor cortex of the rat: cytoarchitecture Alexander GE, Crutcher MD, DeLong MR. 1990. Basal ganglia- and microstimulation mapping. J Comp Neurol 212:76–88. thalamocortical circuits: parallel substrates for motor, oculomotor, Fabri M, Burton H. 1991. Ipsilateral cortical connections of primary so- “prefrontal” and “limbic” functions. Prog Brain Res 85:119–146. matic sensory cortex in rats. J Comp Neurol 311:405–424. Allen GI, Tsukahara N. 1974. Cerebrocerebellar communication systems. Flaherty AW, Graybiel AM. 1991. Corticostriatal transformations in the Physiol Rev 54:957–1006. primate somatosensory system. Projections from physiologically Alloway KD, Mutic JJ, Hoover JE. 1998. Divergent corticostriatal projec- mapped body-part representations. J Neurophysiol 66:1249–1263. tions from a single cortical column in the somatosensory cortex of rats. Flaherty AW, Graybiel AM. 1993. Two input systems for body representa- Brain Res 785:341–346. tions in the primate striatal matrix: experimental evidence in the Alloway KD, Crist J, Mutic JJ, Roy SA. 1999. Corticostriatal projections squirrel monkey. J Neurosci 13:1120–1137. from rat barrel cortex have an anisotropic organization that correlates Flaherty AW, Graybiel AM. 1995. Motor and somatosensory corticostriatal with vibrissal whisking behavior. J Neurosci 19:10908–10922. projection magnifications in the squirrel monkey. J Neurophysiol 74: Alloway KD, Mutic JJ, Hoffer ZS, Hoover JE. 2000. Overlapping cortico- 2638–2648. striatal projections from the rodent vibrissal representations in pri- Gao JH, Parsons LM, Bower JM, Xiong J, Li J, Fox PT. 1996. Cerebellum mary and secondary somatosensory cortex. J Comp Neurol 426:51–67. implicated in sensory acquisition and discrimination rather than motor Alloway KD, Hoffer ZS, Hoover JE. 2003. Quantitative comparisons of control. Science 272:545–547. corticothalamic topography within the ventrobasal complex and the Glickstein M, May JG III, Mercier BE. 1985. Corticopontine projection in posterior nucleus of the rodent thalamus. Brain Res 968:54–68. the macaque: the distribution of labeled cortical cells after large injec- Bajo VM, Merchań MA, Malmierca MS, Nodal FR, Bjaalie JG. 1999. tions of horseradish peroxidase in the pontine nuclei. J Comp Neurol Topographic organization of the dorsal nucleus of the lateral lemniscus 235:343–359. in the cat. J Comp Neurol 407:349–366. Goldman-Rakic PS, Selemon LD. 1990. New frontiers in basal ganglia Berg BG, Almaas TJ, Bjaalie JG, Mustaparta H. 1998. The macroglomeru- research. Trends Neurosci 13:241–244. lar complex of the antennal lobe in the tobacco budworm moth He- Graybiel AM, Kimura M. 1995. Adaptive neural networks in the basal liothis virescens: specified subdivision in four compartments according to information about biologically significant compounds. J Comp ganglia. In: Houk JC, Davis JL, Beiser DG, editors. Models of informa- Physiol A 183:669–682. tion processing in the basal ganglia. Cambridge, MA: MIT Press. p 103–116. Bjaalie JG, Brodal P. 1989. Visual pathways to the cerebellum: segregation in the pontine nuclei of terminal fields from different visual cortical Hall RD, Lindholm EP. 1974. Organization of motor and somatosensory areas in the cat. Neuroscience 29:95–107. neocortex in the albino rat. Brain Res 66:23–38. Bjaalie JG, Leergaard TB. 2000. Functions of the pontine nuclei in cerebro- Hartmann-von Monakow K, Akert K, Ku¨ nzle H. 1981. Projection of precerebellar communication. Trends Neurosci 23:152–153. central, premotor and prefrontal cortex to the basilar pontine grey and to nucleus reticularis tegmenti pontis in the monkey (Macaca fascicu- Bjaalie JG, Daehlen M., Stensby TV. 1997a. Surface modelling from bio- laris). Arch Suisses Neurol Neurochir Psychiat 129:189–208. medical data. In: Daehlen M, Tveito A, editors. Numerical methods and software tools in industrial mathematics. Boston: Birkhauser. p 9–26. Heimer L, Zahm DS, Alheid GF. 1995. Basal ganglia. In: Paxinos G, editor. The rat nervous system. San Diego: Academic Press, Inc. p 579–628. Bjaalie JG, Sudbø J, Brodal P. 1997b. Corticopontine terminal fibres form small scale clusters and large scale lamellae in the cat. Neuroreport He SQ, Dum RP, Strick PL. 1993. Topographic organization of corticospi- 8:1651–1655. nal projections from the frontal lobe: motor areas on the lateral surface Bower JM. 1997. Control of sensory data acquisition. Int Rev Neurobiol of the hemisphere. J Neurosci 13:952–980. 41:489–513. Hoffer ZS, Alloway KD. 2001. Organization of corticostriatal projections Brevik A, Leergaard TB, Svanevik M, Bjaalie JG. 2001. Three-dimensional from the vibrissal representations in the primary motor and somato- computerised atlas of the rat brain stem precerebellar system: ap- sensory cortical areas of rodents. J Comp Neurol 439:87–103. proaches for mapping, visualization, and comparison of spatial distri- Hoffer ZS, Hoover JE, Alloway KD. 2003. Sensorimotor corticocortical bution data. Anat Embryol (Berl) 204:319–332. projections from rat barrel cortex have an anisotropic organization that Brodal P. 1968. The corticopontine projection in the cat. I. Demonstration facilitates integration of inputs from whiskers in the same row. J Comp of a somatotopically organized projection from the primary sensorimo- Neurol 466:525–544. tor cortex. Exp Brain Res 5:210–234. Hoover JE, Strick PL. 1993. Multiple output channels in the basal ganglia. Brodal P. 1978. Principles of organization of the monkey corticopontine Science 259:819–821. projection. Brain Res 148:214–218. Hoover JE, Hoffer ZS, Alloway KD. 2003. Projections from primary somato- Brodal A. 1981. Neurological anatomy in relation to clinical medicine. sensory cortex to the neostriatum: the role of somatotopic continuity in Oxford: Oxford University Press. corticostriatal convergence. J Neurophysiol 89:1576–1587. Brodal P. 1983. Principles of organization of the corticopontocerebellar Inase M, Sakai ST, Tanji J. 1996. Overlapping corticostriatal projections projection to crus II in the cat with particular reference to the parietal from the supplementary motor area and the primary motor cortex in cortical areas. Neuroscience 10:621–638. the macaque monkey: an anterograde double labeling study. J Comp Brodal P. 1987. Organization of cerebropontocerebellar connections as Neurol 373:283–296. studied with anterograde and retrograde transport of HRP-WGA in the Ku¨ nzle H. 1975. Bilateral projections from precentral motor cortex to the 322 T.B. LEERGAARD ET AL.

putamen and other parts of the basal ganglia. An autoradiographic cerebellar hemispheres actively support sensory acquisition and dis- study in Macaca fascicularis. Brain Res 88:195–209. crimination rather than motor control. Learn Mem 4:49–62. Ku¨ nzle H. 1977. Projections from the primary somatosensory cortex to Parthasarathy HB, Schall JD, Graybiel AM. 1992. Distributed but conver- basal ganglia and thalamus in the monkey. Exp Brain Res 30:481–492. gent ordering of corticostriatal projections: analysis of the frontal eye Lanciego JL, Wouterlood FG. 1994. Dual anterograde axonal tracing with field and the supplementary eye field in the macaque monkey. J Neu- Phaseolus vulgaris-leucoagglutinin (PHA-L) and biotinylated dextran rosci 12:4468–4488. amine (BDA). Neurosci Protocols 94-050-06-01-13. Potter RF, Ru¨ egg DG, Wiesendanger M. 1978. Responses of neurons of the Leergaard TB. 2003. Clustered and laminar topographic patterns in rat pontine nuclei to stimulation of the sensorimotor, visual and auditory cerebro-pontine pathways. Anat Embryol (Berl) 206:149–162. cortex of rats. Brain Res Bull 3:15–19. Leergaard TB, Bjaalie JG. 1995. Semi-automatic data acquisition for quan- Ramanathan S, Hanley JJ, Deniau JM, Bolam JP. 2002. Synaptic conver- titative neuroanatomy. MicroTrace—computer programme for record- gence of motor and somatosensory cortical afferents onto GABAergic interneurons in the rat striatum. J Neurosci 22:8158–8169. ing of the spatial distribution of neuronal populations. Neurosci Res 22:231–243. Reiner A, Veenman CL, Medina L, Jiao Y, Del Mar N, Honig MG. 2000. Pathway tracing using biotinylated dextran amines. J Neurosci Meth- Leergaard TB, Lakke EA, Bjaalie JG. 1995. Topographical organization in ods 103:23–37. the early postnatal corticopontine projection: a carbocyanine dye and 3-D computer reconstruction study in the rat. J Comp Neurol 361:77– Rosenquist AC. 1985. Connections of visual cortical areas in the cat. In: 94. Peters A, Jones EG, editors. Cerebral cortex. New York: Plenum Press. p 81–117. Leergaard TB, Alloway KD, Mutic JJ, Bjaalie JG. 2000a. Three- dimensional topography of corticopontine projections from rat barrel Ru¨ egg DG, Wiesendanger M. 1975. Corticofugal effects from sensorimotor cortex: correlations with corticostriatal organization. J Neurosci 20: area I and somatosensory area II on neurones of the pontine nuclei in 8474–8484. the cat. J Physiol 247:745–757. Leergaard TB, Lyngstad KA, Thompson JH, Taeymans S, Vos BP, De Ru¨ egg DG, Sequin JJ, Wiesendanger M. 1977. Effects of electrical stimu- Schutter E, Bower JM, Bjaalie JG. 2000b. Rat somatosensory cerebro- lation of somatosensory and motor areas of the cerebral cortex on pontocerebellar pathways: spatial relationships of the somatotopic map neurons of the pontine nuclei in squirrel monkeys. Neuroscience 2:923– of the primary somatosensory cortex are preserved in a three- 927. dimensional clustered pontine map. J Comp Neurol 422:246–266. Ruigrok TJH, Cella F. 1995. Precerebellar nuclei and red nucleus. In: Paxinos G, editor. The rat nervous system. San Diego: Academic Press, Legg CR, Mercier B, Glickstein M. 1989. Corticopontine projection in the Inc. p 277–308. rat: the distribution of labeled cortical cells after large injections of horseradish peroxidase in the pontine nuclei. J Comp Neurol 286:427– Schmahmann JD. 1997. Rediscovery of an early concept. Int Rev Neurobiol 441. 41:3–27. Lillehaug S, O͞ yan D, Leergaard TB, Bjaalie JG. 2002. Comparison of Schmahmann JD. 1998. Dysmetria of thought: clinical consequences of semi-automatic and automatic data acquisition methods for studying cerebellar dysfunction on cognition and affect. Trends Cogn Sci 2:362– three-dimensional distributions of large neuronal populations and ax- 371. onal plexuses. Network 13:343–356. Schmahmann JD, Pandya DN. 1997. The cerebrocerebellar system. Int Rev Malach R. 1994. Cortical columns as devices for maximizing neuronal Neurobiol 41:31–60. diversity. Trends Neurosci 17:101–104. Schwarz C, Mo¨ck M. 2001. Spatial arrangement of cerebro-pontine termi- Malmierca MS, Leergaard TB, Bajo VM, Bjaalie JG, Merchań MA. 1998. nals. J Comp Neurol 435:418–432. Anatomic evidence of a three-dimensional mosaic pattern of tonotopic Schwarz C, Thier P. 1995. Modular organization of the pontine nuclei: organization in the ventral complex of the lateral lemniscus in Cat. dendritic fields of identified pontine projection neurons in the rat re- J Neurosci 18:10603–10618. spect the borders of cortical afferent fields. J Neurosci 15:3475–3489. Merchań MA, Saldan˜ a E, Plaza I. 1994. Dorsal nucleus of the lateral Schwarz C, Thier P. 1999. Binding of signals relevant for action: towards a lemniscus in the rat: concentric organization and tonotopic projection hypothesis of the functional role of the pontine nuclei. Trends Neurosci to the inferior colliculus. J Comp Neurol 342:259–278. 22:443–451. Mercier BE, Legg CR, Glickstein M. 1990. Basal ganglia and cerebellum Selemon LD, Goldman-Rakic PS. 1990. Topographic intermingling of stria- receive different somatosensory information in rats. Proc Natl Acad Sci tonigral and striatopallidal neurons in the rhesus monkey. J Comp U S A 87:4388–4392. Neurol 297:359–376. Middleton FA, Strick PL. 2000. Basal ganglia and cerebellar loops: motor Takakusaki K, Oohinata-Sugimoto J, Saitoh K, Habaguchi T. 2004. Role of and cognitive circuits. Brain Res Rev 31:236–250. basal ganglia-brainstem systems in the control of postural muscle tone and locomotion. Prog Brain Res 143:231–237. Mihailoff GA, McArdle CB, Adams CE. 1981. The cytoarchitecture, cytol- ogy, and synaptic organization of the basilar pontine nuclei in the rat. Torigoe Y, Blanks RH, Precht W. 1986. Anatomical studies on the nucleus I. Nissl and Golgi studies. J Comp Neurol 195:181–201. reticularis tegmenti pontis in the pigmented rat. II. Subcortical afferents demonstrated by the retrograde transport of horseradish peroxi- Mihailoff GA, Lee H, Watt CB, Yates R. 1985. Projections to the basilar dase. J Comp Neurol 243:88–105. pontine nuclei from face sensory and motor regions of the cerebral cortex in the rat. J Comp Neurol 237:251–263. Vassbø K, Nicotra G, Wiberg M, Bjaalie JG. 1999. Monkey somatosensory cerebrocerebellar pathways: uneven densities of corticopontine neu- Mizuno N, Mochizuki K, Akimoto C, Matsushima R, Sasaki K. 1973. rons in different body representations of areas 3b, 1, and 2. J Comp Projections from the parietal cortex to the brain stem nuclei in the cat, Neurol 406:109–128. with special reference to the parietal cerebro-cerebellar system. Welker WI. 1964. Analysis of the sniffing behavior of the albino rat. J Comp Neurol 147:511–522. Behavior 22:223–244. Nambu A. 2004. A new dynamic model of the cortico-basal ganglia loop. Welker C. 1971. Microelectrode delineation of fine grain somatotopic orga- Prog Brain Res 143:461–466. nization of (SmI) cerebral neocortex in albino rat. Brain Res 26:259– Nambu A, Kaneda K, Tokuno H, Takada M. 2002. Organization of corti- 275. costriatal motor inputs in monkey putamen. J Neurophysiol 88:1830– Welker C. 1976. Receptive fields of barrels in the somatosensory neocortex 1842. of the rat. J Comp Neurol 166:173–189. Neafsey EJ, Bold EL, Haas G, Hurley-Gius KM, Quirk G, Sievert CF, Wiesendanger R, Wiesendanger M. 1982a. The corticopontine system in Terreberry RR. 1986. The organization of the rat motor cortex: a the rat. I. Mapping of corticopontine neurons. J Comp Neurol 208:215– microstimulation mapping study. Brain Res 396:77–96. 226. Oka H, Sasaki K, Matsuda Y, Yasuda T, Mizuno N. 1975. Responses of Wiesendanger R, Wiesendanger M. 1982b. The corticopontine system in pontocerebellar neurons to stimulation of the parietal association and the rat. II. The projection pattern. J Comp Neurol 208:227–238. the frontal motor cortices. Brain Res 93:399–407. Wiesendanger R, Wiesendanger M, Ru¨ egg DG. 1979. An anatomical inves- Parent A, Hazrati LN. 1995. Functional anatomy of the basal ganglia. I. tigation of the corticopontine projection in the primate (Macaca fas- The cortico-basal ganglia-thalamo-cortical loop. Brain Res Rev 20:91– cicularis and Saimiri sciureus)—II. The projection from frontal and 127. parental association areas. Neuroscience 4:747–765. Parsons LM, Fox PT. 1997. Sensory and cognitive functions. Int Rev Wong-Riley M. 1979. Changes in the visual system of monocularly sutured Neurobiol 41:255–271. or enucleated cats demonstrable with cytochrome oxidase histochem- Parsons LM, Bower JM, Gao JH, Xiong J, Li J, Fox PT. 1997. Lateral istry. Brain Res 171:11–28.