The Olivocerebellar Projection in the Cat Studied with the Method of Retrograde Axonal Transport of Horseradish Peroxidase IV. THE PROJECTION TO THE ANTERIOR LOBE ALF BRODAL AND FRED WALBERG Anatomical Institute, ffniuersity of Oslo, Oslo 1 , Norway
ABSTRACT Following injections of horseradish peroxidase (HRP) in the cerebellar cortex of the anterior lobe of the cat, the distribution of labeled cells in the inferior olive was mapped. The findings largely confirm those made previously in studies of olivary retrograde cell loss following cerebellar ablations (Brodal, '40b). In addition, they reveal further olivary areas projecting onto the anterior lobe, and permit a more detailed analysis of the pattern in this projection. Concerning major points the results are in agreement with physiological studies by Armstrong et al. ('74). They bring supporting evidence for a longitudinal zonal pattern in the anterior lobe (fig.6C). The middle zone of the uermis receives its fibers from a large central area in the caudal half of the medial accessory olive, a lateral zone of the vermis from the lateral half of the dorsal accessory olive. Both olivary areas project to the corresponding cerebellar zone throughout lobules V-I. The lateralmost part of the anterior lobe (lobules IV-V) receives afferents from an area in the dorsal lamella of the principal olive. The intermediate part of lobules IV-V receives afferents from the medial half of the dorsal accessory olive and from an area in the rostral half of the medial accessory olive. There is suggestive evidence that the latter projects to a middle zone, the former to a medial and a lateral zone within the intermediate part as found physiologically. Conclusions concerning projections to the intermediate part of lobules 111-11 could not be made. The findings in this and preceding studies with the HRP-method show that the concept of a longitudinal pattern in the cerebellum is scarcely generally valid of the entire olivocerebellar projection. Within the projections of the lateral half of the dorsal accessory olive and the area in the rostral part of the medial accessory olive there appears to be a topical relation with the folial pattern in the anterior lobe. An analysis of the findings with reference to the afferents traced anatomically to the various olivary areas permits some conclusions as to the functional role of the olivary areas. Comparison with Oscarsson's ('73) diagram of the sites of termination of two of the spinal-olivary pathways (his DF-SOCP and VF-SOCP) permits an anatomical explanation as concerns the projections to the vermis, while correlations as concerns the intermediate part are less satisfactory.
The horseradish peroxidase (HRP) method is well suited for anatomical investigations of the pattern of the olivocerebellar projection in the cat (Walberg et al., '76). The present study is a continuation of previous reports dealing with the olivary projections to the paramedian lobule J. COMP. NEUR., 172: 85-108.
(Brodal et al., '751, the uvula (Brodal, '76) and the vermal visual area (Hoddevik et al., '761, in which several new data have been brought forward. According to Brodal's (40b) studies of the retrograde cellular changes in the olive following partial cerebellar ablations in 85
86
ALF BRODAL AND FRED WALBERG
kittens (modified Gudden method, Brodal, '40a1, the anterior lobe receives its olivary afferents only from the accessory olives. More particularly the vermis receives fibers from the lateral half of the dorsal accessory olive and from the latero-caudal part of the caudal half of the medial accessory olive, while the lateral and intermediate part receive fibers from the medial half of the dorsal accessory olive (fig. 6A). Results obtained by autoradiographic tracing of olivocerebellar fibers (Courville, '75) do not contradict these findings. Recent physiological studies have in part confirmed and considerably extended the early anatomical observations. From cerebellar recordings following electrical stimulation of the inferior olive (cat) VanGilder and O'Leary ('70) concluded that: "The dorsal accessory olive appeared to be the sole origin of responses recorded from the paravermal sector of the anterior lobe." Further (their fig. 6) responses recorded from the vermis of the anterior lobe come from the middle and caudal parts of the medial accessory olive. More precise data were obtained by Armstrong et al. ('74) by recording antidromic potentials in identified parts of the olivary complex following electrical stimulation of points on the cerebellar surface. In their summarizing diagram (their fig. 5; see fig. 6B here) they distinguish in the anterior lobe six longitudinal zones and a rostral one, which receive their olivary afferents from particular, in part coinciding, regions of the olive. In addition to areas in the medial and dorsal accessory olives, approximately the same as those found by Brodal ('40b), they describe projections from more rostral parts of the medial accessory olive as well as from restricted parts of the dorsal and ventral lamella (cp. figs. 6A and 6B). As will be shown, the use of the HRPmethod has made it possible to map the olivary projections to the anterior lobe in considerable detail. Projection areas have been identified in addition to those described by Brodal ('40b). Concerning the
main points the present results agree with those of Armstrong et al. ('741, and they bring some additional information as well. MATERIAL AND METHODS
The experimental material consists of some 30 cats of which 21 could be used in the present study. For particulars concerning the methods employed and the principles of presentation, see Walberg et al. ('76) and Brodal et al. ('751, respectively. Briefly the procedure has been as follows: Surgery was performed under deep Nembutal anaesthesia. In parts of the anterior lobe which can be exposed by a craniotomy, largely lobule V and part of lobule IV of Larsell ('701, the injection of HRP was often made through a fine slit in the dura. Injections in other lobules were made stereotactically. Table 1gives the weight, amount and concentration of the HRP fluid injected and the survival time in the various cases. In one case HRP type Serva was used, in the others type Sigma VI P 8375. The animals were killed under deep Nembutal anaesthesia by intracardiac perfusion of a mixture of 0.4% paraformaldehyde and 1.25% glutaraldehyde with phosphate buffer. Dissected blocks were serially cut on the freezing microtome at 50-6Op. From every group of five sections (transverse of the brain stem, sagittal of the cerebellum) two were treated with 3.3'-diaminobenzidine according to the method of Graham and Karnovsky ('66) and mounted on slides. One of the sections was weakly stained with cresyl violet or thionine, the other was left unstained. The distribution of labeled cells in the inferior olive, marked under the microscope in drawings of each section mounted, was transferred to a standard diagram of the olive as seen in 15 equally spaced transverse sections (fig. 1C). In order to facilitate comparisonsbetween cases, these findings were transferred to a diagram of a reconstruction of the olivary complex as imagined unfolded (Brodal, '40b). The occurrence of labeled cells in the olive is shown by dots. The distribution of stained
87
OLIVOCEREBELLAR PHOJECTION TO ANTERIOR LOBE
TABLE 1 Peroxidase injected Cat B.St.L.
Weight
kg
623 625 629 630 648 654 665 689 700 701 704 706 707 709 710 712 713 714 715 723 725
2.3 0.5 2.2 2.4 1.8 1.o 3.0 1.8 2.1 1.9 3.0 3.2 3.2 3.2 2.0 2.5 2.1 2.5 3.5 2.2 4.0
Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Serva Siema VIP
cortical areas is indicated by hatchings in a diagram of the cerebellar surface, taken from Larsell (’70).Areas where the molecular layer is stained are shown in black. The term “vermis” will be used for the cortex overlying the fastigial nucleus. The more lateral parts of the anterior lobe will be referred to as the “intermediate part”, and the lateralmost parts of lobules IV and V will be called the “lateral part”. RESULTS
Following injections of identical amounts of fluid with the same concentration of HRP there may be considerable variations in the extension of the stained cerebellar areas (for particulars, see Walberg et al., ’76). Special attention has been paid to staining of the molecular layer, since the olivary fibers (at least many of them) end as climbing fibers. The staining of the cortex is often irregular, often the tops of some folia are unstained while their sides are well stained. Diffusion of HRP in the cortex occurs particularly easily along the length of the folia. The identification of the particular stained folia and lobules in the anterior lobe offers no problems in sagittal sections. However, determination of the
Concentration
Amount
50%w/v 50%wlv 50%wlv 50%wlv 100%wlv 50%wlv 50%w/v 50%wlv 50%w/v 50%w/v 50%w/v 50%wlv 50%w/v 50%w/v 50%wlv 50%wlv 50%wlv 50%w/v 50%w/v 50%wlv 50%wlv
0.5 pl 0.25 pl 0.5 pl 0.5 pl 0.25 pl 0.75 pl 0.5 pl 0.4 pl 0.2 pl 0.25 pl 0.2 pl 0.2 p1 0.2 pl 0.2 pI 0.2 pl 0.2 p1 0.2 pl 0.2 pl 0.15 pl 0.2 pl 0.2 ul
Survival time in days
3 2 3 2 2 1 1
2 2 1
2 2 2 2 3 2 2 2 2 3 2
extension of cortical staining in the transverse direction, i.e., along the folia, can only be made with some approximation. In some cases there is spreading of the staining from the cortex to the underlying cerebellar nuclei, where some cells may be labeled (Walberg et al., ’76).Since the nuclei receive axons or collaterals of olivocerebellar fibers, spreading of HRP to the nuclei following cortical injections may be a source of error in mapping the distribution of labeled olivary cells. Injections of HRP in the nuclei have, therefore, been made in some cases. The labeled neurons in the olive are easily distinguished, particularly under dark field illumination (Walberget al., ’76). The intensity of labeling varies among cases, and in a particular case one region may often contain chiefly heavily labeled cells, while in another labeled area most cells are only weakly labeled. The distribution within the olivary complex of labeled neurons varies among cases according to the site and extension of the staining of the cortex of the anterior lobe. Two cases with staining of the cortex of the vermis and intermediate part of lobule V serve to illustrate some main features.
88
ALF BRODAL AND FRED WALBERG
In cat l3.t.L. 629 (fig. 1) the injection has resulted in staining of the cortex of the lower folia of lobule V with only weak staining of a small area of the adjoining folium of lobule VI. On the left the staining involves the cortex of the vermis and a considerable portion of the intermediate part, on the right it covers only the medial part of the vermis. The white matter in the lowermost folia of lobule V is stained (fig. 1A). From these folia some stained fibers can be followed to the rostral part of the left fastigial nucleus where a few cells are labeled. The right inferior olive contains labeled cells in both accessory olives, but none in the principal olive (figs. lC,D). In the dorsal accessory olive the caudal half, approximately, contains labeled cells, particularly dense medially, where at some levels (VIII-1x1 practically all cells are labeled. In the caudal half of the medial accessory olive there is a large central area which contains many heavily labeled cells, while relatively fewer and less heavily labeled cells occur laterally, caudally and medially to this area. Finally, there is an area of labeled cells medially in the rostral half of the medial accessory olive. In the left inferior olive less intensely labeled cells than in the right olive occur only in the caudal half of the medial accessory olive, mainly in the central part of the area containing labeled cells on the right. Very similar findings are made in cat l3.St.L. 714 (fig. 1, inset). There is cortical staining of the left anterior lobe almost as in the previous case, but the stained area is situated a little more rostrally with no diffusion to lobule VI and it reaches only to the midline. A few labeled cells are found in the left nucleus interpositus anterior. Labeled cells are found only in the right olive, in regions which correspond almost completely to those found in the right olive in cat B.St.L. 629 (fig. 1D).
The projection to the vermis of the anterior lobe Cases with HRP-staining restricted to
the vermis show that of the olivary areas containing labeled cells in the preceding two cases (fig. 11, only those in the caudal half of the medial accessory olive and in the lateral part of the dorsal accessory olive project to the vermis. While the former area projects to the vermis of all lobules of the anterior lobe, lobules I-V differ somewhat with regard to their projection from the dorsal accessory olive. Abbreuiations @,Nucleus @ (Brodal, '40b) of inferior olive D, Dorsal accessory olive d.cap., Dorsal cap d.l., Dorsal lamella of principal olive dm.c.col., Dorsomedial cell column (Mareschal, '34) Flocc., Flocculus F.pr., Primary fissure l., Lateral Lobant., Anterior lobe L.pm., Lobulus paramedianus M, Medial accessory olive m., Medial nucl. p, Nucleus @ (Brodal, '40b) of inferior olive P.fl.d., Paraflocculus dorsalis P.fl.v., Paraflocculus ventralis v.l., Ventral lamella of principal olive v.I.o., Ventrolateral outgrowth I-V, Lobules of anterior lobe (Larsell, '70) Fig. 1 Presentation of the findings in cat B.St.L. 629 with HRP-injection in lobule V. Above to the left (A) a drawing of a nearly midsagital section of the cerebellum, to the right (B) a diagram of part of the cerebellar surface (taken from Larsell, '70). The stained regions are hatched, black denotes staining of the molecular layer. In the middle (C)a standard diagram of a series of 15 equally spaced transverse sections through the inferior olive (taken from Brodal, '40b). The various subdivisionsof the olivary complex are indicated (see list of Abbreuiations). Dots show the occurrence of labeled cells, transferred from drawings of particular sections to the corresponding levels of the diagram. Density of dots indicates density of labeling of olivary neurons (each dot does not correspond to one labeled cell). Below (D) a diagram (from Brodal, '40b) of the inferior olive imagined unfolded into one plane by pulling it apart as indicated by arrows in the insets (1,2,3) below. In C and D the thin broken lines refer to borders between projection areas as determined by Brodal ('40b). The dotted areas are those shown in the row of sections (C).The entire right olive, projecting onto the left cerebellar half, is shown in D, of the left olive only the medial accessory olive is reproduced. Inset in lower left comer shows an almost corresponding injection of the cortex in another case. Compare text. For Abbreviations see list above.
89
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LOBE
R
3J Q -. C vL
cI
caudal Right
Left
PRINCIPAL OLIVE
DORSAL ACC OLIVE
Left
2
I
Right
I
z
Figure 1
3
I
90
ALF BRODAL AND FRED WALBERG
B.St.L.689
MEDIAL
MEDIAL
ACC. OLIVE ACC.OLIVE -~
rostral
PRINCIPAL OLIVE
DORSAL ACC.OLIVE
d
n
caudal
Left
-
Right
B.S t.L.654
i) Left
MEDIAL ACC.OLIVE
Right
MEDIAL ACC OLIVE
rostral n
coudol Left
Right
Fig. 2 Diagrams of the findings in three cases (cats B. St.L. 689, and 648) with HRP-injections in the vermis of lobule V. For each case is shown a drawing of the cerebellar surface, some representative transverse sections through the olive, and a diagram of the unfolded olivary complex. Symbols as in figure 1. For Abbreviations see list on p. 88.
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LORE
In cat B.St.L. 689 (fig.2A) the injection has resulted in staining of lobule V, on the left extending to the medial border of the intermediate part, on the right side involving the medialmost part of the vermis only. Some labeled fibers can be traced to the nucleus fastigii where they break up into terminals. There are no labeled cells in the nucleus. Both inferior olives contain labeled cells. On the right most of them are found centrally in the caudal half of the medial accessory olive, some in the dorsal accessory olive at caudal levels. In the left olive there are only a few labeled cells in the area of the medial accessory olive which contains labeled cells in the right olive. In cat B.St.L. 654 (fig.2B) there is good staining of a circumscribed area in the vermis of the caudal part of lobule V, almost symmetrical,with staining of the molecular layer except for the summits of most of the stained folia. The lateralmost parts of the vermis are free. There is no visible staining of the white matter outside the folia, and no labeled cells in the fastigial nucleus. Faintly labeled cells occur in practically identical areas centrally in the caudal half of both medial accessory olives, a little more marked on the right than on the left. On the right in addition two labeled cells were found caudally in the dorsal accessory olive at level V. An even more restricted distribution of labeled olivary cells is found in cat B.St.L. 648 (fig. 2C), where 0.25p1 of a concentrated suspension of HRP (100%w/v) was injected in the rostralmost folia of lobule V near the midline. The staining of the cortex extends a little more laterally on the left than on the right. A few weakly labeled cells are found in the caudal part of the right medial accessory olive, most of them in a small cluster at level VI. At the corresponding site on the left there are a few scattered labeled cells in the sections mounted. Following injections of the vermal region of lobules IV, 111, I1 and I, the distribution of labeled cells in the olive resembles that described above for lobule V, apart from
91
minor differences in the dorsal accessory olive. In cat I3.St.L. 709 (fig. 3A) a stereotactic injection through lobule V has resulted in staining of a restricted part of the v e m i s of lobule ZV, mainly to the right of the midline, particularly of the molecular layer. A few stained fibers can be followed to the right fastigial nucleus, but there are no labeled cells in this. In the inferior olive there are (mainly weakly) labeled cells on both sides. On the left, they occur in a central area in the caudal half of the medial accessory olive and almost throughout the rostrocaudal extent of the lateral half of the dorsal accessory olive, with a maximum at levels VII-IX. In the right olive labeled cells occur only in the central part of the caudal half of the medial accessory olive. In cat B.St.L. 723 (fig. 3A, inset) following a stereotactically made injection (HRP type Serva) through lobule VI there is staining of the caudalmost folia of the vermis of lobule ZVand the rostralmost folium of lobule V. On the right staining reaches approximately to the lateral border of the vermis, on the left it involves only the medialmost part of the vermis. There is no staining and no labeled cells occur in the cerebellar nuclei. In the inferior olive the findings correspond largely to those made in the previous case (B.St.L. 709, fig. 3A). Labeled cells occur on both sides centrally in the caudal part of the medial accessory olive, scanty on the right, and in the caudal part of the lateral half of the left dorsal accessory olive. In another cat (B.St.L. 706, not illustrated) the findings in the cerebellum and the olive correspond almost completely to those in the two preceding cases (fig.3A). In cat B.St.L. 701 (fig. 3B) the stereotactically made injection has resulted in good staining of the vermis of the rostral part of lobule ZZZ and the caudal part of lobule ZZ, most marked near the midline but extending laterally almost symmetrically. There are a few labeled cells in the rostral part of the fastigial nuclei. The labeling of neurons in the olive is on the whole rather weak,
92
ALF BRODAL AND FRED WALBERG
B.St.L.709
A
Left DORSAL ACC OLIVE
rdstral
PRINCIPAL OLIVE
MEDIAL ACC OLIVE
PRINCIPAL OLIVE
DORSAL ACC OLIVE
Right
MEDIAL ACC OLIVE
~
w XI
m caudal
1
B.St.L. 70 I
B
MEDIAL ACC OLIVE
m xv xnr XI m wr M
vm M VII
m N
m N caudal
I
B.St.L.700
I Right
and
Left
1B.St.L.725
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LOBE
but labeled cells occur on both sides, almost symmetrically, in one area centrally in the caudal half of the medial accessory olive and a more extensive area in the lateral part of the dorsal accessory olive, with a faint maximum at levels IX-X. Following an injection which results in staining of the vermis of some folia of lobule ZZZ (cat B3t.L. 704, fig. 3B, inset), the distribution of on the whole faintly labeled cells corresponds closely to that shown in the diagram of the olive in cat B.St.L. 701 (fig.3B).The number of labeled cells in the dorsal accessory olive is largest between levels XII-VIII. A further case (cat B.St.L. 725, fig. 3B, inset) with an injection of the vermis of lobules 11-111 shows a corresponding distribution of labeled olivary neurons as found in cat B.St.L. 704. Cat B.St.L. 700 (fig.3C). The tip of the stereotactically inserted needle has been in the white matter just beneath the base of lobule ZZ. On both sides of the midline, almost symmetrically, there is considerable staining of the white matter and very good staining of the molecular layer of the verma1 parts of the folia of lobule ZZ, especially anteriorly, and of the adjoining cortex of lobule 1. In the rostralmost part of the fastigial nuclei there is a light brownish color, and a few cells are labeled. In both olives well labeled neurons occur symmetrically in a rather circumscribed central part of the caudal half of the medial accessory and in almost the entire lateral half of the dorsal accessory olive. In the latter labeling extends practically to its rostral end. The majority of labeled cells occurs between levels X and XIII. In cat B.St.L. 665 (fig. 3C, inset) the staining of the cortex is almost identical to that in the former case, and the findings in the olive correspond closely to those described above. In cat B3t.L. 715 (fig. 3C, inset) staining
93
is restricted to the vermis of lobule 11, mainly on the right. It does not extend quite to the lateral border of the vermis. In the inferior olive well labeled cells occur bilaterally in the areas showing labeling in the two previous cases (B.3.L. 700 and 665, fig. 3C1, on the whole more abundant on the left than on the right, particularly in the dorsal accessory olive. On both sides this shows a maximum of labeling at levels XI-XIII.
Zntermediate and lateral parts of the anterior lobe Injections in the anterior lobe lateral to the vermis result in a distribution of labeled olivary cells different from that following injections in the vermis. In cat l3.St.L. 630 (fig. 4A) cortical staining, mainly of the molecular layer, is present in the lateralmost part of the left lobule V, while the vermal zone and most of the intermediate part are free. There is no staining of the cerebellar nuclei and there are no labeled cells in these. In the olive labeled cells are present only on the right side and in two fairly well circumscribed regions. One of them, containing weakly labeled cells, is found in the dorsal accessory olive, the other in the dorsal lamella of the principal olive, where on the whole, labeling is more marked. In cat B.St.L. 713 (fig. 4A, inset) a restricted injection in the lateralmost part of lobule V, with a little spreading to lobule VI, resulted in labeling of a few cells only in the part of the contralateral dorsal lamella labeled in cat B.St.L. 630 (fig.4A). In cat B.St.L. 625 (fig. 4B) an injection in the intermediate part of the left lobule V has spread to the adjoining folia of lobule VI. Some stained fibers can be traced towards the cerebellar nuclei, and a few labeled-cellsoccur in the left nucleus interpositus posterior. In the inferior olive labeled cells are found in two regions only on Fig. 3 Diagrams of the findings in 3 cases with in- the right side. One covers a considerable jections in the vermis of lobules IV-I1 (cats B.St.L. part of the medial half of the dorsal ac709, 701 and 7001,presented as in figure 2. Insets cessory otiue, the other is found medially in show stained areas in other cases with findings in the the rostral half of the medial accessory olive which are closely similar to those in the corresponding main case (see text). For Abbreuiatim see olive. An injection in the lateral and intermelist on p. 88.
94
ALF BRODAL AND FRED WALBERG
B.St.L.630
Left
Right
rost ral
caudal
-
Right
Fig. 4 Diagrams of the findings in two cases (cats B.St.L.630 and 625) with injections in the lateral-intermediatepart of lobule V, presented as in figure 3. Inset shows staining of cortex in a third case (see text). For Abbreviations see list on p. 88.
diate parts of the left lobule ZVkat B.St.L. 623, fig. 5A) gives a distribution of labeled cells resembling that following injections in the lateral and intermediate parts of lobule V when taken together (figs. 4A,B). There are no signs of diffusion of HRP from the restricted stained area of the cortex to the nuclei. The right olive shows three areas of labeled cells, one in the medial half of the dorsal accessory olive, another in the dorsal lamella, and a third in the rostra2 half
of the medial accessory olive. These areas overlap those containing labeled cells following injections in the lateral-intermediate parts of lobule V, but are a little differently placed. Labeling is somewhat stronger in the medial accessory olive than in the two other regions. An almost completely corresponding injection of the left lateral-intermediate part of lobule IV was achieved in cat B.St.L. 710 (fig. 5A, inset). Some stained fibers can be
95
OLIVOCEREBELLAR PROJECTION TO ANTERIOR COBE
6.St.L. 623
-
,
I
MEDIAL ACC OLIVE
B.St.L.707
B
PRINCIPAL OLIVE
DORSAL ACC OLIVE
M
Left MEDIAL ACC OLIVE
PRINCIPAL OLIVE
rostra1
Right
DORSAL ACC OLIVE
caudal Right
Fig. 5 Diagrams of two cases (cats B.St.L. 623 and 707) with injections in the lateral-intermediate part of lobule IV, presented as in figure 3. Inset shows staining of the cortex in a third case. For Abbreuiations see list on p. 88.
traced from the injected area to the nucleus interpositus anterior and posterior, which contain a few labeled cells. Well labeled neurons are found in the right inferior olive in regions which are almost identical to those found in the previous case (fig. 5A). A somewhat different pattern in the distribution of labeled olivary neurons is seen
when the lateral part of lobule IV (its molecular layer) is not stained. In cat. B.St.L. 707 (fig. 5B) a stereotactic injection through lobule VI has resulted in staining of folia of lobule ZVon the left side. The stained cortex (particularly the molecular layer) does not include the lateral end of the lobule, medially it encroaches a little on the vermis. There is considerable stain-
96
ALF BRODAL AND FRED WALBERG
ing of the white matter of lobule IV. A central part of the nucleus interpositus anterior is faintly brownish stained and contains some labeled cells, as well as labeled fibers entering from lobule IV. The fastigial nucleus is not affected. In the olive labeled cells are found only in two regions on the right. There is a restricted area containing mainly weakly labeled cells caudally in the rostral half of the medial accessory oliue, and a larger area centrally in the dorsal accessory oliue, where there is a marked maximum of labeled cells at levels VIII-IX. In four cats attempts were made to achieve staining of the intermediate part of the anterior lobe of lobules 11-111. Conclusions concerning olivary afferents to this part could not be drawn because in all cases there was considerable spreading of HRP to the dentate nucleus.
Znterpretation of findings Labeled olivary neurons can be identified with great certainty and their topographical distribution can be mapped with precision. In spite of the difficulties referred to above involved in determination of the mediolateral extent of the injected cerebellar area, it is striking to note the almost complete correspondence between olivary areas labeled in different cases in which the same cerebellar area has been stained by HRP. This fact further shows that the errors which might be induced by transferring the findings in the olive in individual cases to a standard diagram, can be only minor ones. An analysis of our material gives a picture of the projection as shown in figure 6C. The olivary projections to the vemnis of the anterior lobe (fig. 6C) Whenever the vermis of the anterior lobe is included in the cortical area stained following injections of HRP, there is labeling of cells in a central area in the caudal half of the medial accessory olive. This is the case for each of the lobules V-I (figs. 1-31. Apart from minor variations in size of the area and in the density of labeled cells, this area is almost identical in all cases.
In many of our cases with vermal injections there is in addition labeling of cells in the lateral half, approximately, of the dorsal accessory olive. However, this occurs only when the staining (molecular layer) extends to the lateral border of the vermis. This is clear as concerns lobules V and IV (compare the two sides of the injected cerebellar areas in figs. 1,2A and 3A). (The labeling in the medial part of the right dorsal accessory olive in fig. 1 is due to staining of the intermediate part of lobule V, see below). In lobules 1-111staining entirely restricted to the medial part of the vermis has not been achieved. However, from a comparison of these cases it appears that the labeling in the lateral half of the dorsal accessory olive is more marked the further laterally in the vermis the staining has spread (for example, cats B.St.L. 704, fig. 3B, inset; 715, fig. 3C, inset). In cases with staining mainly of the intermediate part of Fig. 6 Diagrams of the olivocerebellar projection onto the anterior lobe. A, As determined with the modified Gudden method by Brodal (‘40b). Above a diagram of the anterior lobe (after Larsell, ’70), and some representative transverse sections through the olive from caudal (I) to rostral (XV) (cp. fig. 1). Below a reconstruction of the olivary complex as imagined unfolded in one plane (cp. fig. 1). B. The projection as Found by Armstrong et al. (‘74).To the left their drawing of the anterior lobe, in the middle the olivary areas projecting to different zones in the anterior lobe. The order of the sections has been reversed in order to Correspond to the sequence used here. The Roman numerals indicate the approximate levels of the olive (as marked here) to which the levels reproduced in Armstrong et al.’s diagram correspond. On the basis of this, their findings are transferred to our diagram of the unfolded olive (to the right). C. The projection as determined in the present study. To the left surface of the anterior lobe, to the right a series of representative transverse sections through the olive. Below the projection seen as entered in the diagram of the unfolded olive. The intermediate part is shown as receiving fibers from the medial half of the dorsal accessory olive and a region in the rostral part of the medial accessory olive (open rings). Not shown is the suggestive conclusion that the latter projects to a middle zone of the intermediate part, the former to a medial and a lateral zone, as in figure 6B. Information of olivary projections to areas left white in the cerebellum was not obtained. See text. For AbbreoMtions see list on p. 88.
97
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LOBE Ivernxs
lvermis
I
xm
caudal
m
rostra1
MEDIAL ACC OLIVE
PRINCIPAL OLIVE
rostrol
caudal
Figure 6
DORSAL ACC OLIVE
98
ALF BRODAL AND FRED WALBERG
Fig. 7 Diagrams to illustrate the topographical pattern in the projection of the medial part of the rostral medial accessory olive and the lateral part of the dorsal accessory olive onto the anterior lobe. See text.
lobules 11-111,there is labeling in the lateral half of the dorsal accessory olive when staining involves the lateral zone of the vermis (cat B.St.L. 707, fig. 5B). It appears thus that the two projection zones in the vermis extend at least to lobule I1 as indicated in figure 6C. It follows from these findings that the vermis of the anterior lobe throughout its length receives fibers from a central area in the caudal half of the medial accessory olive and from the lateral half of the dorsal accessory olive. Furthermore, the former area projects onto the medial zone of the vermis, the latter to a lateral zone of the uermis (fig. 6C). While no evidence has been found for a topical arrangement within the projection to the vermis from the medial accessory olive, there is evidence for an arrangement of this type in the projection from the lateral half of the dorsal accessory olive (fig. 7). In spite of considerable overlapping, the area labeled in this is on the whole situated progressively further caudally as the injection is placed more posteriorly, from lobules 1-11 to lobule V. Furthermore, the part of the area which contains the greatest proportion of labeled cells shows a corresponding variation.
The olivary projections to the intermediate-lateral part of the anterior lobe (fig. 6C) These arise from other olivary areas than do the projections to the vermis. Following injections of the intermediate-lateral part, labeling is found in three places, one area in the rostral half of the medial accessory
olive, another in the medial h a l , approximately, of the dorsal accessory olive, and a third in part of the dorsal lamella of the principal olive. These three olivary areas appear to project to different regions of the anterior lobe, apparently different longitudinal zones. Most convincing in this respect is the evidence for the projection from the area in the dorsal lamella (fig. 6C), since labeling occurs here only when the staining of the cortex (molecular layer) of the anterior lobe includes its most lateral regions (figs. 4A, 5A). There is some evidence from cases not described that the lateralmost part of lobule I11 likewise receives fibers from this olivary area. Since we have not been able to obtain reliable information concerning the projection to the intermediate part of lobules 11-111, these are left without symbols in figure 6C. The two areas in the accessory olives mentioned above, are most often both labeled following staining of the intermediate part. In the dorsal accessory olive the area labeled in our relevant cases (figs. 1, 4A,B, lobule V; figs. 5A,B, lobule IVI, when taken together cover its entire medial half (open rings, fig. 6C). It appears that this area overlaps somewhat that of the lateral half of the dorsal accessory olive projecting onto the lateral zone of the vermis (open triangles). No clear topographical relations can be discerned in the projection from the medial half of the dorsal accessory olive. In the area in the rostral half of the medial accessory olive, however, it appears (fig. 7) that fibers to lobule V take origin mainly from its medial part, while those to lobule IV arise more laterally (Cp. figs. 1 and 4B with figs. 5A,B). Whether these projections from the dorsal and medial accessory olives supply different zones of the intermediate part cannot be finally decided. However, in two cases with injections involving the intermediate part there is labeling in the dorsal accessory olive but not in the rostral part of the medial accessory olive. In one of these cases (cat B.St.L. 720, not illustrated) only the medialmost region of the intermediate part is stained. This may suggest, as con-
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LOBE
cluded by Armstrong et al. ('74,see fig.6B here), that there are in the intermediate part a lateral and a medial zone (open rings in fig. 6B) receiving fibers from the dorsal accessory olive on each side of a middle zone (vertical hatching) which receives fibers from the rostral part of the medial accessory olive. This tentative conclusion is not indicated in the diagram figure 6C. In some of the cases with injections of the anterior lobe there is evidence of spreading of some HRP to the cerebellar nuclei. To some extent the olivary areas projecting to the nuclei interpositi (Brodal et al., '75, their fig. 6) and to the fastigial nucleus (Hoddevik et al., '76, their fig. 4) coincide with those projecting to the intermediate-lateral part or the vermis of the anterior lobe, respectively. However, the diffusion of HRP to the nuclei in some of our cases is too modest to explain the marked labeling in these common projecting olivary areas. Furthermore, the same olivary areas are found labeled in cases without evidence of diffusion to the nuclei (cats B.St.L. 623, 630, 648, 654, 665, 689, 704, 706, 709, 713, 715). We feel, therefore, that the moderate spreading of HRP to the nuclei in some of our cases does not invalidate the conclusions drawn above. DISCUSSION
As discussed elsewhere (Walberg et al., '761, the use of the HRP method permits a precise mapping of the sites of labeled cells in the inferior olive following injections of HRP in the cerebellar cortex. The cortical area where axon terminals have been exposed to HRP, as witnessed by brown staining of the tissue, can be less precisely determined. However, when a number of cases with similar and different injection sites are compared, it is possible to outline some main features in the olivocerebellar projection to the anterior lobe as described above. The present material as well as previous studies (Brodal et al., '75; Brodal, '76; Hoddevik et al., '76) indicate that distribution of labeling in the olive is correlated to the extent of staining of the molecular layer.
99
From studies in other regions of the CNS it has been concluded that the distribution of retrogradely labeled neurons is smaller than might be predicted from the extent of diffusion of HRP as witnessed by the peroxidase positive area at the injection sites of their terminals (Bunt et al., '76, see this paper for further references). The correspondence in our material between the size of the stained part of the cerebellar cortex and the size of the labeled areas of the olive indicates that the situation may be different in the olivocerebellar projection, As found by a number of authors, our studies confirm that the olivocerebellar projection to the anterior lobe is purely crossed. The frequently observed astonishingly sharp transition between an olivary area containing labeled cells towards one free of such cells, shows that the olivocerebellar localization must be very precise. When there are more gradual transitions, this may be due to the fact that in the periphery of an injected area the concentration of HRP must be expected to be less than in the center, and accordingly not all terminals ending here may have taken up sufficient HRP to show labeling. The same fact may explain that the cells in the center of a labeled area are often more heavily labeled than those in the periphery. The uncontrollable spreading of HRP within the cortex prevents us from deciding whether cortex deep in the sulci differs in its projection from the superficial cortex.
1. Topography of the olivocerebellar projection to the anterior lobe It is of particular interest to consider our findings in relation to the electrophysiological studies of Armstrong et al. ('74) and to the data on a longitudinal zonal subdivision of the cerebellum. Even if minor errors are apt to occur in our transposition of the findings of Armstrong et al. ('74)to our diagram of the unfolded olive (fig. 6B), this will largely reveal itself in slightly incorrect markings of the borders of particular olivary projection areas. Concerning main points there is good agreement between
100
ALF BRODAL AND FRED WALBERG
the results of the two kinds of study. Discrepancies on certain points are probably largely due to differences between the methods used. The HRP method has made it possible to obtain information of some points which could not be clarified by the method of antidromic recording, and others which were not demonstrated with the anatomical method used by Brodal (’40b). The vermis of the anterior lobe was found by Brodal (’40b) to receive olivary afferents from the lateral half of the dorsal accessory olive and the caudolateral part of the medial accessory olive (fig. 6A). The former area corresponds to that determined with the HRP method (fig. 6C), the latter is found here to be situated somewhat more medially. Armstrong et al. (fig. 6B) likewise found two areas to project to the vermis of the anterior lobe (oblique hatching and triangles), one caudally in the medial accessory olive, the other caudolaterally in the dorsal accessory olive. The former area corresponds well with our mapping, except that we found only few labeled cells most laterally (this region projects largely to lobule VIII, Hoddevik et al., ’7611. The area in the dorsal accessory olive is indicated by Armstrong et al. as comprising only the caudal part of the lateral half of the dorsal accessory olive, while we find this area to extend practically to its rostra1 end. This is most likely due to the fact that Armstrong et al. did not explore the most anterior folia (1-111, probably in part IV). This assumption is strengthened by our observation that there is a topical pattern, although with considerable overlapping, in the projections to lobules I-V (fig. 7) from this part of the olive. Evidence that the lateral part of the dorsal accessory olive projects to anterior regions of the intermediate part as found by Armstrong et al. (‘741, dotted area in fig. 6B) was not obtained in our study. It cannot be seen from their diagram (their fig. 5, fig. 6B here) to what lobule the dotted area belongs. If this site of stimulation has been more medially than supposed in lobule I11 or IV, the dottings in the dorsal accessory
olive in their diagram (fig. 6B here) would agree completely with our findings (fig. 6C). Apart from the latter discrepancy there appears to be a satisfactory agreement between the findings of Armstrong et d. (’74) and our own concerning the total olivary areas projecting onto the anterior lobe vermis. However, while we found the projections to the medial and lateral vermal zones to take origin from separate olivary regions (oblique hatching and triangles, fig. 6C), Armstrong et al. found each of these regions to project to both zones (fig. 6B). We feel that our findings concerning separate projections are fairly decisive. Even if the current at the electrode site will not stimulate terminals more than 1.3 mm away (Armstrong et al., ’741, it is conceivable that spread of current to the other vermal zone on stimulation of one of them may explain their findings and has obscured the separate projections to the two zones of the vermis. Our interpretation receives support from the study of Voogd et al. (‘751, who with anterograde degeneration and tritiated leucine transport methods found “that the caudal parts of the medial and dorsal accessory olives project to zones A and B, respectively”, i.e., medial and lateral vermal zones. With regard to the olivary projection to the intermediate-lateral part of the anterior lobe, the HRP-study has confirmed Brodal’s (’40b) conclusion (fig. 6A) that the medial half, approximately, of the dorsal accessory olive projects to the intermediate part, at least as concerns lobules IV and V. VanGilder and O’Leary (’70) likewise concluded that all olivary afferents to this part come from the dorsal accessory olive. 1 In their fgure 2 Armstrong et al. (‘74) indicate antidromic potentials in the lateralmost part of the medial accessory olive (at about level IV of our diagram) following stimulationof the midvermis of lobule V as well as of what appears to be lobule VIII. It is of interest that according to their description (p. 2951 it appears “that the two stimuli were to a considerable extent exciting the axons of different olivary neurones,”while some were antidromicaly invaded following stimulation of either zone. This is readily explained if the site of their recording electrode in the latter situation has been at the transitionbetween the olivary areas projecting to lobules VIII and V, respectively, according to the HRP findings.
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LOBE
The HRP-method, in addition, has revealed two other olivary areas projecting to the intermediate-lateral parts, one in the rostral half of the medial accessory olive, another in the dorsal lamella (fig. 6C). Somewhat smaller, but largely similarly placed areas (fig.6B) were found by Armstrong et al. ('74). The small differences in size of the areas may be due to the different techniques employed. In contrast to Armstrong et al. ('741, however, we did not find evidence for a projection from the ventral lamella of the principal olive. Not a single labeled cell was ever observed here in our cases. The ventral lamella is only a thin sheet of cells closely adjacent to the medial accessory olive (which at these levels projects to the intermediate part), and careful microscopic scrutiny is necessary to decide in which of the two olivary subdivisions a particular labeled cell is situated. We are inclined to believe that slight misinterpretation of the placement of the recording electrode may explain Armstrong et al.'s conclusions about projections from the ventral lamella. Armstrong et al. (fig. 6B) indicate different origins for the projections to the longitudinal zones of the intermediate-lateral part of the anterior lobe. The HRP-method (fig. 6C) gives concordant results as concerns the lateralmost part which receives fibers from an area in the dorsal lamella. (This is further compatible with the brief statement by Voogd et al. ('75) that the principal olive projects to Voogd's zone D). However, we have no evidence that cells in the ventral lamella (see above) or in the rostral half of the medial accessory olive are labeled following injections involving the lateralmost cerebellar zone. As concerns the projections to the intermediate part from the medial half of the dorsal accessory olive and from the rostral medial accessory olive, our material does not permit us to establish clear topographical relations to the three longitudinal zones of the intermediate part shown in the diagram of Armstrong et al. (fig. 6B). However, as mentioned above, there is a suggestion that, as concluded by these au-
101
thors, the area in the medial accessory olive projects to a middle zone of the intermediate part, flanked by one zone medially and one laterally which receive fibers from the medial part of the dorsal accessory olive. According to Voogd et al. ('75) the rostral part of the medial accessory olive projects to his zone q, the rostral part of the dorsal accessory olive to zones C,-C,. 2. The longitudinal subdivision of the cerebellum and the olivocerebellar projection to the anterior lobe The idea of a longitudinal subdivision of the cerebellum suggested 50 years ago by Hayashi ('24) and Jakob ('281, received support from studies of the corticonuclear projection (Jansen and Brodal, '40, '42, and others). In recent years anatomical (Voogd, '64, '67, '69; Korneliussen, '68; '69; Courville, '75; Voogd et al., '75) and physiological (Oscarsson, '69, '73; Armstrong et al., '73a,b,c, '74) studies have demonstrated a far more detailed longitudinal zonal subdivision. Unfortunately, due to the absence of reliable surface landmarks in the anterior lobe the particular zones outlined by different authors are difficult to compare. With the HRP-method conclusions concerning the projection from the olive to particular longitudinal zones can only be made with some approximation. Nevertheless, as appears from the preceding section and figure 6C, on many points the HRP-findings agree with those made by others concerning zonal projections. Thus, in agreement with Voogd ('64, '69) and Armstrong et al. ('74) two zones can be distinguished in each hemivemis of the anterior lobe? Furthermore, our material shows that the vermal zones extend throughout the anterior lobe as maintained by Voogd ('69)and Korneliussen ('68, '69). In contrast to the conclusion of Armstrong et al. ('74) but in agreement with Voogd ('69)and Voogd et al. ('75) the two vermal * While Voogd ('64, '691 distinguishes two zones (A and B)in each half of the vermis of the anterior lobe, Komeliussen ('68, '69) subdivides his medial zone (correspondingapparently to the vermid in three subzones. Physiologically Oscarsson distinguishes two, three or four zones, depending on the type of afferent input from the olive (Oscarsson, '73, his frg. 8).
102
ALF BRODAL AND FRED WALBERG
zones receive their fibers from different olivary regions (fig. 6C). As concerns the intermediate-lateral part of the anterior lobe,3 for technical reasons we have been able with the HRPmethod to outline definitely a zonal projection only to the lateralmost part (presumably corresponding to Voogd's zones D, and D2, Korneliussen's lateral zone) from part of the dorsal lamella. As mentioned above, it cannot be decided whether the two olivary regions indicated by open circles in figure 6, project to different cerebellar zones (fig. 6B) as found by Armstrong et al. ('741, but there is suggestive evidence for an anatomical confirmation of this point. Both of these projections supply the anterior lobe at least as far forward as lobule IV. Even if our data support the idea of a longitudinal organization of the anterior lobe, our HRP-studies performed so far clearly show that all olivary areas do not project to longitudinal strips throughout the cerebellum. Thus, as concerns the vermis, the nucleus 8 and the dorsomedial cell column (which project heavily to the large medial zone of the uvula, Brodal, '76, and to the fastigial nucleus, Hoddevik et al., '76) do not contain labeled cells either following injections in the anterior lobe or in vermal lobules VI,VII and VIII, except for a few in the case of lobule VII (Hoddevik et al., '76). Further, vermal lobules VI-VTII (Hoddevik et al., '76) receive their olivary afferents from parts of the medial accessory olive which, in spite of some overlapping, surround the area projecting to the vermis of lobules I-V. To some extent the same is seen also from the diagram of Armstrong et al. ('741, where parts of the medial accessory olive (their fig. 5, levels 12.5, 13.5, 14.5) are indicated as projecting only to the vermis of the anterior lobe. The zonal subdivisions of the intermediate-lateral part of the anterior lobe according to i
ABSTRACT Following injections of horseradish peroxidase (HRP) in the cerebellar cortex of the anterior lobe of the cat, the distribution of labeled cells in the inferior olive was mapped. The findings largely confirm those made previously in studies of olivary retrograde cell loss following cerebellar ablations (Brodal, '40b). In addition, they reveal further olivary areas projecting onto the anterior lobe, and permit a more detailed analysis of the pattern in this projection. Concerning major points the results are in agreement with physiological studies by Armstrong et al. ('74). They bring supporting evidence for a longitudinal zonal pattern in the anterior lobe (fig.6C). The middle zone of the uermis receives its fibers from a large central area in the caudal half of the medial accessory olive, a lateral zone of the vermis from the lateral half of the dorsal accessory olive. Both olivary areas project to the corresponding cerebellar zone throughout lobules V-I. The lateralmost part of the anterior lobe (lobules IV-V) receives afferents from an area in the dorsal lamella of the principal olive. The intermediate part of lobules IV-V receives afferents from the medial half of the dorsal accessory olive and from an area in the rostral half of the medial accessory olive. There is suggestive evidence that the latter projects to a middle zone, the former to a medial and a lateral zone within the intermediate part as found physiologically. Conclusions concerning projections to the intermediate part of lobules 111-11 could not be made. The findings in this and preceding studies with the HRP-method show that the concept of a longitudinal pattern in the cerebellum is scarcely generally valid of the entire olivocerebellar projection. Within the projections of the lateral half of the dorsal accessory olive and the area in the rostral part of the medial accessory olive there appears to be a topical relation with the folial pattern in the anterior lobe. An analysis of the findings with reference to the afferents traced anatomically to the various olivary areas permits some conclusions as to the functional role of the olivary areas. Comparison with Oscarsson's ('73) diagram of the sites of termination of two of the spinal-olivary pathways (his DF-SOCP and VF-SOCP) permits an anatomical explanation as concerns the projections to the vermis, while correlations as concerns the intermediate part are less satisfactory.
The horseradish peroxidase (HRP) method is well suited for anatomical investigations of the pattern of the olivocerebellar projection in the cat (Walberg et al., '76). The present study is a continuation of previous reports dealing with the olivary projections to the paramedian lobule J. COMP. NEUR., 172: 85-108.
(Brodal et al., '751, the uvula (Brodal, '76) and the vermal visual area (Hoddevik et al., '761, in which several new data have been brought forward. According to Brodal's (40b) studies of the retrograde cellular changes in the olive following partial cerebellar ablations in 85
86
ALF BRODAL AND FRED WALBERG
kittens (modified Gudden method, Brodal, '40a1, the anterior lobe receives its olivary afferents only from the accessory olives. More particularly the vermis receives fibers from the lateral half of the dorsal accessory olive and from the latero-caudal part of the caudal half of the medial accessory olive, while the lateral and intermediate part receive fibers from the medial half of the dorsal accessory olive (fig. 6A). Results obtained by autoradiographic tracing of olivocerebellar fibers (Courville, '75) do not contradict these findings. Recent physiological studies have in part confirmed and considerably extended the early anatomical observations. From cerebellar recordings following electrical stimulation of the inferior olive (cat) VanGilder and O'Leary ('70) concluded that: "The dorsal accessory olive appeared to be the sole origin of responses recorded from the paravermal sector of the anterior lobe." Further (their fig. 6) responses recorded from the vermis of the anterior lobe come from the middle and caudal parts of the medial accessory olive. More precise data were obtained by Armstrong et al. ('74) by recording antidromic potentials in identified parts of the olivary complex following electrical stimulation of points on the cerebellar surface. In their summarizing diagram (their fig. 5; see fig. 6B here) they distinguish in the anterior lobe six longitudinal zones and a rostral one, which receive their olivary afferents from particular, in part coinciding, regions of the olive. In addition to areas in the medial and dorsal accessory olives, approximately the same as those found by Brodal ('40b), they describe projections from more rostral parts of the medial accessory olive as well as from restricted parts of the dorsal and ventral lamella (cp. figs. 6A and 6B). As will be shown, the use of the HRPmethod has made it possible to map the olivary projections to the anterior lobe in considerable detail. Projection areas have been identified in addition to those described by Brodal ('40b). Concerning the
main points the present results agree with those of Armstrong et al. ('741, and they bring some additional information as well. MATERIAL AND METHODS
The experimental material consists of some 30 cats of which 21 could be used in the present study. For particulars concerning the methods employed and the principles of presentation, see Walberg et al. ('76) and Brodal et al. ('751, respectively. Briefly the procedure has been as follows: Surgery was performed under deep Nembutal anaesthesia. In parts of the anterior lobe which can be exposed by a craniotomy, largely lobule V and part of lobule IV of Larsell ('701, the injection of HRP was often made through a fine slit in the dura. Injections in other lobules were made stereotactically. Table 1gives the weight, amount and concentration of the HRP fluid injected and the survival time in the various cases. In one case HRP type Serva was used, in the others type Sigma VI P 8375. The animals were killed under deep Nembutal anaesthesia by intracardiac perfusion of a mixture of 0.4% paraformaldehyde and 1.25% glutaraldehyde with phosphate buffer. Dissected blocks were serially cut on the freezing microtome at 50-6Op. From every group of five sections (transverse of the brain stem, sagittal of the cerebellum) two were treated with 3.3'-diaminobenzidine according to the method of Graham and Karnovsky ('66) and mounted on slides. One of the sections was weakly stained with cresyl violet or thionine, the other was left unstained. The distribution of labeled cells in the inferior olive, marked under the microscope in drawings of each section mounted, was transferred to a standard diagram of the olive as seen in 15 equally spaced transverse sections (fig. 1C). In order to facilitate comparisonsbetween cases, these findings were transferred to a diagram of a reconstruction of the olivary complex as imagined unfolded (Brodal, '40b). The occurrence of labeled cells in the olive is shown by dots. The distribution of stained
87
OLIVOCEREBELLAR PHOJECTION TO ANTERIOR LOBE
TABLE 1 Peroxidase injected Cat B.St.L.
Weight
kg
623 625 629 630 648 654 665 689 700 701 704 706 707 709 710 712 713 714 715 723 725
2.3 0.5 2.2 2.4 1.8 1.o 3.0 1.8 2.1 1.9 3.0 3.2 3.2 3.2 2.0 2.5 2.1 2.5 3.5 2.2 4.0
Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Sigma VIP Serva Siema VIP
cortical areas is indicated by hatchings in a diagram of the cerebellar surface, taken from Larsell (’70).Areas where the molecular layer is stained are shown in black. The term “vermis” will be used for the cortex overlying the fastigial nucleus. The more lateral parts of the anterior lobe will be referred to as the “intermediate part”, and the lateralmost parts of lobules IV and V will be called the “lateral part”. RESULTS
Following injections of identical amounts of fluid with the same concentration of HRP there may be considerable variations in the extension of the stained cerebellar areas (for particulars, see Walberg et al., ’76). Special attention has been paid to staining of the molecular layer, since the olivary fibers (at least many of them) end as climbing fibers. The staining of the cortex is often irregular, often the tops of some folia are unstained while their sides are well stained. Diffusion of HRP in the cortex occurs particularly easily along the length of the folia. The identification of the particular stained folia and lobules in the anterior lobe offers no problems in sagittal sections. However, determination of the
Concentration
Amount
50%w/v 50%wlv 50%wlv 50%wlv 100%wlv 50%wlv 50%w/v 50%wlv 50%w/v 50%w/v 50%w/v 50%wlv 50%w/v 50%w/v 50%wlv 50%wlv 50%wlv 50%w/v 50%w/v 50%wlv 50%wlv
0.5 pl 0.25 pl 0.5 pl 0.5 pl 0.25 pl 0.75 pl 0.5 pl 0.4 pl 0.2 pl 0.25 pl 0.2 pl 0.2 p1 0.2 pl 0.2 pI 0.2 pl 0.2 p1 0.2 pl 0.2 pl 0.15 pl 0.2 pl 0.2 ul
Survival time in days
3 2 3 2 2 1 1
2 2 1
2 2 2 2 3 2 2 2 2 3 2
extension of cortical staining in the transverse direction, i.e., along the folia, can only be made with some approximation. In some cases there is spreading of the staining from the cortex to the underlying cerebellar nuclei, where some cells may be labeled (Walberg et al., ’76).Since the nuclei receive axons or collaterals of olivocerebellar fibers, spreading of HRP to the nuclei following cortical injections may be a source of error in mapping the distribution of labeled olivary cells. Injections of HRP in the nuclei have, therefore, been made in some cases. The labeled neurons in the olive are easily distinguished, particularly under dark field illumination (Walberget al., ’76). The intensity of labeling varies among cases, and in a particular case one region may often contain chiefly heavily labeled cells, while in another labeled area most cells are only weakly labeled. The distribution within the olivary complex of labeled neurons varies among cases according to the site and extension of the staining of the cortex of the anterior lobe. Two cases with staining of the cortex of the vermis and intermediate part of lobule V serve to illustrate some main features.
88
ALF BRODAL AND FRED WALBERG
In cat l3.t.L. 629 (fig. 1) the injection has resulted in staining of the cortex of the lower folia of lobule V with only weak staining of a small area of the adjoining folium of lobule VI. On the left the staining involves the cortex of the vermis and a considerable portion of the intermediate part, on the right it covers only the medial part of the vermis. The white matter in the lowermost folia of lobule V is stained (fig. 1A). From these folia some stained fibers can be followed to the rostral part of the left fastigial nucleus where a few cells are labeled. The right inferior olive contains labeled cells in both accessory olives, but none in the principal olive (figs. lC,D). In the dorsal accessory olive the caudal half, approximately, contains labeled cells, particularly dense medially, where at some levels (VIII-1x1 practically all cells are labeled. In the caudal half of the medial accessory olive there is a large central area which contains many heavily labeled cells, while relatively fewer and less heavily labeled cells occur laterally, caudally and medially to this area. Finally, there is an area of labeled cells medially in the rostral half of the medial accessory olive. In the left inferior olive less intensely labeled cells than in the right olive occur only in the caudal half of the medial accessory olive, mainly in the central part of the area containing labeled cells on the right. Very similar findings are made in cat l3.St.L. 714 (fig. 1, inset). There is cortical staining of the left anterior lobe almost as in the previous case, but the stained area is situated a little more rostrally with no diffusion to lobule VI and it reaches only to the midline. A few labeled cells are found in the left nucleus interpositus anterior. Labeled cells are found only in the right olive, in regions which correspond almost completely to those found in the right olive in cat B.St.L. 629 (fig. 1D).
The projection to the vermis of the anterior lobe Cases with HRP-staining restricted to
the vermis show that of the olivary areas containing labeled cells in the preceding two cases (fig. 11, only those in the caudal half of the medial accessory olive and in the lateral part of the dorsal accessory olive project to the vermis. While the former area projects to the vermis of all lobules of the anterior lobe, lobules I-V differ somewhat with regard to their projection from the dorsal accessory olive. Abbreuiations @,Nucleus @ (Brodal, '40b) of inferior olive D, Dorsal accessory olive d.cap., Dorsal cap d.l., Dorsal lamella of principal olive dm.c.col., Dorsomedial cell column (Mareschal, '34) Flocc., Flocculus F.pr., Primary fissure l., Lateral Lobant., Anterior lobe L.pm., Lobulus paramedianus M, Medial accessory olive m., Medial nucl. p, Nucleus @ (Brodal, '40b) of inferior olive P.fl.d., Paraflocculus dorsalis P.fl.v., Paraflocculus ventralis v.l., Ventral lamella of principal olive v.I.o., Ventrolateral outgrowth I-V, Lobules of anterior lobe (Larsell, '70) Fig. 1 Presentation of the findings in cat B.St.L. 629 with HRP-injection in lobule V. Above to the left (A) a drawing of a nearly midsagital section of the cerebellum, to the right (B) a diagram of part of the cerebellar surface (taken from Larsell, '70). The stained regions are hatched, black denotes staining of the molecular layer. In the middle (C)a standard diagram of a series of 15 equally spaced transverse sections through the inferior olive (taken from Brodal, '40b). The various subdivisionsof the olivary complex are indicated (see list of Abbreuiations). Dots show the occurrence of labeled cells, transferred from drawings of particular sections to the corresponding levels of the diagram. Density of dots indicates density of labeling of olivary neurons (each dot does not correspond to one labeled cell). Below (D) a diagram (from Brodal, '40b) of the inferior olive imagined unfolded into one plane by pulling it apart as indicated by arrows in the insets (1,2,3) below. In C and D the thin broken lines refer to borders between projection areas as determined by Brodal ('40b). The dotted areas are those shown in the row of sections (C).The entire right olive, projecting onto the left cerebellar half, is shown in D, of the left olive only the medial accessory olive is reproduced. Inset in lower left comer shows an almost corresponding injection of the cortex in another case. Compare text. For Abbreviations see list above.
89
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LOBE
R
3J Q -. C vL
cI
caudal Right
Left
PRINCIPAL OLIVE
DORSAL ACC OLIVE
Left
2
I
Right
I
z
Figure 1
3
I
90
ALF BRODAL AND FRED WALBERG
B.St.L.689
MEDIAL
MEDIAL
ACC. OLIVE ACC.OLIVE -~
rostral
PRINCIPAL OLIVE
DORSAL ACC.OLIVE
d
n
caudal
Left
-
Right
B.S t.L.654
i) Left
MEDIAL ACC.OLIVE
Right
MEDIAL ACC OLIVE
rostral n
coudol Left
Right
Fig. 2 Diagrams of the findings in three cases (cats B. St.L. 689, and 648) with HRP-injections in the vermis of lobule V. For each case is shown a drawing of the cerebellar surface, some representative transverse sections through the olive, and a diagram of the unfolded olivary complex. Symbols as in figure 1. For Abbreviations see list on p. 88.
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LORE
In cat B.St.L. 689 (fig.2A) the injection has resulted in staining of lobule V, on the left extending to the medial border of the intermediate part, on the right side involving the medialmost part of the vermis only. Some labeled fibers can be traced to the nucleus fastigii where they break up into terminals. There are no labeled cells in the nucleus. Both inferior olives contain labeled cells. On the right most of them are found centrally in the caudal half of the medial accessory olive, some in the dorsal accessory olive at caudal levels. In the left olive there are only a few labeled cells in the area of the medial accessory olive which contains labeled cells in the right olive. In cat B.St.L. 654 (fig.2B) there is good staining of a circumscribed area in the vermis of the caudal part of lobule V, almost symmetrical,with staining of the molecular layer except for the summits of most of the stained folia. The lateralmost parts of the vermis are free. There is no visible staining of the white matter outside the folia, and no labeled cells in the fastigial nucleus. Faintly labeled cells occur in practically identical areas centrally in the caudal half of both medial accessory olives, a little more marked on the right than on the left. On the right in addition two labeled cells were found caudally in the dorsal accessory olive at level V. An even more restricted distribution of labeled olivary cells is found in cat B.St.L. 648 (fig. 2C), where 0.25p1 of a concentrated suspension of HRP (100%w/v) was injected in the rostralmost folia of lobule V near the midline. The staining of the cortex extends a little more laterally on the left than on the right. A few weakly labeled cells are found in the caudal part of the right medial accessory olive, most of them in a small cluster at level VI. At the corresponding site on the left there are a few scattered labeled cells in the sections mounted. Following injections of the vermal region of lobules IV, 111, I1 and I, the distribution of labeled cells in the olive resembles that described above for lobule V, apart from
91
minor differences in the dorsal accessory olive. In cat I3.St.L. 709 (fig. 3A) a stereotactic injection through lobule V has resulted in staining of a restricted part of the v e m i s of lobule ZV, mainly to the right of the midline, particularly of the molecular layer. A few stained fibers can be followed to the right fastigial nucleus, but there are no labeled cells in this. In the inferior olive there are (mainly weakly) labeled cells on both sides. On the left, they occur in a central area in the caudal half of the medial accessory olive and almost throughout the rostrocaudal extent of the lateral half of the dorsal accessory olive, with a maximum at levels VII-IX. In the right olive labeled cells occur only in the central part of the caudal half of the medial accessory olive. In cat B.St.L. 723 (fig. 3A, inset) following a stereotactically made injection (HRP type Serva) through lobule VI there is staining of the caudalmost folia of the vermis of lobule ZVand the rostralmost folium of lobule V. On the right staining reaches approximately to the lateral border of the vermis, on the left it involves only the medialmost part of the vermis. There is no staining and no labeled cells occur in the cerebellar nuclei. In the inferior olive the findings correspond largely to those made in the previous case (B.St.L. 709, fig. 3A). Labeled cells occur on both sides centrally in the caudal part of the medial accessory olive, scanty on the right, and in the caudal part of the lateral half of the left dorsal accessory olive. In another cat (B.St.L. 706, not illustrated) the findings in the cerebellum and the olive correspond almost completely to those in the two preceding cases (fig.3A). In cat B.St.L. 701 (fig. 3B) the stereotactically made injection has resulted in good staining of the vermis of the rostral part of lobule ZZZ and the caudal part of lobule ZZ, most marked near the midline but extending laterally almost symmetrically. There are a few labeled cells in the rostral part of the fastigial nuclei. The labeling of neurons in the olive is on the whole rather weak,
92
ALF BRODAL AND FRED WALBERG
B.St.L.709
A
Left DORSAL ACC OLIVE
rdstral
PRINCIPAL OLIVE
MEDIAL ACC OLIVE
PRINCIPAL OLIVE
DORSAL ACC OLIVE
Right
MEDIAL ACC OLIVE
~
w XI
m caudal
1
B.St.L. 70 I
B
MEDIAL ACC OLIVE
m xv xnr XI m wr M
vm M VII
m N
m N caudal
I
B.St.L.700
I Right
and
Left
1B.St.L.725
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LOBE
but labeled cells occur on both sides, almost symmetrically, in one area centrally in the caudal half of the medial accessory olive and a more extensive area in the lateral part of the dorsal accessory olive, with a faint maximum at levels IX-X. Following an injection which results in staining of the vermis of some folia of lobule ZZZ (cat B3t.L. 704, fig. 3B, inset), the distribution of on the whole faintly labeled cells corresponds closely to that shown in the diagram of the olive in cat B.St.L. 701 (fig.3B).The number of labeled cells in the dorsal accessory olive is largest between levels XII-VIII. A further case (cat B.St.L. 725, fig. 3B, inset) with an injection of the vermis of lobules 11-111 shows a corresponding distribution of labeled olivary neurons as found in cat B.St.L. 704. Cat B.St.L. 700 (fig.3C). The tip of the stereotactically inserted needle has been in the white matter just beneath the base of lobule ZZ. On both sides of the midline, almost symmetrically, there is considerable staining of the white matter and very good staining of the molecular layer of the verma1 parts of the folia of lobule ZZ, especially anteriorly, and of the adjoining cortex of lobule 1. In the rostralmost part of the fastigial nuclei there is a light brownish color, and a few cells are labeled. In both olives well labeled neurons occur symmetrically in a rather circumscribed central part of the caudal half of the medial accessory and in almost the entire lateral half of the dorsal accessory olive. In the latter labeling extends practically to its rostral end. The majority of labeled cells occurs between levels X and XIII. In cat B.St.L. 665 (fig. 3C, inset) the staining of the cortex is almost identical to that in the former case, and the findings in the olive correspond closely to those described above. In cat B3t.L. 715 (fig. 3C, inset) staining
93
is restricted to the vermis of lobule 11, mainly on the right. It does not extend quite to the lateral border of the vermis. In the inferior olive well labeled cells occur bilaterally in the areas showing labeling in the two previous cases (B.3.L. 700 and 665, fig. 3C1, on the whole more abundant on the left than on the right, particularly in the dorsal accessory olive. On both sides this shows a maximum of labeling at levels XI-XIII.
Zntermediate and lateral parts of the anterior lobe Injections in the anterior lobe lateral to the vermis result in a distribution of labeled olivary cells different from that following injections in the vermis. In cat l3.St.L. 630 (fig. 4A) cortical staining, mainly of the molecular layer, is present in the lateralmost part of the left lobule V, while the vermal zone and most of the intermediate part are free. There is no staining of the cerebellar nuclei and there are no labeled cells in these. In the olive labeled cells are present only on the right side and in two fairly well circumscribed regions. One of them, containing weakly labeled cells, is found in the dorsal accessory olive, the other in the dorsal lamella of the principal olive, where on the whole, labeling is more marked. In cat B.St.L. 713 (fig. 4A, inset) a restricted injection in the lateralmost part of lobule V, with a little spreading to lobule VI, resulted in labeling of a few cells only in the part of the contralateral dorsal lamella labeled in cat B.St.L. 630 (fig.4A). In cat B.St.L. 625 (fig. 4B) an injection in the intermediate part of the left lobule V has spread to the adjoining folia of lobule VI. Some stained fibers can be traced towards the cerebellar nuclei, and a few labeled-cellsoccur in the left nucleus interpositus posterior. In the inferior olive labeled cells are found in two regions only on Fig. 3 Diagrams of the findings in 3 cases with in- the right side. One covers a considerable jections in the vermis of lobules IV-I1 (cats B.St.L. part of the medial half of the dorsal ac709, 701 and 7001,presented as in figure 2. Insets cessory otiue, the other is found medially in show stained areas in other cases with findings in the the rostral half of the medial accessory olive which are closely similar to those in the corresponding main case (see text). For Abbreuiatim see olive. An injection in the lateral and intermelist on p. 88.
94
ALF BRODAL AND FRED WALBERG
B.St.L.630
Left
Right
rost ral
caudal
-
Right
Fig. 4 Diagrams of the findings in two cases (cats B.St.L.630 and 625) with injections in the lateral-intermediatepart of lobule V, presented as in figure 3. Inset shows staining of cortex in a third case (see text). For Abbreviations see list on p. 88.
diate parts of the left lobule ZVkat B.St.L. 623, fig. 5A) gives a distribution of labeled cells resembling that following injections in the lateral and intermediate parts of lobule V when taken together (figs. 4A,B). There are no signs of diffusion of HRP from the restricted stained area of the cortex to the nuclei. The right olive shows three areas of labeled cells, one in the medial half of the dorsal accessory olive, another in the dorsal lamella, and a third in the rostra2 half
of the medial accessory olive. These areas overlap those containing labeled cells following injections in the lateral-intermediate parts of lobule V, but are a little differently placed. Labeling is somewhat stronger in the medial accessory olive than in the two other regions. An almost completely corresponding injection of the left lateral-intermediate part of lobule IV was achieved in cat B.St.L. 710 (fig. 5A, inset). Some stained fibers can be
95
OLIVOCEREBELLAR PROJECTION TO ANTERIOR COBE
6.St.L. 623
-
,
I
MEDIAL ACC OLIVE
B.St.L.707
B
PRINCIPAL OLIVE
DORSAL ACC OLIVE
M
Left MEDIAL ACC OLIVE
PRINCIPAL OLIVE
rostra1
Right
DORSAL ACC OLIVE
caudal Right
Fig. 5 Diagrams of two cases (cats B.St.L. 623 and 707) with injections in the lateral-intermediate part of lobule IV, presented as in figure 3. Inset shows staining of the cortex in a third case. For Abbreuiations see list on p. 88.
traced from the injected area to the nucleus interpositus anterior and posterior, which contain a few labeled cells. Well labeled neurons are found in the right inferior olive in regions which are almost identical to those found in the previous case (fig. 5A). A somewhat different pattern in the distribution of labeled olivary neurons is seen
when the lateral part of lobule IV (its molecular layer) is not stained. In cat. B.St.L. 707 (fig. 5B) a stereotactic injection through lobule VI has resulted in staining of folia of lobule ZVon the left side. The stained cortex (particularly the molecular layer) does not include the lateral end of the lobule, medially it encroaches a little on the vermis. There is considerable stain-
96
ALF BRODAL AND FRED WALBERG
ing of the white matter of lobule IV. A central part of the nucleus interpositus anterior is faintly brownish stained and contains some labeled cells, as well as labeled fibers entering from lobule IV. The fastigial nucleus is not affected. In the olive labeled cells are found only in two regions on the right. There is a restricted area containing mainly weakly labeled cells caudally in the rostral half of the medial accessory oliue, and a larger area centrally in the dorsal accessory oliue, where there is a marked maximum of labeled cells at levels VIII-IX. In four cats attempts were made to achieve staining of the intermediate part of the anterior lobe of lobules 11-111. Conclusions concerning olivary afferents to this part could not be drawn because in all cases there was considerable spreading of HRP to the dentate nucleus.
Znterpretation of findings Labeled olivary neurons can be identified with great certainty and their topographical distribution can be mapped with precision. In spite of the difficulties referred to above involved in determination of the mediolateral extent of the injected cerebellar area, it is striking to note the almost complete correspondence between olivary areas labeled in different cases in which the same cerebellar area has been stained by HRP. This fact further shows that the errors which might be induced by transferring the findings in the olive in individual cases to a standard diagram, can be only minor ones. An analysis of our material gives a picture of the projection as shown in figure 6C. The olivary projections to the vemnis of the anterior lobe (fig. 6C) Whenever the vermis of the anterior lobe is included in the cortical area stained following injections of HRP, there is labeling of cells in a central area in the caudal half of the medial accessory olive. This is the case for each of the lobules V-I (figs. 1-31. Apart from minor variations in size of the area and in the density of labeled cells, this area is almost identical in all cases.
In many of our cases with vermal injections there is in addition labeling of cells in the lateral half, approximately, of the dorsal accessory olive. However, this occurs only when the staining (molecular layer) extends to the lateral border of the vermis. This is clear as concerns lobules V and IV (compare the two sides of the injected cerebellar areas in figs. 1,2A and 3A). (The labeling in the medial part of the right dorsal accessory olive in fig. 1 is due to staining of the intermediate part of lobule V, see below). In lobules 1-111staining entirely restricted to the medial part of the vermis has not been achieved. However, from a comparison of these cases it appears that the labeling in the lateral half of the dorsal accessory olive is more marked the further laterally in the vermis the staining has spread (for example, cats B.St.L. 704, fig. 3B, inset; 715, fig. 3C, inset). In cases with staining mainly of the intermediate part of Fig. 6 Diagrams of the olivocerebellar projection onto the anterior lobe. A, As determined with the modified Gudden method by Brodal (‘40b). Above a diagram of the anterior lobe (after Larsell, ’70), and some representative transverse sections through the olive from caudal (I) to rostral (XV) (cp. fig. 1). Below a reconstruction of the olivary complex as imagined unfolded in one plane (cp. fig. 1). B. The projection as Found by Armstrong et al. (‘74).To the left their drawing of the anterior lobe, in the middle the olivary areas projecting to different zones in the anterior lobe. The order of the sections has been reversed in order to Correspond to the sequence used here. The Roman numerals indicate the approximate levels of the olive (as marked here) to which the levels reproduced in Armstrong et al.’s diagram correspond. On the basis of this, their findings are transferred to our diagram of the unfolded olive (to the right). C. The projection as determined in the present study. To the left surface of the anterior lobe, to the right a series of representative transverse sections through the olive. Below the projection seen as entered in the diagram of the unfolded olive. The intermediate part is shown as receiving fibers from the medial half of the dorsal accessory olive and a region in the rostral part of the medial accessory olive (open rings). Not shown is the suggestive conclusion that the latter projects to a middle zone of the intermediate part, the former to a medial and a lateral zone, as in figure 6B. Information of olivary projections to areas left white in the cerebellum was not obtained. See text. For AbbreoMtions see list on p. 88.
97
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LOBE Ivernxs
lvermis
I
xm
caudal
m
rostra1
MEDIAL ACC OLIVE
PRINCIPAL OLIVE
rostrol
caudal
Figure 6
DORSAL ACC OLIVE
98
ALF BRODAL AND FRED WALBERG
Fig. 7 Diagrams to illustrate the topographical pattern in the projection of the medial part of the rostral medial accessory olive and the lateral part of the dorsal accessory olive onto the anterior lobe. See text.
lobules 11-111,there is labeling in the lateral half of the dorsal accessory olive when staining involves the lateral zone of the vermis (cat B.St.L. 707, fig. 5B). It appears thus that the two projection zones in the vermis extend at least to lobule I1 as indicated in figure 6C. It follows from these findings that the vermis of the anterior lobe throughout its length receives fibers from a central area in the caudal half of the medial accessory olive and from the lateral half of the dorsal accessory olive. Furthermore, the former area projects onto the medial zone of the vermis, the latter to a lateral zone of the uermis (fig. 6C). While no evidence has been found for a topical arrangement within the projection to the vermis from the medial accessory olive, there is evidence for an arrangement of this type in the projection from the lateral half of the dorsal accessory olive (fig. 7). In spite of considerable overlapping, the area labeled in this is on the whole situated progressively further caudally as the injection is placed more posteriorly, from lobules 1-11 to lobule V. Furthermore, the part of the area which contains the greatest proportion of labeled cells shows a corresponding variation.
The olivary projections to the intermediate-lateral part of the anterior lobe (fig. 6C) These arise from other olivary areas than do the projections to the vermis. Following injections of the intermediate-lateral part, labeling is found in three places, one area in the rostral half of the medial accessory
olive, another in the medial h a l , approximately, of the dorsal accessory olive, and a third in part of the dorsal lamella of the principal olive. These three olivary areas appear to project to different regions of the anterior lobe, apparently different longitudinal zones. Most convincing in this respect is the evidence for the projection from the area in the dorsal lamella (fig. 6C), since labeling occurs here only when the staining of the cortex (molecular layer) of the anterior lobe includes its most lateral regions (figs. 4A, 5A). There is some evidence from cases not described that the lateralmost part of lobule I11 likewise receives fibers from this olivary area. Since we have not been able to obtain reliable information concerning the projection to the intermediate part of lobules 11-111, these are left without symbols in figure 6C. The two areas in the accessory olives mentioned above, are most often both labeled following staining of the intermediate part. In the dorsal accessory olive the area labeled in our relevant cases (figs. 1, 4A,B, lobule V; figs. 5A,B, lobule IVI, when taken together cover its entire medial half (open rings, fig. 6C). It appears that this area overlaps somewhat that of the lateral half of the dorsal accessory olive projecting onto the lateral zone of the vermis (open triangles). No clear topographical relations can be discerned in the projection from the medial half of the dorsal accessory olive. In the area in the rostral half of the medial accessory olive, however, it appears (fig. 7) that fibers to lobule V take origin mainly from its medial part, while those to lobule IV arise more laterally (Cp. figs. 1 and 4B with figs. 5A,B). Whether these projections from the dorsal and medial accessory olives supply different zones of the intermediate part cannot be finally decided. However, in two cases with injections involving the intermediate part there is labeling in the dorsal accessory olive but not in the rostral part of the medial accessory olive. In one of these cases (cat B.St.L. 720, not illustrated) only the medialmost region of the intermediate part is stained. This may suggest, as con-
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LOBE
cluded by Armstrong et al. ('74,see fig.6B here), that there are in the intermediate part a lateral and a medial zone (open rings in fig. 6B) receiving fibers from the dorsal accessory olive on each side of a middle zone (vertical hatching) which receives fibers from the rostral part of the medial accessory olive. This tentative conclusion is not indicated in the diagram figure 6C. In some of the cases with injections of the anterior lobe there is evidence of spreading of some HRP to the cerebellar nuclei. To some extent the olivary areas projecting to the nuclei interpositi (Brodal et al., '75, their fig. 6) and to the fastigial nucleus (Hoddevik et al., '76, their fig. 4) coincide with those projecting to the intermediate-lateral part or the vermis of the anterior lobe, respectively. However, the diffusion of HRP to the nuclei in some of our cases is too modest to explain the marked labeling in these common projecting olivary areas. Furthermore, the same olivary areas are found labeled in cases without evidence of diffusion to the nuclei (cats B.St.L. 623, 630, 648, 654, 665, 689, 704, 706, 709, 713, 715). We feel, therefore, that the moderate spreading of HRP to the nuclei in some of our cases does not invalidate the conclusions drawn above. DISCUSSION
As discussed elsewhere (Walberg et al., '761, the use of the HRP method permits a precise mapping of the sites of labeled cells in the inferior olive following injections of HRP in the cerebellar cortex. The cortical area where axon terminals have been exposed to HRP, as witnessed by brown staining of the tissue, can be less precisely determined. However, when a number of cases with similar and different injection sites are compared, it is possible to outline some main features in the olivocerebellar projection to the anterior lobe as described above. The present material as well as previous studies (Brodal et al., '75; Brodal, '76; Hoddevik et al., '76) indicate that distribution of labeling in the olive is correlated to the extent of staining of the molecular layer.
99
From studies in other regions of the CNS it has been concluded that the distribution of retrogradely labeled neurons is smaller than might be predicted from the extent of diffusion of HRP as witnessed by the peroxidase positive area at the injection sites of their terminals (Bunt et al., '76, see this paper for further references). The correspondence in our material between the size of the stained part of the cerebellar cortex and the size of the labeled areas of the olive indicates that the situation may be different in the olivocerebellar projection, As found by a number of authors, our studies confirm that the olivocerebellar projection to the anterior lobe is purely crossed. The frequently observed astonishingly sharp transition between an olivary area containing labeled cells towards one free of such cells, shows that the olivocerebellar localization must be very precise. When there are more gradual transitions, this may be due to the fact that in the periphery of an injected area the concentration of HRP must be expected to be less than in the center, and accordingly not all terminals ending here may have taken up sufficient HRP to show labeling. The same fact may explain that the cells in the center of a labeled area are often more heavily labeled than those in the periphery. The uncontrollable spreading of HRP within the cortex prevents us from deciding whether cortex deep in the sulci differs in its projection from the superficial cortex.
1. Topography of the olivocerebellar projection to the anterior lobe It is of particular interest to consider our findings in relation to the electrophysiological studies of Armstrong et al. ('74) and to the data on a longitudinal zonal subdivision of the cerebellum. Even if minor errors are apt to occur in our transposition of the findings of Armstrong et al. ('74)to our diagram of the unfolded olive (fig. 6B), this will largely reveal itself in slightly incorrect markings of the borders of particular olivary projection areas. Concerning main points there is good agreement between
100
ALF BRODAL AND FRED WALBERG
the results of the two kinds of study. Discrepancies on certain points are probably largely due to differences between the methods used. The HRP method has made it possible to obtain information of some points which could not be clarified by the method of antidromic recording, and others which were not demonstrated with the anatomical method used by Brodal (’40b). The vermis of the anterior lobe was found by Brodal (’40b) to receive olivary afferents from the lateral half of the dorsal accessory olive and the caudolateral part of the medial accessory olive (fig. 6A). The former area corresponds to that determined with the HRP method (fig. 6C), the latter is found here to be situated somewhat more medially. Armstrong et al. (fig. 6B) likewise found two areas to project to the vermis of the anterior lobe (oblique hatching and triangles), one caudally in the medial accessory olive, the other caudolaterally in the dorsal accessory olive. The former area corresponds well with our mapping, except that we found only few labeled cells most laterally (this region projects largely to lobule VIII, Hoddevik et al., ’7611. The area in the dorsal accessory olive is indicated by Armstrong et al. as comprising only the caudal part of the lateral half of the dorsal accessory olive, while we find this area to extend practically to its rostra1 end. This is most likely due to the fact that Armstrong et al. did not explore the most anterior folia (1-111, probably in part IV). This assumption is strengthened by our observation that there is a topical pattern, although with considerable overlapping, in the projections to lobules I-V (fig. 7) from this part of the olive. Evidence that the lateral part of the dorsal accessory olive projects to anterior regions of the intermediate part as found by Armstrong et al. (‘741, dotted area in fig. 6B) was not obtained in our study. It cannot be seen from their diagram (their fig. 5, fig. 6B here) to what lobule the dotted area belongs. If this site of stimulation has been more medially than supposed in lobule I11 or IV, the dottings in the dorsal accessory
olive in their diagram (fig. 6B here) would agree completely with our findings (fig. 6C). Apart from the latter discrepancy there appears to be a satisfactory agreement between the findings of Armstrong et d. (’74) and our own concerning the total olivary areas projecting onto the anterior lobe vermis. However, while we found the projections to the medial and lateral vermal zones to take origin from separate olivary regions (oblique hatching and triangles, fig. 6C), Armstrong et al. found each of these regions to project to both zones (fig. 6B). We feel that our findings concerning separate projections are fairly decisive. Even if the current at the electrode site will not stimulate terminals more than 1.3 mm away (Armstrong et al., ’741, it is conceivable that spread of current to the other vermal zone on stimulation of one of them may explain their findings and has obscured the separate projections to the two zones of the vermis. Our interpretation receives support from the study of Voogd et al. (‘751, who with anterograde degeneration and tritiated leucine transport methods found “that the caudal parts of the medial and dorsal accessory olives project to zones A and B, respectively”, i.e., medial and lateral vermal zones. With regard to the olivary projection to the intermediate-lateral part of the anterior lobe, the HRP-study has confirmed Brodal’s (’40b) conclusion (fig. 6A) that the medial half, approximately, of the dorsal accessory olive projects to the intermediate part, at least as concerns lobules IV and V. VanGilder and O’Leary (’70) likewise concluded that all olivary afferents to this part come from the dorsal accessory olive. 1 In their fgure 2 Armstrong et al. (‘74) indicate antidromic potentials in the lateralmost part of the medial accessory olive (at about level IV of our diagram) following stimulationof the midvermis of lobule V as well as of what appears to be lobule VIII. It is of interest that according to their description (p. 2951 it appears “that the two stimuli were to a considerable extent exciting the axons of different olivary neurones,”while some were antidromicaly invaded following stimulation of either zone. This is readily explained if the site of their recording electrode in the latter situation has been at the transitionbetween the olivary areas projecting to lobules VIII and V, respectively, according to the HRP findings.
OLIVOCEREBELLAR PROJECTION TO ANTERIOR LOBE
The HRP-method, in addition, has revealed two other olivary areas projecting to the intermediate-lateral parts, one in the rostral half of the medial accessory olive, another in the dorsal lamella (fig. 6C). Somewhat smaller, but largely similarly placed areas (fig.6B) were found by Armstrong et al. ('74). The small differences in size of the areas may be due to the different techniques employed. In contrast to Armstrong et al. ('741, however, we did not find evidence for a projection from the ventral lamella of the principal olive. Not a single labeled cell was ever observed here in our cases. The ventral lamella is only a thin sheet of cells closely adjacent to the medial accessory olive (which at these levels projects to the intermediate part), and careful microscopic scrutiny is necessary to decide in which of the two olivary subdivisions a particular labeled cell is situated. We are inclined to believe that slight misinterpretation of the placement of the recording electrode may explain Armstrong et al.'s conclusions about projections from the ventral lamella. Armstrong et al. (fig. 6B) indicate different origins for the projections to the longitudinal zones of the intermediate-lateral part of the anterior lobe. The HRP-method (fig. 6C) gives concordant results as concerns the lateralmost part which receives fibers from an area in the dorsal lamella. (This is further compatible with the brief statement by Voogd et al. ('75) that the principal olive projects to Voogd's zone D). However, we have no evidence that cells in the ventral lamella (see above) or in the rostral half of the medial accessory olive are labeled following injections involving the lateralmost cerebellar zone. As concerns the projections to the intermediate part from the medial half of the dorsal accessory olive and from the rostral medial accessory olive, our material does not permit us to establish clear topographical relations to the three longitudinal zones of the intermediate part shown in the diagram of Armstrong et al. (fig. 6B). However, as mentioned above, there is a suggestion that, as concluded by these au-
101
thors, the area in the medial accessory olive projects to a middle zone of the intermediate part, flanked by one zone medially and one laterally which receive fibers from the medial part of the dorsal accessory olive. According to Voogd et al. ('75) the rostral part of the medial accessory olive projects to his zone q, the rostral part of the dorsal accessory olive to zones C,-C,. 2. The longitudinal subdivision of the cerebellum and the olivocerebellar projection to the anterior lobe The idea of a longitudinal subdivision of the cerebellum suggested 50 years ago by Hayashi ('24) and Jakob ('281, received support from studies of the corticonuclear projection (Jansen and Brodal, '40, '42, and others). In recent years anatomical (Voogd, '64, '67, '69; Korneliussen, '68; '69; Courville, '75; Voogd et al., '75) and physiological (Oscarsson, '69, '73; Armstrong et al., '73a,b,c, '74) studies have demonstrated a far more detailed longitudinal zonal subdivision. Unfortunately, due to the absence of reliable surface landmarks in the anterior lobe the particular zones outlined by different authors are difficult to compare. With the HRP-method conclusions concerning the projection from the olive to particular longitudinal zones can only be made with some approximation. Nevertheless, as appears from the preceding section and figure 6C, on many points the HRP-findings agree with those made by others concerning zonal projections. Thus, in agreement with Voogd ('64, '69) and Armstrong et al. ('74) two zones can be distinguished in each hemivemis of the anterior lobe? Furthermore, our material shows that the vermal zones extend throughout the anterior lobe as maintained by Voogd ('69)and Korneliussen ('68, '69). In contrast to the conclusion of Armstrong et al. ('74) but in agreement with Voogd ('69)and Voogd et al. ('75) the two vermal * While Voogd ('64, '691 distinguishes two zones (A and B)in each half of the vermis of the anterior lobe, Komeliussen ('68, '69) subdivides his medial zone (correspondingapparently to the vermid in three subzones. Physiologically Oscarsson distinguishes two, three or four zones, depending on the type of afferent input from the olive (Oscarsson, '73, his frg. 8).
102
ALF BRODAL AND FRED WALBERG
zones receive their fibers from different olivary regions (fig. 6C). As concerns the intermediate-lateral part of the anterior lobe,3 for technical reasons we have been able with the HRPmethod to outline definitely a zonal projection only to the lateralmost part (presumably corresponding to Voogd's zones D, and D2, Korneliussen's lateral zone) from part of the dorsal lamella. As mentioned above, it cannot be decided whether the two olivary regions indicated by open circles in figure 6, project to different cerebellar zones (fig. 6B) as found by Armstrong et al. ('741, but there is suggestive evidence for an anatomical confirmation of this point. Both of these projections supply the anterior lobe at least as far forward as lobule IV. Even if our data support the idea of a longitudinal organization of the anterior lobe, our HRP-studies performed so far clearly show that all olivary areas do not project to longitudinal strips throughout the cerebellum. Thus, as concerns the vermis, the nucleus 8 and the dorsomedial cell column (which project heavily to the large medial zone of the uvula, Brodal, '76, and to the fastigial nucleus, Hoddevik et al., '76) do not contain labeled cells either following injections in the anterior lobe or in vermal lobules VI,VII and VIII, except for a few in the case of lobule VII (Hoddevik et al., '76). Further, vermal lobules VI-VTII (Hoddevik et al., '76) receive their olivary afferents from parts of the medial accessory olive which, in spite of some overlapping, surround the area projecting to the vermis of lobules I-V. To some extent the same is seen also from the diagram of Armstrong et al. ('741, where parts of the medial accessory olive (their fig. 5, levels 12.5, 13.5, 14.5) are indicated as projecting only to the vermis of the anterior lobe. The zonal subdivisions of the intermediate-lateral part of the anterior lobe according to i
