Intra- and interhemispheric collateral branching in the rat pontocerebellar system, a fluorescence double-label study

Neuroscicwcr Vol. IO. No. I, pp. 141-160, Printed in Great Britain

0306-4522/83 $3.00+0.00 Pergamon Press Ltd Q IBRO

1983

INTRA- AND INTERHEMISPHERIC COLLATERAL BRANCHING IN THE RAT PONTOCEREBELLAR SYSTEM, A FLUORESCENCE DOUBLE-LABEL STUDY G.

Department

of Cell Biology,

A.

MIHAILOFF

The University of Texas Health Science Center, Dallas, TX 75235, U.S.A.

5323 Harry

Hines

Blvd.

Abstract-A double-labeling method which employed the injection and subsequent retrograde transport of two different fluorescent dyes was utilized to investigate the possibility that some pontocerebellar neurons might give rise to collateral branches which distributed to more than one lobule of the contralateral lateral cerebellar hemisphere (intrahemispheric branching) or bilaterally to homotopic or heterotopic lobules in both lateral hemispheres (interhemispheric branching). With regard to intrahemispheric branching, these studies revealed the largest number of double-labeled neurons when the dye injections involved a combination of crus I and the paraflocculus. A considerable number of doublelabeled cells were also observed in simplexxrus II cases while a modest number were noted in crus I-paramedian and simplex-paramedian combinations. Evidence for interhemispheric branching was also apparent but the number of double-labeled cells was generally less than that observed in the intrahemispheric experiments. Following bilateral injections, a modest number of double-labeled cells was noted with simplex~simplex and crus ICrus I homotopic injections and crus II-paramedian heterotopic combinations. In contrast to those cases with unilateral injections, not all bilateral injection combinations produced double-labeling, the most conspicuous in this regard being homotopic and heterotopic injections involving the paraflocculus. Overall, the population of double-labeled cells included many varieties of pontine neurons ranging from small, spindle-shaped to large multipolar. No clear topographic patterns emerged when the location or distribution of either intrahemispheric or interhemispheric double-labeled cells was compared. It was noted, however, that most double-labeled neurons were situated in zones of overlap that occurred between single-labeled neurons projecting to one or the other of the injected lobules. Conversely, in some situations, sharp borders were maintained between adjacent groups of single-labeled neurons. These studies have demonstrated the existence of both intrahemispheric and interhemispheric branching in the pontocerebellar system. The interhemispheric category included pontine neurons which distributed bilaterally to either homotopic or heterotopic lobules in each hemisphere. The overall distribution pattern of double-labeled pontine neurons, those whose axons distribute to more than one hemispheral lobule, is extremely complex and, although some trends were noted, no system-wide topographic organization was apparent. The category of double-labeled neurons included a wide variety of shapes and sizes ranging from the smallest to the largest types of pontine neurons. Most double-labeled neurons were intermixed with single-labeled pontine cells which projected to one or the other of the two injected lobules. No cluster of pontine neurons composed exclusively of double-labeled ceils was ever observed although on some occasions, only one population of single-labeled pontine neuron was interspersed with double-labeled cells.

One of the issues central to an understanding of the pattern of organization in the pontocerebellar system concerns the degree to which pontocerebellar axons might collateralize and distribute to multiple hemispheral lobules on one side of the cerebellum (in trahemispheric branching). Similarly, it is of interest to determine whether pontocerebellar axons might distribute bilaterally to specific locations in both the right and the left cerebellar hemispheres (interhemispheric branching). These two issues, although not the major focus of interest, were addressed in a recent HRP study of the pontocerebellar system.” Although the observations suggested that both intraand interhemispheric branching existed, no conclusions could be drawn since the HRP methodology

cannot accurately define whether or not a given labeled cell actually projects in a collateral fashion to a location other than the area included within the injection site. Comparing the locations of labeled cells across a number of cases with different injection sites might allow one to suggest that collateral branching did, or did not exist, but the definitive resolution of the question could best be achieved through the use of a double-labeling method. Such a method employing the retrograde transport of two different fluorescent dye substances each of which effectively labels either the nuclear or cytoplasmic compartment of the neuron has been developed by Kuypers and colleagues. *’ Briefly, the technique involves the injection of a fluorescent dye into one of the suspected axonal termination sites of the cell group in question and in the same animal, the injection of a second dye substance into another spatially separate (non-overlapping) region also

Abbreviations: FB, Fast Blue; NY, Nuclear Yellow; PI, propidium iodide: BPN, basilar pontine nuclei; NRTP, nucleus reticularis tegmenti pontis. I41

142

G. A. ~ihailoff

thought to receive axon terminals from the same neuron or cell group. The dyes fluoresce at different excitation wavelengths or exhibit different colors within the same excitation spectrum and are selected such that one is ‘a nuclear marker and the other a cytoplasmic marker. Thus, if a neuron provided axon colaterals to both injection sites, it would be doublelabeled, i.e. exhibit both nuclear and cytoplasmic fluorescence. Recently, this methodology was applied to the question of interhemispheric branching of ponto~rebelIar axons in the cat2* The authors reported that “double-1a~led pontine neurons were seen up to 30% in some microscopic fields” following very large injections which involved portions of crus I and crus II in both hemispheres. While it demonstrated that some interhemispheric collaterals did exist, this study was not designed to explore in a systematic manner the specific interhemispheric relationships mediated by the pontine nuclei. In addition, the possible existence of intrahemispheric pontocerebellar branching was not investigated. No other reports in the literature appear to have addressed specifically the issue concerning the possible existence of intra- or interhemispheric branching of pontocerebellar axons. Accordingly the present study, through the use of a double-labeling method which involved combined (but non-overlapping) unilateral fluorescent dye injections in a variety of hemispheral lobules, revealed the presence of some double-ladled pontine cells and thus indicated the existence of intrahemispher~c pontocerebellar collateralization. In a second series of experiments. bilateral dye injections in corresponding (homotopic) and non-corresponding (heterotopic) lobules of the right and left lateral cerebellar hemispheres each demonstrated some double-labeled pontine cells and thus indicat~‘the existence of interhemispheric pontocerebeliar branching. EXPERIMENTAL Two

PROCEDURES

different fluorescent dye combinations were used in

the present $tudy.

In the early phase

of the investigation,

Nuclear Ytillow (NY) was used in conjunction with Propidium iodide (PI). Each dye was prepared as a 3% solution in distilled water and injected via a glass micropipette (30-60fim tip diameter) attached to a 1~1 Hamilton syringe. Small injection sites with a medial-lateral extent ranging from 0.345 mm were achieved with volumes of 0.04-0.06 ~1 injected over a 15-20 min period. Although PI appeared to diffuse mare than NY, the zone of effective uptake and/or transport rate of PI was somewhat iess than NY. This was evidenced by a decreased number of PIlabeled cells relative to the number of NY-labeled cells observed when comparing cases in which injection site locations were reversed with respect to each other but injection volume was nearly equivalent. In addition, the two dyes required different excitation wavelengths (PI, 550 nm; NY, 360 nm) thus n~ssitating a constant alteration be-

tween microscope filter systems which occasionally resulted in considerable difkulty in establishing the identity of double-labeled cells when embedded in a cluster of 3&40 labeled neurons. For these and other reasons, the protocol was changed such that NY was paired with a different cytoplasmic

marker, Fast Blue (FB). The uptake and transport properties of FB and NY were quite similar in our hands with NY perhaps being slightly more effective than FB. Nonetheless. care was taken to reverse the cerebellar dye locations in several of the lobular combinations investigated in order to control for differences in the uptake, transport and labeling capability of the two dye substances. Both dyes were prepared in aqueous solution (5”:,) and pressure injected during the same surgical procedure through a glass micropipette (2&60pm tip) in volumes ranging from 0.04-O. 10~1. In some cases, two injections of the same dye (O.O&U.OS/rl) separated by a distance of approximately 0.5-1.0 mm were made into a single lobule in an attempt to produce a maximal amount of labeling while still restricting the dye to a single lobule. After a survival period which ranged from 20 to 26 h in the successful cases, animals were anesthetized and perfused intracardially with cold 4”,, paraformaldehyde in 0.12 M phosphate buffer. The brain was removed immediately, immersed intact in the same fixative for 24 h then transferred to an equal volume mixture 01‘ fixative and 15:: sucrose-phosphate buffer overnight. On the following morning, the cerebellum (containing the injection sites) was separated from the brainstem and divided into right and left halves by a midline sagittal knife cut. One or both halves of the cerebellum (depending on injection locations) were then sectioned in the sagittal plane at 35 /tm on a freezing microtome and every fourth section collected in phosphate buffer over ice. The brainstem block containing the basilar pons was also sectioned transversely at 35 pm and every section collected in phosphate buffer over ice. The sections were mounted with minimal delay on clean nonsubbed slides, dried thoroughly with gentle heat on a warming table and viewed after coverslipping with Entallen (Merck) or DPX (Gallard-Schlessinger). A Zeiss epifluorescence system consisting of a IOO-watt mercury lamp, the standard UV (360 nm) excitation litter system (4~7702) and a high perfo~man~e interference filter set (487713) with a transmittance range between 500 and 560 nm was used throughout the study. Represented in Fig. I are dorsal (A) and posterior (B) views of the rat cerebellum which illustrate the terminology employed in this study as adopted from Larsell.” A parasagittal section through the lateral cerebellum is shown in (C) to reveal the internal ar~ngemenl of the hem~sphe~~i lobules under investigation while representative injection sites are illustrated in (D) and (E). A total of I69 Long-Evans black-hooded rats (males and females, l4&240 g) were used throughout this study. Of this group of animals. 34 were omitted because of overlapping injection sites. or extension of injected dye beyond the intended lobule. Both NY (Fig. ID) and FB (Fig. 1E) injection sites typically consisted of an intensely fluorescent core which included the molecular, Purkinje and granule ceil layers and occasionally the white matter near the summit of the injected lobule. The central. densely stained area was surrounded by a region where only weak fluorescence was evident. This peripheral zone of weak fluorescence apparently did not constitute a part of the effective uptake zone as judged by the fact that the pattern of pontine labeling was not altered when this region extended into an adjacent lobule. Glial cell iabehng was extensive in the central core area (particularly at NY sites) but decreased considerably in the surrounding region and labeled axons could be seen coursing away from the injected lobule toward the medullary core of the cerebellum. Figure 2 is provided to illustrate the major basilar pontine su~ivisions previously described in the rat” and also utilized in the present study. On the right side of the section (Fig. 28). rectangular areas indicate those pontine regions shown in the color photomicrographs comprising Fig. 3. As demonstrated in Fig. 3(A), NY fluorescence was invariably confined to the cell nucleus and with the survival period employed. no signilicant migration of the dye into

A.

PI -

Fig. 1. Dorsal (A), posterior (B) and parasagittal (C) views of the rat cerebellum which illustrate the terminology employed in describing injection sites. A representative nuclear yellow (NY] injection site in the anterior lobule of crus I is shown in D while a typical fast blue (FB) injection site in the anterior lobule of crus II is illustrated in E. Abbreviations: CrI, crus I; CrIIa, crus II anterior; CrIIp, crus II posterior; Pf, paraflocculus; Pm. paramedian; Sx, simplex (adopted from Larsell 2’).

143

Fig. 3. Representative examples of single-labeled pontine neurons rlrc shorr n rn A and B (NY, nuclear yellow; FB, fast blue). Examples of double-labeled pontine neurons are presented in C-F. Each exhibits intense NY nuclear labeling and moderate to weak FB cytoplasmic fluorescence. Some (E.F) exhibit bright cytoplasmic granules. Several examples of pontine labeling patterns are shown at low magnificabon in G-K. The location of each photomicrograph is indicated in Fig. 2. See text for details.

144

Pontocerebellar

collateral distribution ABREVIATIONS

dl dM dPd lnl

Coil

tat MCP M,d ML MLF NRTP PEO SCP Vent

Fig. 2. Orientation diagrams which illustrate the terminology employed in describing subdivisions of the basilar pontine nuclei (adopted from Mihailoff et aZ.r5).Rectangular areas illustrated on the right side of the section in B indicate the approximate location of the color photomicrographs shown in Fig. 3.

surrounding neuronal or ghal elements was evident. Pontine neurons retrogradely labeled with FB exhibited a brilliant blue cytoplasm (Fig. 3B) with no nuclear labeling. In some, but not all cases, the cytoplasm of FB labeled cells also contained bright pink or silver granules. Rapid quenching of FB fluorescence was a frequent occurrence under high magnification whereas NY fluorescence was quite stable. Double-labeled neurons were easily distinguished by their yellow nucleus and blue cytoplasm (Figs 3C,D). Some double-labeled neurons also exhibited the bright pink or silver granules embedded within blue cytoplasm (Figs 3E,F). The population of single- and double-labeled neurons inchided a heterogenous mixture of sizes and shapes ranging from small round or spindle-shaped cells to large multipolar neurons. Several representative examples of the pattern of pontine labeling observed in various cases are also shown in Fig. 3 (G-K). The labeling pattern demonstrated in G is characteristic of that observed in NRTP following combined injec-

tions which involved crus I and crus II. Cells in the central portion of the nucleus project to crus II and exhibit NY fluorescence while the dorsal, dorsolateral, and particularly the dorsomedial perimeter cells project to crus I and exhibit FB fluorescence. The segregation of pontine cell groups projecting to crus II and the paraflocculus is illustrated in H whereas the overlap and juxtaposition of pontine cells projecting to crus II and the paramedial lobule is demonstrated in I. Note especially the medial pontine region where pontoparamedian cells (FB) are clustered medially but lie immediately adjacent to (with little overlap) a slightly more lateral group of cells (NY) which project to crus II. The same spatial relationship is illustrated in J and K taken from another case involving a combined injection of NY and propidium iodide (PI). The medial group of NY-labeled ponto-paramedian cells is shown in J using the 360nm excitation wavelength while in K the same field is visualized with the 550nm filter to reveal the adjacent group of PI-labeled pontine cells which project to crus II.

C. A. Mihailoff

146 RESULTS

Before reporting the details of findings concerning intrahemispheric pontocerebellar branching, three observations of a general nature are presented below. First, the number of double-labeled pontine cells was quite variable and was dependent upon a number of factors including the combination of lobules injected and of course the volume of dye delivered to the injected lobule The largest number of double-labeled cells was consistently observed when the dye injections involved a combination of crus I and the ipsilateral paraflocculus. A considerable number were also observed in cases with combined injections in lobulus simplex and crus II, whereas only a modest number of double-labeled cells was noted in cases which combined injections in crus I and crus II. Table I below summarizes the relative number of doublelabeled cells observed following each combination of injections. A second observation of general relevance was the fact that not all of the separate pontine cell groups which projected to either of the two injected lobules in a given case exhibited double-labeled cells. If one assumes that this is not simply a technical limitation in the sense that the injected dye was not made available to all the collateralized pontocerebellar axon terminals, it might then be suggested that at least two subsets of pontocerebellar neurons exist, one of which projects to a preferred cerebellar cortical location with very little or no collateral branching to other lobules in the same hemisphere and another group (double-labeled) which does collateralize intrahemispherically to the two injected lobules. The third observation was the finding that in general the population of double-labeled pontine cells included many different varieties of size and shape. Similarly, it was noted that even within individual cases, the subset of double-labeled neurons resulting from combined dye injections could not be categorized as a morphologically unique group of cells but rather included several varieties of size and shape. Table I. Summary of double-labeled pontine neurons observed following the indicated intrahemispheric combina-

tion of injections. Each asterisk (*) is equivalent to approximately 35 doubly-labeled neurons. The black circle (0) indicates that this intrahemispheric category involved an injection in crus IIa combined with a second in crus Hp. The number number

in the lower right corner of each box denotes the of successful cases used in the analysis of that

combination

of injections. The total number of intra-

hemispheric

cases with useful injections

was 57.

The following observations are derived from a number of cases in which two fluorescent dyes were unilaterally injected into different lobules of the same cerebellar hemisphere. All the possible combinations of injections were made and the six cases described below are representative of those which exhibited the largest number of double-labeled cells. Crus I-pur&cculus combination. A large number of double-labeled pontine cells were observed in cases with a combination of dye injections that involved portions of crus I and the paraflocculus (Figs 4,A,B,C). At each site, the spread of injected dye is extensive but does not extend beyond the boundaries of either lobule (Figs 4A,B). Within the basilar pontine gray contralateral to the cerebetlar injections, the double-labeled cells (A) were confined to three regions: a ventrolateral group (arrow 1. Fig. 4C) was present through the rostra1 two-thirds of the pons; a more medial group (arrow 2, Fig. 4C) consisted of double-labeled cells situated near the ventral surface of the pons adjacent to, and among the fibers comprising the brachium pontis; a third group (arrow 3, Fig. 4C) located dorsally in the rostra1 one-half of the pontine gray in an area where the dorsal peduncular nucleus of the pons is in continuity with a ventrolateral extension of cells forming a part of the nucleus reticularis tegmenti pontis (NRTP). In each of these regions, the double-labeled ceils were situated in zones of overlap which occurred between clusters of single-labeled pontine cells projecting to either crus I or the para~occulus. However. not all zones of overlap exhibited double-labeled cells. Moreover, never was a single, isolated cluster of cells composed exclusively of double-labeled neurons, nor were any double-labeled cells observed in the pontine gray ipsilateral to the injections, or in the central or dorsomedial portions of NRTP on either side. Puramedian-par~~o~Lulus combination. In general, the number of double-labeled cells noted with this combination of dye injections was considerably less than that observed in the preceding (crus I-paraflocculus) combination. The injection sites again were purposely quite extensive but did not extend beyond the limits of the lobules in question (Figs 4D,E,F). The primary locus of double-labeled pontine neurons was restricted to a rather thin lamina of cells near the central surface of the pons (arrows, Fig. 4F). As in the preceding crus I~para~~culus combination, the location of the double-labeled cells corresponded to a zone of overlap that occurred between pontine cell groups projecting to both the paramedian lobule and the paraflocculus. Although single-labelled neurons (from either dye injection) were observed in the ipsilateral pontine gray, no double-labeled cells were noted there, nor were they observed in the NRTP on either side. Crus II @osterior) ~pur~~~edi~~n ~~j~~bi~Qt~~n. This combination represents a situation where immediately adjacent cerebellar lobules were injected (Figs SA,B). It was thus imperative to prevent mixing of

Pontocerebellar

- Crl

l

Pf

A Double

collateral

141

distribution

-Pm lPf

ADouble

Fig. 4. Two cases illustrating the pontine labeling resulting from unilateral intrahemispheric injections which involved crus I-paraflocculus (A,B,C) and paramedian-paraflocculus (D,E,F). Labeling scheme: single-labeled cells (.) single-labeled cells (0); double-labeled cells (A). See text for details.

148

G.A. Mihailoff

A. , Crllp-Pm B.~

~'~-- CrtE"

Crllp

Pm

~

~

X

-- ~"'~ • Pm -Crllp

C.

~

~L~

I

Pm

X

'~Double

- Crl

oCrllp

,~Double

F. ,;

2

~

-

-

~

..

..

,

ji

""

3

'

Fig. 5. Two cases illustrating the pontine labeling resulting from unilateral intrahemispheric injections which involved crus II posterior-paramedian (A,B,C) and crus I-crus II posterior (D,E,F). Labeling scheme: single-labeled cells (.); single-labeled cells (O); double-labeled cells (A). See text for details.

~

J

Ponto~re~ilar

collateral dist~bution

the two dyes at the injection sites in order to preclude the production of false-positive double-labeling if both dyes were made available to the population of pontocerebellar (mossy) axon terminals in both labules. Within the pontine gray, four different regions contralateral to the injected hemisphere contained double-labeled cells, while a fifth region was observed within the central portion of the ipsilateral NRTP. The four pontine regions were restricted to the middle one-third of the pons, and, as in preceding combinations, the double-labeled cells were interspersed among single-labeled cells in zones where overlap occurred between cellular aggregates projecting to either crus IIp or the paramedian lobule. These areas included a medial pontine region (arrow 1, Fig. SC), a zone near the ventral pontine perimeter (arrow 2, Fig. 5C), an area in the mid-ventral region (arrow 3, Fig. SC), and the dorsomedial pontine area near the medial lemniscus (arrow 4, Fig. SC). The central portion of NRTP ipsilateral to the injections which exhibits double-labeled cells (arrow 5, Fig. 5C), also corresponds to an area containing single-labeled cells from either injection. It is also of interest to note the distinct segregation of, and the clear boundaries between, the single-labeled cells in the medial and lateral pontine aggregates which project to either lobule (open arrows, section 5, Fig. SC). Crus I-Crus II (posterior) combination. Fluorescent dye injections restricted to crus I and posterior lobule of crus II (Fig. 5DE) produced doublelabelling of cells in two regions of the rostra1 one-half of the contralateral pontine gray. The most rostra1 group of double-labeled cells was present in the dorsal peduncular region of the pons which is inserted between the medial lemniscus dorsally and the cerebral peduncle ventrally (arrow 1, Fig. 5F). This aggregate of double-labeled cells was unique in that it was the only group observed in a pontine zone which did not represent a site of overlap of those pontocerebellar neurons projecting to the injected lobules, that is, in this location, only two (rather than three) populations of labeled cells were present, those single-labeled neurons which apparently projected to crus I and the double-labeled cells which collateralized and distributed to both crus I and crus II posterior. There were no cells in this region which were single-labeled by the crus II posterior injection. This is in contrast to all of the previously noted groups of double-labeied cells where, in addition to the double-labeled neurons, two populations of single-labeled cells, were present, each of which projected separately to the two injected lobules. The latter type of organization was evident, however, in the second group of double-labeled cells in this series (arrow 2, Fig. 5F). This extreme ventral region of the pontine gray, in addition to the double-labeled ceils, also exhibited single-labeled cells which projected to either crus I or crus II posterior. Crus II (anterior)-crus

II (posterior) combination.

149

This combination of injections represents another situation involving adjacent lobules, however, unlike the crus II posterior-paramedian cases, it was less difficult to restrict the spread of injected dye into the neighboring lobule because the more dorsal position of these lobules made them much more accessible at the time of surgery. Although there is considerable overlap in the location of pontine cells projecting to crus IIa and crus IIp, nonetheless, there are clusters of single-labeled cells projecting to either lobule which are intermingled in some areas (left side, sections 3-5, Fig. 6C) and largely segregated in other regions (open arrows, Fig. 6C). Double-labeled neurons were observed within overlap zones at rostra1 and mid pontine levels in the central portion of the contralateral ventral pontine region (arrow 1, Fig. 6C) Somewhat more caudally, double-labeled cells were present bilaterally in the dorsomedial area (arrow 2) and two mirror-image groups in medial and ventrolateral pontine regions (arrows 3 and 4) which are characteristic of the pontocerebellar projections originating from crus II. 25 Double-labeled neurons were also present biiaterally in the caudal one-half of NRTP (arrow 5, Fig. 6C). Crus I-parumedian corn~j~~tjo~. Relatively little overlap exists in the location of pontine cell groups projecting to crus I and the paramedian lobule. However, where overlap did exist, double-labeled cells were present. These included the dorsolateral region (arrow 1, Fig. 6F) and scattered neurons near the ventral surface of the ponting gray (arrow 2, Fig. 6F). The double-labeling was bilateral in the dorsolateral area but only ipsilateral in the ventral location. No double-labeled cells were observed in NRTP. Lob&s simplex-crus II (anterior) comb~nution. In this combination of injections, the location of doublelabeled cells was of particular interest because such neurons were distributed in the pontine gray both ipsilateral and contralateral to the injected hemisphere. This is evident in Fig. 7(C), where arrow 1 (sections 2, 3, and 4) indicates the location of doublelabeled cells (ipsilateral to the injections} among a mixture of single-labeled cells, the majority of which project to simplex. The number of ipsilateral doublelabeled cells was sparse, however, in comparison with the larger number present in the contralateral pontine gray as indicated by arrow 2 in Fig. 7(C). Similar to the preceding cases, the contralateral double-labeled cells were located for the most part near the ventral surface of the pontine gray although some were situated in the ventral peduncular region (arrow 3, Fig. 7C) near a cleft in the peduncle which allows continuity between dorsal and ventral pontine regions. Athough in all instances the doubIe-labeled cells were interspersed among an overlapping cluster of single-labeled neurons projecting to either simplex or crus II anterior, there were other overlap zones in the dorsomedial pontine region and in NRTP (sec-

150

G . A . Mihailoff

A.

Crlla- Crllp

B.

D.

Crl - Pm

i Crl~

C ~ ~ ~ C~pI

E~

l • Crlla

.Crllp

c.

x

~Double

~ . 1

"'---:Z__~ j. .,.

_ - ~ _5 ~.

<;>

F. ~_

2~

- ~ . - ~ _ ~

Pm

• Crl

~ 1

k\

3~

/

j

) 5

5J'f //i/

6

i{

a

Fig. 6. Two cases illustrating the pontine labeling resulting from unilateral intrahemispheric injections which involved crus II anterior-crus II posterior (A,B,C) and crus I-paramedian (D,E,F). Labeling scheme: single-labeled cells (.); single-labeled cells (O); double-labeled cells (A). See text for details.

Ponto~rebcllar

151

collateral dist~bution

Sx-Cdl

. Sx

~SX

lCrlle~Double

F.

*Pm aDouble

c)

,’ -..-.~__. ’

1

Fig. 7. Two cases illustrating the pontine labeling resulting from unilateral intrahemispheric injections which involved simplex-crus II (A,B,C) and simplex-paramedian (D,E,F). Labeling scheme: single-labeled cells (.); single-labeled cells (0); double-labeled cells (A). See text for details.

G. A. Mihailoff

IS2

tions 4 and 5, Fig. 7C) where no double-labeled cells were apparent. Lobulus simplex-paramedian combination. The case selected to represent this combination of injections (Figs 7D,E,F) exhibited two groups of double-labeled cells, each of which was situated contralateral to the injected hemisphere. The largest number of doublelabeled cells were located in the central portion of the ventral pontine region (arrow 1, Fig. 7F) an area which also contained a mixture of single-labeled cells projecting to either the simplex or paramedian lobule. In addition, double-labeled cells were observed in the ventral peduncular region (arrow 2, Fig. 7F) an area which also exhibited double-labeled cells in the previous combination involving simplex and crus II anterior. In spite of the fact that simplex and the medial portion of the paramedian lobule each receive considerably more ipsilateral pontocerebellar input than many of the other hemispheral lobules as evidenced in this study (sections 2-5. Fig. 7F) and in a previous report,*’ no ipsilateral double-labeled cells were noted in this, or any other case with a similar combination of injected lobules. Interhemispheric branching This portion of the study was concerned with the question of whether individual pontocerebellar axons collateralize to provide input to both (left and right) cerebellar hemispheres. Further, it was of interest to determine, if interhemispheric branching did exist, whether or not such interconnections involved homotopic (corresponding) and/or heterotopic (noncorresponding) hemispheral lobules. Table 2 and the following series of cases illustrate that both homotopic and heterotopic pontocerebellar branching do exist but the richness of the collateral branching varies from lobule to lobule. Paramedian-paramedian combination. As indicated in Fig. 8(A), the injections in this case involved the medial portion of the left and the right paramedian lobule. The location of single-labeled pontine neurons illustrated in B of Fig. 8 corresponds to the pattern described previously in those cases with a unilateral injection in the paramedian lobule. A Table 2. Summary of double-labeled pontine neurons observed following the indicated interhemispheric homotopic or heterotopic combinationof injections. Each asterisk is equivalent to approximately 45 double-labeled neurons. The number in the lower right corner of each box denotes the number of successful cases used in the analysis of that combination of injections. The total number of interhemispheric cases with useful injections was 78. s,

Crl

C,IIa

Crllp

Pn

Pf 1

modest number of double-labeled pontine neurons were observed in several locations throughout the cuadal two-thirds of the basilar pons. The more rostra1 double-labeled cells were situated bilaterally in the dorsal and ventral peduncular regions (Fig. 8B, sections 4, 5). At more caudal levels, double-labeled cells were present in medial and ventrolateral pontine regions Fig. 8B, section 7). Single and double-labeled cells were also present bilaterally in the central portion of NRTP. Other cases with larger and more laterally placed injections did not exhibit an increase in the ratio of double to single-labeled pontine (or NRTP) neurons, nor was there a difference in the distribution pattern of the labeled neurons. Crus I-crus I combination. In this case (Fig. 8C) two dye placements were made in crus I in the right and the left hemisphere in an attempt to label maxima!ly any branching pontocerebellar axons distributing to both lobules. Despite the rather extensive injection sites, once again, as evidenced in D of Fig. 7, only a modest number of double-labeled neurons were observed. Although some were clearly located in the pontine nuclei, many were situated in various regions of NRTP. Similar findings were noted in cases where the dye placements involved different portions of crus I. Crus II-crus II combination. This group of cases was divided into two categories, those with homotopic injections in the anterior lobule of crus II, and those with dye placements involving the posterior lobule or crus II in each hemisphere. The findings were quite similar in the two groups and overall each case exhibited more double-labeled cells than either of the previous combinations involving the paramedian lobule or crus I. In the anterior lobule cases, double-labeled cells were frequently present in the dorsal peduncular region of the pons (open arrows. Fig. 9B, section 5) but were not observed in that location in posterior lobule cases. In posterior lobule cases, double-labeled cells were numerous within an oval cluster of cells in a medial pontine region (solid arrow, Fig. 9D, section 5) which has been shown in previous studies to be a characteristic source of axons projecting to crus II. Double-labeled cells were less frequently encountered in this region in anterior lobule cases. With regard to NRTP, double-labeled cells were distributed bilaterally in nearly equivalent numbers and similar locations following anterior or posterior lobule dye placements. Simplex-simplex combinations. The injection sites in lobulus simplex (Fig. IOA) were of necessity positioned within the more intermediate or medial portion of the posterior lobule since the anterior and lateral regions were largely inaccessible due to their location beneath the transverse sinus. Double-labeled neurons were quite numerous bilaterally in the mid ventral pontine region, beginning at mid-rostrocaudal levels and extending nearly to the caudal limit of the pontine gray (Fig. IOB, sections 46). A

Pontocerebellar

APm 7 A.a

.d\

vs

collateral distribution

153

Pm0 z’ 1: ’ Pm

ONY

AFB l

Double

Fig. 8. Two cases illustrating the pontine labeling resulting from bilateral homotopic injections which involved the paramedian lobules (A,B) and crus I (C,D). Labeling scheme: single-labeled cells (0); singIe-ladled cells (a); double-labeled cells (.). See text for details.

considerable number of double-labeled cells were also observed in NRTP. They were present bilaterally in the central portion of the nucleus just dorsal to the medial Iemniscus and extended for some distance through the caudal one-half of the pontine gray and slightly beyond (Fig. lOB, sections S-7). Having determined that homotopic (correspond-

ing) lobules of the right and left cerebellar hemispheres receive branched axonal projections from single basilar pontine neurons, the possibility that pontine neurons might project in collateral fashion to heterotopic (non-corresponding) cerebellar hemispheral lobules was investigated. The following five cases will describe the projection pattern observed in

154

A.

G. A. Mihailoff

A

Cr

II, vsCr iI0 0

m Y--P

-

/---f’

i’,, .’

,-

__--

’

A

._.-a

a

u

.

0

Cr lla

o NY A

F8

l

Double

u D.

B.

Fig. 9. Two cases illustrating the pontine labeling resulting from bilateral homotopic injections which involved crus II anterior (A,B) and crus II posterior (C,D). Labeling scheme: single-labeled cells (0); single-labeled cells (A); double-labeled cells (.). See text f’or details.

Pontocerebellar collateral distribution OSX

vs

A Sx vs Crllpo

SXA

NY FB o Double

0

A

Fig. 10. Two cases illustrating the pontine labeling resulting from bilateral homotopic (simplex-simplex. A and B) or heterotopic (simplex-crus II posterior, C and D) injections. Labeling scheme: singie-~~~~ cells (A); single-labeled cells (0); double-labeled cells (.). See text for details.

those animals which received bilateral dye injections in heterotopic hemispheral locations and clearly exhibited evidence of collateral branching. Simplex-crus II posterior combination. The dye injections in this case involved the intermediate portion of lobulus simplex in the left hemisphere and the laterai region of crus II posterior in the right hemisphere (Fig. IOC). Single-labeled neurons from either

injection were distributed bilaterally throughout much of the caudal two-thirds of the pontine gray. Double-labeled neurons were also evident bilaterally and were located in mid-ventral or ventral peduncular regions scattered among single-labeled neurons in those areas (Fig. IOD, sections 4-6). A few doublelabeled neurons were also present bilaterally in the central

portion

of NRTP

immediately

dorsal

to the

156

A.

/ / m 0Pm

/

-

vs

vs

Cr lla 0

*

Crl

Pm

* Pm

CrlA

A’

r

G. A. Mihailoff

oNY A t-‘B . Double

Fig. 11. Two cases i~~ustrafing the pontine labeling resulting from bilateral heterotopic injections which involved param~ian~~s I (A,B) and paramedian-crus II anterior (C,D). Labeling scheme: single-labeled cells (Of; single-labeled cells (A); double-labeled cells (,). See text for details.

medial lemniscus (Fig. NRTP and the pontine clusters or aggregates only of double-labeled

IOD, sections 4 and 6). In both

gray there were no discernible of cells that were comprised neurons. Paramedian-crus I combimtim Double-tabeled neurons were bilaterally distributed within four primary pontine locations. These included the dor-

solateral area (Fig. I IB, section 4), a ventrolateral zone (Fig. 1lB, sections 3-5) as well as dorsomedial and dorsal peduncular regions (Fig. 11B, section 5). No double-labeled cells were observed in the NRTP in this, or any other cases with this injection site combination. Further, when the dye substance injected at each location was reversed, there was no

Ponto~rebellar

A.

0 Cr

II,

vs

157

collateral dist~bution

Cr Ilp A

o Pm VS Crlip A C

.

ONY AFB

Fig. 12. Two cases illustrating the pontine labeling resulting from bilateral heteotopic injections which involved crus II anterior-crus II posterior (A,B) and paramedian-Crus II posterior (C,D). Labeling

scheme: single-labeled cells (0); single-labeled cells (A); double-labeled cells (.). See text for details. change in the number labeled cells.

or distribution

of doubie-

Par~edia~-cry II anterior ~orn~i~ti5~. Considerably more double-labeled neurons were observed in this group of cases than in the preceding combination. The majority were located in the medial and lateral aspects of the ventral pontine region (Fig.

IlD, sections 4, 5) at mid rostro-caudal levels. Double-ladled neurons were also present in the dorsal peduncular region and in NRTP. The bilateral distribution of the double-labeled neurons indicates that the pontine neurons with branched axons are more or less randomly dispersed on both sides of the midline.

G. A. Mihailoff

158

Crus II anterior-crus II posterior combination. As indicated in Fig. 12(A), this is an example of a case in which two dye placements were made in each lobule in an attempt to visualize maximally the complement of bilaterally branching pontocerebellar neurons. The two injections in each lobule were spaced far enough apart to prevent accumulation of dye in immediately adjacent lobules. Double-labeled neurons were distributed bilaterally in several locations at mid pontine levels. The majority were located in two areas, the medial and lateral portions of the ventral pontine gray. which represent the two characteristic cell groups projecting to crus II (Fig. l2B, section 4, arrows 1 and 2). In addition, several double-labeled cells were noted in the dorsal peduncular zone and the dorsomedial pontine region between the midline and the medial lemniscus. The rostra1 portion of the NRTP also exhibited doublelabeled neurons located biiaterahy in the area just dorsal to the medial lemniscus (Fig. t2B, sections 43). P~ru~edia~-cru.~ II p~~steri~~r co~nt~~~ut~(~~, In this particular case, the dye injections were placed in the medial portion of both the left pardmedian lobule and the right posterior lobule of crus II. Doublelabeled neurons were present bilaterally in moderate numbers at mid-rostro-caudal levels mainly in ventral and ventrolateral pontine regions (Fig. l2D, sections 4-6). Other double-labeled cells were located in the dorsal peduncular nucleus (arrow 1) and the medial peduncular region (arrow 2). Double-labeled neurons were also evident bilaterally in NRTP.

based upon

HRP studies

which have demonstrated

the bilateral nature of the pontocerebellar system in the primate,” cat,6~‘~i7~‘x~“‘~‘Y opossum” and rat.4.‘0.‘4.‘5.‘h.2hThese findings are also in agreement with an earlier degeneration study which reported the existence of a bilateral component in the pontocerebellar systems of the rabbit and cat.’ The question that arises. however. concerns the underlying pattern of organization among pontocerebellar neurons which might account for this apparent bilaterality. Could it be that within any given pontine region there exists a mixture of pontocerebeilar neurons with some projecting only to the left cerebellar hemisphere while other, immediately adjacent cells project only to the right hemisphere? Or might the observed bilaterality be due to the fact that some pontocerebellar neurons give rise to divergent axon collaterals which branch in a bilateral manner to reach the left and right cerebellar hemispheres? This issue has been addressed in a preliminary manner in the cat where, on the basis of fluorescent double-label studies, it was shown that double-labeled neurons were present in the BPN following large bilateral dye injections which involved portions of crus I and crus 11 in each lateral hemisphere.” The implication of these findings is that at least some BPN neurons give rise to axons which distribute to both the right and left cerebellar hemispheres. The present report has confirmed and extended these initial observations by demonstrating that pontocerebellar axons distribute bilaterally to homotopic as well as heterotopic lobules and further. that pontine neurons project unilaterally in a divergent manner to more than one lobule in a given lateral hemisphere.

DISCUSSION

The present study has sought to investigate the issue of intrahemispheric and interhemispheric branching of cerebellar afferent axons arising from neurons located within the basilar pontine nuclei (BPN) and the nucleus reticularis tegmenti pontis (NRTP). Combined (but non-overlapping) unilateral fluorescent dye injections in various lobules of the lateral cerebellum revealed the presence of doublelabeled BPN and NRTP cells although the number of such neurons was somewhat less than had been anticipated on the basis of previous HRP studies of the pontocerebellar system.26 Similarly, bilateral dye injections in corresponding (homotopic) and noncorresponding (heterotopic) lobules of the left and right hemispheres also revealed double-labeled cells in the BPN and NRTP. These observations have. therefore, been interpreted as supportive of the suggestion that some individual pontocerebellar neurons give rise to axons which might collateralize to more than one lobule in the same hemisphere (intrahemispheric branching) or distribute bilaterally to lobules in both the right and left lateral hemispheres (interhemispheric branching). The existing literature contains several reports

Double-labeled neurons were observed in greater or lesser numbers in all of the combinations involving two unilateral cerebellar dye injections. Primarily because of a variety of factors which made it difficult to deliver precisely the same quantity of dye at all injection sites, a comparison of the absolute number of double-labeled neurons observed in the various combinations has not been reported. Nonetheless, a careful attempt was made to represent the relative number of double-labeled cells in each combination. The results presented in Table I indicate that the simplex-crus 11 and crus Icombinations paraflocculus consistently produced the largest number of double-labeled cells. However, no clear pattern was apparent in comparing the number or location of double-Iabeled cells observed in all the various combinations of dye injections, that is. there was no overt topography in the projection of pontocerebehar axon collaterals to lobules within the same hemisphere. Similarly. there was no discernible pattern with regard to the location of double-labeled neurons in the BPN or NRTP although it was noted that. with the exception of the crus Ilp--crus I combination, doublclabeled BPN cells were most frequently tocated in

Pontocerebellar

areas where overlap existed between pontine cell groups projecting to the injected lobules. Therefore, one might suggest that some of the double-labeling might have been due to diffusion or migration of the dyes between the two types of single-labeled neurons present in the overlap zone. However, there was no visible evidence of outward migration of the dye, i.e. no faint halo around somata nor was there any glial cell labeling and, considering the relatively short survival time (24 h or less, in most cases), it seems unlikely that much of the double-labeling represents a false-positive observation. Moreover, in many instances, sharp borders were maintained between adjacent clusters of neurons singly labeled with either of the two injected dyes. Another interesting observation with regard to intrahemispheric branching was the fact that most double-labeled cells in the BPN were located contralateral to the injected hemisphere. An exception to this trend was noted in simplexxrus IIa and crus IIaxrus IIp combinations, where a considerable number of ipsilateral double-labeled BPN neurons were evident. Curiously these two lobules (simplex and crus II) exhibited the largest ipsilateral component in HRP studies of the rat pontocerebellar system.*’ Conversely, in NRTP the distribution of double-labeled neurons was typically bilateral. Interhemispheric

159

collateral distribution

branching

As indicated in Table 2, unlike the unilateral injections, not all combinations of bilateral cerebellar dye injections produced double-labeling of BPN or NRTP neurons. Moreover, in those combinations which did produce positive findings (doublelabeling), the number of double-labeled cells generally was proportionately lower than that observed after unilateral injections. These observations correlate well with electrophysiological studies in the rat which describe a multiple, patch-like or mosaic somatotopic organizational pattern for the inputs from somatosensory cortex’ and tactile periphery’O to the lateral cerebellar hemispheres. According to the studies of Bower et al.,’ a representation of the body surface (primarily perioral regions) can be demonstrated in the paramedian lobule, both lobules comprising crus II, crus I, and lobulus simplex following microstimulation of the contralateral somatosensory cortex. For example, stimulation of the cortical representation for the animal’s upper lip produces a response in the contralateral paramedian lobule, both lobules of crus II, crus I and lobulus simplex. The most likely anatomic substrate for this finding would involve transmission of the signal from somatosensory cortex to the ipsilateral BPN, the cells of which then project contralaterally to the paramedian, crus II, crus I, and simplex lobules. The results of the present double-label study suggest that a considerable number of individual pontine cells might project divergent collateral branches to at least two, if not all the cerebellar lobules in question. Moreover, a small

number of ipsilateral receptive fields were evident in each lobule and this correlates with the finding from the present study suggesting that fewer pontine cells branch in an interhemispheric manner compared to the number which collateralize intrahemispherically. Following homotopic injections (corresponding lobule in each hemisphere), the largest number of double-labeled cells was observed in simplex or crus I cases while in heterotopic cases (non-corresponding lobules), the greatest number of double-labeled cells was apparent when a paramedian injection was paired with either a crus IIa or a crus IIp injection. It should be noted that overall, there was no clear difference between the number of homotopically and heterotopically double-labeled cells. An earlier study which examined interhemispheric pontocerebellar branching28 reported that “doublelabeled cells were seen up to 30% in single microscope fields.” If one interprets this statement to mean that 30% of all cells in the field were double-labeled, then that number would be quite disparate relative to the findings in the present study. If, however, one interprets the number given by Rosina et a/.28 as 30% of the labeled cells in the field to be double-labeled, that number would be quite similar to those of the present study. Rosina et a1.28 also indicated that the population of double-labeled cells included only small to medium sized pontine neurons, whereas in the present study, all sizes of pontine neurons, including the largest variety, were among the double-labeled population in one case or another. Possible ,finctional

implications

If one compares the location of the projection fields of all the known pontine afferent systems with the location of the double-labeled cells identified in the present study, no clear connectional relationships emerge. This perhaps is largely a result of the wide diversity of afferent inputs to the pontine nuclei coupled with the diffuse distribution of doublelabeled neurons. In fact double-labeled neurons are situated within the terminal fields of nearly all the afferent systems described in the rat. These include sensorimotor cortex,** visual cortex,” the deep cerebellar nuclei,32 superior colliculus’* mammillary nuclei,13 nucleus of the optic tract,3’ and lateral geniculate body.*’ However, it should be noted that although no specific patterns of connectivity based on the congruency of afferent projection fields and the locations of double-labeled neurons was evident, the possibility remains that certain functional linkages might be operational and simply are not revealed by the anatomical methodology employed in this study. Electrophysiological studies in the cat have indicated both that the corticopontine’.3 the and pontocerebellar’ systems are each composed of two functional sets of fibers (fast and slow) differentiated by their conduction velocity. It seems reasonable to suggest that this duality of input and output organization might in some way be related to whether

G. A. Mihailoff

160

or not a given BPN or NRTP neuron distributes in a restricted fashion to a single lobule or functional unit, or collateralizes to reach multiple lobules or functional cerebellar zones.

Acknowledgements-The author is grateful for the technical assistance provided by MS C. Stricklin and Mr Jahangir Imani. This work was supported in part by USPHS grant NS 12644, NSF grant BNS-80-04853 and funds of the Biologic Humanics Foundation.

REFERENCES

1. Allen G. I., Kom H. and Oshima T. (197%~) The mode of synaptic linkage in the cerebra-ponto-cerebellar of the cat--I. Responses in the brachium pontis. Exp/ Bruin Res. 24, l-14.

pathway

2. Allen G. I., Kom H., Oshima T. and Toyama K. (19756) The mode of synaptic linkage in the cerebra-ponto-cerebellar pathway of the cat-II. Responses of single cell in the pontine nuclei. Expi Bruin Res. 24, 15-36. 3. Allen G. I., Oshima T. and Toyoma K. (I 977) The mode of synaptic linkage in the cerebra-ponto-cerebellar pathway investigated with intracellular recording from pontine cells of the cat. Expl Bruin Res. 29, 123-136. 4. Azizi S. A., Mihailoff G. A., Burne R. A. and Woodward D. J. (1981) The pontocerebellar system, an HRP study in the rat-I. Projections to the posterior vermis. J. cotnp. Neural. 197, 543-558. 5. Bower J. M., Beerman D. H., Gibson J. M., Shambes G. M. and Welker W. (I 98 I) Principles of organization of a cerebro-cerebellar circuit. Brain Behuu. Evol. 18, l-18. 6. Brodal A. and Hoddevik G. H. (1978) The pontocerebellar projection to the uvula in the cat. Expi. Bruin Res. 32, 105-l 16.

I. Brodal A. and Jansen J. (1946) The ponto-cerebellar projection in the rabbit and cat. J. camp. Neural. 84, 31-l 18. 8. Brodal P. (1979) The pontocerebellar projection in the rhesus monkey: an experimental study with retrograde axonal transport of horseradish peroxidase. Neuroscience 4, 193-208. 9. Brodal P. and Walberg F. (1977) The pontine projection to the cerebellar anterior lobe. An experimental study in the cat with retrograde transport of horseradish peroxidase. Expl Bruin Res. 29, 233-248. 10. Burne R. A., Eriksson M. E., Saint-Cyr J. A. and Woodward D. J. (1978) The organization of the pontine projection to lateral cerebellar areas in the rat. Dual zones in the pons. Bruin Res. 139, 34@347. 11. Burne R. A., Mihailoff G. A. and Woodward D. J. (1978) Visual corticopontine input to the paraflocculus: a combined autoradiographic and horseradish peroxidase study. Bruin Res. 143, 139-146. 12. Bume R. A., Azizi S. A., Mihailoff G. A. and Woodward D. J. (1981) The tectopontine projection in the rat with comments on visual pathways to the basilar pans. J. camp. Neural. 202, 287-307. 13. Cruce J. A. F. (1977) An autoradiographic study of the descending connections of the mammillary nuclei of the rat. J. camp. Neurol. 176, 631-644. 14. Eisenman L. M. (1980) Pontocerebellar projections to the paraflocculus in the rat. Bruin Res. 188, 550-554. 15. Eisenman L. M., and Noback C. R. (1980) The pontocerebellar projection in the rat: differential projections to sublobules of the uvula. Expl Bruin Res. 38, 11-18. 16. Eisenman L. M. (1981) Pontocerebellar projections to the pyramis and copula pyramidis in the rat: evidence for a mediolateral topography. J. camp. Neural. 199, 77-86. 17. Hoddevik G. H. (1975) The pontocerebellar projection onto the paramedian lobule in the cat: an experimental study with the use of horseradish peroxidase as a tracer. Bruin Res. 95, 291-307. 18. Hoddevik G. H. (1977) The pontine projection to the flocculonodular lobe and the paraflocculus studied by means of the retrograde axonal transport of horseradish peroxidase in the rabbit. Expl Bruin Res. 30, 51 l-526. 19. Hoddevik G. H., Brodal A., Kawamura K. and Hashikawa T. (1977) The pontine projection to the cerebellar vermal visual area studied by the means of retrograde axonal transport of HRP. Bruin Res. 123, 209-227. 20. Kuypers H. G. J. M., Catsman-Benevoets C. E. and Padt R. E. (1977) Retrograde axonal transport of fluorescent substances in the rat’s forebrain. Neurosci. Letts 6, 127-135. 21. Larsell 0. (1952) The morphogenesis and adult pattern of lobules and fissures of the cerebellum of the white rat. J. camp. Neural. 97, 281-356.

22. Mihailoff G. A., Burne R. A. and Woodward D. J. (1978) Projections of the sensorimotor cortex to the basilar pontinc nuclei in the rat: an autoradiographic study. Bruin Res. 145, 347-354. 23. Mihailoff G. A., Martin G. F. and Linauts M. (1980) The pontocerebellar system in the opossum Dide~phis virginiunu. a horseradish peroxidase study. Bruin Behuo. Evol. 17, 179-208. 24. Mihailoff G. A. and McArdle C. B. (1981) The cytoarchitecture, cytology and synaptic organisation of the basilar pontine nuclei in the rat--II. Electron microscopic studies. J. camp. Neural. IS, 203-219. 25. Mihailoff G. A., McArdle C. B. and Adams C. E. (1981) The cytoarchitecture, cytology and synaptic organization of the basilar pontine nuclei in the rat--I. Nissl and Golgi studies. J. camp. Neurol. 195, 181-201. 26. Mihailoff G. A., Burne R. A., Azizi S. A., Norell G. and Woodward D. J. (1981) The pontocerebellar system in the rat, and HRP study-II. Hemispheral components. J. camp. Neural. lW, 559-577. 27. Ribak C. E. and Peters A. (1975) An autoradiographic study of the projections from the lateral geniculate body of the rat. Bruin Res. 92, 341-368. 28. Rosina A., Provini L., Bentivoglio M. and Kuypers H. G. J. M. (1980) Pontoneocerebellar axonal branching as revealed by double fluorescent retrograde labeling technique. Bruin Res. 195, 461466. 29. Rosina A. and Provini L. (1981) Pontine projections to crus I and crus II of the cat cerebellum. A horseradish peroxidase study. Neuroscience 6, 2613-2624. 30. Shambes G. M., Gibson J. M. and Welker W. (1978) Fractured somatotopy in granule cell tactile areas of rat cerebellar hemispheres revealed by micromapping. Bruin Behuv. Evol. 15, 94140. 31. Terasawa K., Otani K. and Yamada J. (1979) Descending pathways of the nucleus of the optic tract in the rat. Bruin Res. 173, 405417.

32. Watt C. B. and Mihailoff G. A. (1982) The cerebellopontine

system in the rat-i.

(Accepted 9 February 1983)

Autoradiographic

studies.