An autoradiographic study of the organization of the efferent connections of the hippocampal formation in the rat.

An Autoradiographic Study of the Organization of the Efferent Connections of the Hippocampal Formation in the Rat’ L. W. SWANSON AND W. M. COWAN Department of Anatomy and Neurobiology, Washington University School of Medicine, St. Louis, Missouri 631 10

ABSTRACT The efferent connections of the hippocampal formation of the rat have been re-examined autoradiographically following the injection of small quantities of 3H-amino acids (usually 3H-proline) into different parts of Ammon’s horn and the adjoining structures. The findings indicate quite clearly that each component of the hippocampal formation has a distinctive pattern of efferent connections and that each component of the fornix system arises from a specific subdivision of the hippocampus or the adjoining cortical fields. Thus, the precommissural fornix has been found to originate solely in fields CAl-3 of the hippocampus proper and from the subiculum; the projection to the anterior nuclear complex of the thalamus arises more posteriorly in the pre- and/or parasubiculum and the postsubicular area; the projection to the mammillary complex which comprises a major part of the descending columns of the fornix has its origin in the dorsal subiculum and the pre- and/or parasubiculum; and finally, the medial cortico-hypothalamic tract arises from the ventral subiculum. The lateral septal nuclei (and the adjoining parts of the posterior septal complex) constitute the only subcortical projection field of the pyramidal cells in fields CAI-3of Ammon’s horn. There is a rostral extension of the pre-commissural fornix to the bed nucleus of the stria terminalis, the nucleus accumbens, the medial and posterior parts of the anterior olfactory nucleus, the taenia tecta, and the infralimbic area, which appears to arise from the temporal part of field CA, or the adjacent part of the ventral subiculum. The projection of Ammon’s horn upon the lateral septal complex shows a high degree of topographic organization (such that different parts of fields CAI and CA, project in an ordered manner to different zones within the lateral septal nucleus). The septal projection of “CA,” and field CA, is bilateral, while that of field CA, is strictly unilateral. In addition to its subcortical projections, the hippocampus has been found to give rise to a surprisingly extensive series of intracortical association connections. For example, all parts of fields CAI,CA, and CA, project to the subiculum, and at least some parts of these fields send fibers to the pre- and parasubiculum, and to the entorhinal, perirhinal, retrosplenial and cingulate areas. From the region of the preand parasubiculum there is a projection to the entorhinal cortex and the parasubiculum of both sides. That part of the postsubiculum (= dorsal part of the presubiculuml which we have examined has been found to project to the cingulate and retrosplenial areas ipsilaterally, and to the entorhinal cortex and parasubiculum bilaterally.

Since the efferent C!oMections of the hippocampal formation 2 have been studied for more than a century’ in a variety Of species and with almost every available neuroanatomical method, it might be assmed that by this time our understanding of their organization would be reasonably J. COMP. NEUR., 172: 49-84.

I This work was supported in part by Grants NS-10943 and MH-24604 from the National Institutes of Health. * In this, as in our previous papers (Gottlieb and Cowan. ’73, Schlessinger, Cowan and Gottlieb, ’75) we have used the general term “hippocampal formation”to include the various fields of Ammon’s horn, the dentate gyrus, and the related cortical fieldsof the subicular and entorhind regions. The unqualifield terms “hippocampus” or ”hippocampus proper” are used synonymously with h m o n ’ s horn.

49

50

L. W. SWANSON AND W. M. COWAN

complete. However, it is evident from a perusal of the relevant literature that there still are a number of issues about which comparatively little is known, and several about which we have reason to believe our knowledge may be based on erroneous, or at least misleading, evidence. As an example of the former we might cite the paucity of information about the caudally-directed efferents of the hippocampus. Because of the widely-held view that the fimbria and dorsal fornix represent the major, if not the sole, efferent pathways of the hippocampus, until recently comparatively little attention had been paid to the possibility that a substantial proportion of the output of Ammon’s horn may be directed caudally to the subicular, entorhinal, and retrosplenial areas of the cerebral cortex. Similarly, insufficient consideration has been given to the possibility that some of the efferent projections of the hippocampus that have been described in experimental material (followingthe placement of destructive lesions in the fimbria, the dorsal fornix or the various subfields of Ammon’s horn) could be due to the interruption of pathways arising in more remote regions of the hippocampal formation. At the same time this problem -which is inherent in the use of all degeneration techniques - has made the analysis of the projection of individual cytoarchitectonic fields within the hippocampal formation extremely difficult (Raisman et al., ’66) and has even led some, as recently as 1974, to suggest that there may be no distinctive organization in the efferent projection of the different fields of Ammon’s horn, at least to the septa1 complex (Siege1 et al., ’74). With the introduction of the autoradiographic method for tracing connections in the central nervous system (Cowan et al., ’72)it seemed opportune to re-examine the connections of the hippocampal formation from this point of view. We felt encouraged to pursue this when, during a study of the commissural connections of the hippocampus and dentate gyrus (Cottlieb and Cowan, ’73)we found, to our surprise, that in many experiments in which we had in-

jected fairly large amounts of either 3Hleucine or 3H-proline into one or the other of the two major cytoarchitectonic fields of Ammon’s horn,3 there was no evidence of transported label in either the mammillary complex or in the anterior thalamic nuclei, which for some years have been regarded as the principal subcortical projection fields of the hippocampus. On the other hand, in several cases in which the subicular region had been inadvertently labeled by the isotope injection, we could consistently find transported label in the hypothalamus and thalamus. It thus seemed appropriate to systematically explore the efferent connections of the entire hippocampal formation using small, localized injections of tritium-labeled amino acids to determine whether each cytoarchitectonic field into which the formation has been divided is characterized by a distinctive pattern of efferent connections, and to establish more precisely what these connections are. Although we have not succeeded in discretely labeling all the various fields of the subicular and entorhinal regions, we have a sufficient number of cases with injections localized to each of the principal fields of Ammon’s horn to justify the publication of our findings at this time. From these successful experiments it has been possible to establish that there is a high degree of topographic organization within the efferent projections of the hippocampal region and to show that the contribution of the fornix system to the hypothalamus and thalamus arises not within Ammon’s horn, as is generally believed, but from the dorsal and ventral parts of the subiculum and the adjoining pre- and parasubiculum.4 MATERIALS AND METHODS

The brains of more than 80 rats were 3 Fields CA, and CA, of Lorente de N6 (‘34).Field CA, cDrresponds fairly closely to the regio superior of Cajal (‘11); field CA,, with the adjoining fields CA, and CA,, to Cajal’s regio inferior. The latter is distinguished cytoarchitectonically by the presence of generally larger pyramidal cells than those in the regio superior. A preliminary report of this last finding has already been published (Swanson and Cowan, ’75).

HIPPOCAMPAL EFFERENTS

available for this study. In most of these animals a small injection of tritiated proline (L 2,3-3H-proline, specific activity 24 Ci/ mmole, New England Nuclear) was made stereotaxically into one or another part of the hippocampal formation. In a few cases a comparable volume (10-40 nl) of a mixture of 3H-proline, 3H-leucine and 3H-lysine was injected. The injections were made through a specially beveled needle over a period of 15to 20 minutes, using the stereotaxic coordinates of Konig and Klippel (’63).After survival periods of 24 to 72 hours the animals were perfused with a buffered 10% formalin solution, and the brains processed for autoradiography in the manner described by Cowan et al. (’72). From each brain a 1-in-10, or a 1in-5, series of autoradiographs was prepared throughout the extent of the forebrain and the rostral part of the midbrain. RESULTS

In order to provide a reasonably complete account of the efferent connections of the hippocampal formation injections of the labeled amino acids were aimed at all parts of Ammon’s horn and the adjoining subiculum; in addition various parts of the pre-, the post-, and the parasubiculum, and the medial and lateral entorhinal areas were labeled, incidentally in the first group of experiments and intentionally when it became clear that these areas contributed significantly to the projections of the fimbria and dorsal fornix. In the following account the “rostral” projections of the hippocampal formation (by way of the fimbria and dorsal fornix) will be described first, and this will be followed by a description of the “caudal” connections of Ammon’s horn and the subicular complex, to the adjoining cortical fields. But before presenting the experimental results some remarks should be made about the topology of the hippocampal formation in the rat since the interpretation of the experiments rests heavily upon an understanding of the topographic relationships of the various cytoarchitectonic fields into which this region can be divided.

51

I. Some features of the topology of the hippocampal formation Although the normal cytoarchitecture of the hippocampal formation in most of the common laboratory mammals has been described in detail on a number of occasions (see for example Cajal, ’11; M. Rose, ’27, ’29; Lorente de N6, ’34; Blackstad, ’56; White, ’59; Hjorth-Simonsen and Jeune, ’72) the three-dimensional configuration of the region remains notoriously difficult to visualize and even more difficult to describe. In an attempt to clarify the topographic relationships of the various fields three-dimensional reconstructions of the stratum pyramidale of fields CA, and CA, of Ammon’s horn, and of the subiculum, were prepared from a series of 20 equallyspaced tracings of frontal sections through the hippocampal formation. The tracings, which had been made at 300-m intervals, were from sections stained by either the Bodian or Nissl methods. For the sake of clarity, the dentate gyrus was omitted from these reconstructions, but its location can readily be envisaged by reference to figure 1, since it surrounds virtually the whole length of the free edge of fields CA, and C&. Fields CA,, CA, and the subiculum were chosen for illustration because, as Rose (’27, ’29) and Lorente de N6 1’34) have pointed out these areas may be considered to consist of a single major cellular lamina (the stratum pyramidale), whereas the adjacent presubiculum, parasubiculum, postsubiculum, and entorhinal cortex consist of at least two well-defined cellular layers (the so-called laminae externa and interna principalis). Figure 1, which is a drawing of one of these reconstructions, shows the disposition of the stratum pyramidale of the right hippocampal formation from its medial aspect. It is evident from this that the subiculum and the two major regions of the hippocampus (the regio superior and the regio inferior, corresponding roughly to fields CA, and CA,,,, respectively) comprise three contiguous strips of cortex. The topographic relationships of the three cortical strips is further shown in figure 2 (a

52


drawing of the lateral aspect of the reconstruction) from which it is apparent that no single account of a lesion or of an isotope injection site, in terms of the “dorsal” or “ventral” hippocampus is really meaningful. For this reason we have been at pains to define as precisely as possible, the location of each isotope injection both with respect to its involvement of specific cytoarchitectonic fields, and also its distribution within each field. Because the hippocampal formation has this complex form it is necessary to indicate, in each case, the rostro-caudal (or more correctly the septotemporal) distribution of injected isotope and, at the same time, its spread along the medio-lateral (or subiculo-dentate) dimension. These two dimensions respectively correspond to the longitudinal and transverse axes of each cortical field. The actual

ACB, Nucleus accumbens AD, Anterodorsal nucleus AH, Anterior hippocampal rudiment ALC, Anterior limbic area AM, Anteromedial nucleus AON, Anterior olfactory nucleus ARC, Arcuate nucleus AV, Anteroventral nucleus BAC, Bed nucleus of the anterior commissure (Gurdjianl CAI-4,Hippocampal fields CAI-4 Cg, Cingulate area CLA, Claustrum CoAp, Cortical amygdaloid nucleus, p. posterior CP, Caudate-putamen DBB, Nucleus of the diagonal band DG, Dentate gyrus DMH, Dorsomedial nucleus ENT, Entorhinal area (28) FC, Fasciola cinerea HM, Medial habenular nucleus HL, Lateral habenular nucleus IC, Inferior colliculus IG, Induseum griseum LD, Lateral dorsal nucleus LGN, Lateral geniculate nucleus LH, Lateral habenular nucleus LHA, Lateral hypothalamic area LM, Lateral mammillary nucleus LS, Lateral septal complex LP, Lateral posterior nucleus

disposition of the various fields is perhaps best illustrated by the two photomicrographs shown in figures 3 and 4. Figure 3 is from a horizontally-cut brain at about the level H in figure 2; while figure 4 is from a frontally-sectioned brain taken at the level marked F in figure 2.

11. The principal subcortical projections of the hippocampal formation A. The dentate gyrus and field C& The results of our autoradiographic experiments with injections involving the dentate gyrus and field CA., of the hippocampus are in essential agreement with the findings of Blackstad et al. (’70) who fist experimentally confirmed the observation of Cajal 1’11)and Lorente de N6 (‘34) that the axons of the dentate granule cells

Abbreviations MD, Mediodorsal nucleus MeAp, Medial amygdaloid nucleus, p. posterior MeM, Medial mammillary nucleus, p. medianus MGB, Medial geniculate body MM, Medial mammillary nucleus MPO, Medial preoptic area MS, Medial septal nucleus OB, Olfactory bulb OT, Olfactory tubercle PARA, Parasubiculum (49) PERI, Perirhinal area (35) PIR, Pirifom cortex PM, Premammillary nucleus PRES, Presubiculum (27) Ps, Postsubiculum (48) Pt, Parataenial nucleus PVT, Periventricular nucleus of the thalamus RF, Rhinal fissure RN, Red nucleus Rs, Retrosplenial area RT, Reticular complex SC, Superior colliculus SCh, Suprachiasmatic nucleus SCO, Subcommissuralorgan SEP, Subependymal plate SF, Septofimbrialnucleus SFO, Subfornical organ SH, Septohippocampal nucleus SN, Substantia nigra SUB, Subiculum

SUM, Supramammillary area TS, Triangular nucleus lT,Taenia tecta VB, Ventrobasal complex VL, Lateral ventricle VMH, Ventromedial nucleus VPL, Ventral posterior nucleus, p. lateralis V3, Third ventricle ZI, Zona incerta ab, Angular bundle ac, Anterior commissure alv, Alveus cc, Corpus callosum cing, Cingulate bundle cp, Cerebral peduncle df, Dorsal fornix dhc, Dorsal hippocampal commissure fi, Fimbria fx, Fornix hf, Hippocampal fissure ic, Internal capsule Id, Lamina dissecans ml, Medial lemniscus mt, Mammillothalamictract och, Optic chiasm ot, Optic tract pc, Posterior commissure pp, Perforant path sm, Stria medullaris st, Stria terminalis vhc, Ventral hippocampal commissure


53

VENT. Fig. 1 A drawing of a reconstruction of the pyramidal cell layer of Ammon’s horn and the subiculum of the right hemisphere of a rat, as seen from the midline. The dentate gyrus is not shown but would lie along the free edge of the regio inferior.Tracings of coronal sections at the four indicated levels through the hippocampal formation are shown in the insets A through D.

do not project outside the hippocampus, stratum moleculare of the dentate gyrus. but terminate in a highly-ordered manner We also have several experiments with inwithin the stratum lucidum of field CA,. A jections involving field C&, but not subdetailed account of the organization of this field CA,, which suggest that the only mossy fiber projection will be presented efferent projection of any length arising elsewhere; here we simply wish to empha- from field (2% is included within this size that the dentate gyrus does not con- bilateral hippocampo-dentate connection; tribute to the extrinsic connections of the thus, like the dentate gyrus, field C& aphippocampal region. We have also con- pears to have no subcortical projection. firmed Zimmer’s (’71) and Gottlieb and Two experiments will serve to illustrate Cowan’s (’73)identification of the hilar re- this point. In experiment R23 the septa1 gion of the dentate gyrus (which includes portion of field C% was heavily labeled subfield CA,, and field CAJ as the source together with some cells in the adjoining of the bilateral projection from the hip- dorsal subiculum; the cells in subfield CA,, pocampus to the inner one-third of the were not labeled. The autoradiographs of

54

L. W. SWANSON AND W. M. COWAN I

DORSAL

POST.

Fig. 2 A lateral view of the reconstructionshown in figure 1; the small inset indicates the location of the hippocampal formation in the hemisphere. Note that to a first approximation the fields CA,, CA,, and the subiculum can be regarded as a series of concentric cortical strips with a longitudinal (or septo-temporal)and a transverse (or subiculo-dentate)dimension.The broken lines H and F mark the approximate location of the photomicrographs shown in figures 3 and 4, respectively.

this brain show a heavy bilateral projection graphically ordered input to the lateral to the inner one-third o f the stratum mo- septal complex o f both sides. We have leculareof the dentate gyrus (as well as the examined six experiments in which the characteristic unilateral projection of the isotope injections appear to be confined dorsal subiculum to the lateral septal com- entirely to field CA,, and one in which the plex, but not the bilateral projection char- injection was limited to the junctional zone acteristic of field CA,; see below). In the “CA,”; in several other brains fields CA, second case (RH 16) field C& was heavily and “CA,” were labeled together with labeled at a more temporal level, together varying amounts of the adjacent cortical with a small number o f cells in the adjoin- fields. The findings in one brain (RH 38) in ing subfield CA,,. In this brain there is a which the injection was strictly confined to very light bilateral projection to the lateral the septal portion of field CA, are shown in septum and again, heavy, bilateral labeling figure 5. Leaving aside the labeled intraof the inner part of the molecular layer of hippocampal and retrohippocampal prothe dentate gyrus.

B. The subcortical projections of fields CA, and “CA, ” 5 All o f our experiments are consistent with the view that the only subcortical projection o f fields CA, and “CA,” is a topo-

s We have used the expression “CA,” because of the continued uncertainty as to the precise limits of this field. In Nissl preparations it is particularly difficult to distinguish the pyramidal cells of this field from those in field CA,, and there are reasons for thinking that field CA, probably has a very limited extent in the rat. “CA,” is thus used in a purely topographic sense, to mark the narrow transition zone between fields CA, and CA, .


Fig. 3 A low-power photomicrograph of a Lux01 fast blue-cresyl violet-stained horizontal section through the temporal region of the rat hippocampus at the level marked by the broken line H in figure 2. Each of the four major divisions of the hippocampal formation (entorhinal cortex, subicular complex, Ammon’s horn, and dentate gyrusl and their various subfields are clearly shown. Scale: 250 m .

55

56


Fig. 4 A low-power photomicrograph of a frontal section through the septal one-third of the rat hippocampus.At this level (which corresponds to the broken line F in fig. 2) only Ammon’s horn and the dentate gyrus are present. Lux01 fast blue-cresyl violet preparation. Scale: 500 pin.

jections illustrated in figure 5, it is evident jection involved this field, are illustrated in that the only rostrally-directed fibers that figure 6. Six features of this relationship are are labeled in this case are confined to the not immediately apparent from the ilposterior part of the septum of both sides, lustration and should be emphasized. (1) the fibers to the contralateral septum The bilateral projection to the lateral sepcrossing the midline in the ventral hippo- tal complex in each case is symmetrically campal commissure. On reaching the sep- located, with the important qualification tum the labeled fibers pass rostrally that the labeling of the ipsilateral terminal through part of the septofimbrial and field is generally slightly more extensive, triangular septal nuclei (where some may and decidedly heavier, than that to the terminate; it is not possible from our mate- contralateral side. (2) It appears that the rial to establish this point) to end in a well- entire lateral septal complex may be in circumscribed region of the dorsal and receipt of fibers from field CA,. (3) While intermediate divisions of the lateral septal each injection resulted in the labeling of a complex. In none of our experiments does discrete terminal area within the lateral the input to the lateral septal complex from septum, the labeled areas appear to overfield CA, extend as far rostrally as the an- lap to some extent; thus when adjacent terior hippocampal rudiment. parts of field CA, were injected (as in exIn the other experiments with injections periments R H 24 and RH 38) the distribuconfined to field CA, the bilateral input to tion of silver grains in the septum of the the septum was similar, but interestingly, two sides showed a significant area of overdifferent parts of field CA, have been lap (fig. 6: G, H, and I). (4) More septally found to project to different areas within placed injections in field CA, result in lathe lateral septal complex. The topo- beling in more dorsal portions of the lateral graphic relationships between field CA, septal complex, while temporally located and the posterior and lateral septal nuclei, injections give rise to terminal labeling in which have emerged from the analysis of the ventral part of the septum (compare, all our experiments in which the isotope in- for example, brains T10 vs RH 15 vs RH 23:


57

58


W

59


PROJECTIONS

\

TERMINAL AREAS

INJECTIONS

IN CA2

Fig. 7 A schematic representation of the location of the injection sites in four experiments involving “CA,” (drawings C and D), their projections through the fimbria (drawing A), and the patterns of terminal labeling to which they give rise in the septum (drawing B). As far as can be determined, the organization of these connections is similar to that seen following injections of field CA, [see fig. 6).

fig. 6). (5) There is a similar medial to lateral arrangement in the projection from the septal part of CA, where this field has a distinct transverse dimension in frontal sections (fig. 6: G, H, and I). (6) Each level within field CA, projects through a restricted segment of the fimbria (fig. 6C). Thus at the level of the rostra1 pole of AmFig. 6 A schematic representation of the injection sites involving field CA, in each of several experiments (D, E, F, H, I) and their projections through the fimbria (drawing C) to terminal fields in the septum (drawings A, B and GI. For each experiment the symbol used hatching, stippling, etc.) is carried through from injection site to terminal field. Note that the efferent fibers in the h b r i a are organized from medial to lateral depending upon whether the injection site in field CAI involved its septal or temporal part, and that injections in the septal part of the field (as shown in D, Hand I) result in the transport of label to the dorsal part of the lateral septal complex of both sides, while injections in the temporal part of the field result in bilateral labeling in the ventral part of the lateral septum.

mon’s horn it is evident that the fibers which arise from temporal parts of field CA, are confined to the lateral edge of the fimbria (RH 23: fig, 6C) whereas those arising from the septal pole of the field lie immediately adjacent to the stratum oriens (RH 5: fig. 6C);intermediate regions within CA, project through the mid-portion of the h b r i a (experiment T 10: fig. 6C). As far as we can determine from our material, the subcortical projection of “CA,” is identical to that of field CA, in its general organization. The only experiment we have in which the isotope injection appears to be confined to this field is R 105 (fig. 7). The other injections illustrated in figure 7 involved, in addition to “,A2,” varying numbers of pyramidal cells in the adjacent field CA,. The incidental involvement of field CA, does not present a serious problem since the projection of this field upon the septum is strictly unilateral (and to the

60


Fig. 8 A low-power photomicrograph of the injection site in experiment R 102. Note that the region of the injected isotope (marked by arrows) includes the midportion of field CA, at this level and the contiguous portion of the dorsal blade of the dentate gyrus. The asterisk marks the location of the needle tract immediately dorsal to the hippocampus; the double arrowheads mark the junction between fields CA, and CAI. Scale: 500 pm; thionin preparation.

same side). Thus the finding of bilaterally distributed label within the septal complex can safely be attributed to the labeling of pyramidal cells in field “CA,.” As may be seen from figure 7, there is a septo-temporal relationship between “CA,” and the lateral septal complex, and an organization of the efferent fibers within the fimbria, comparable to that described above for field CA,. Thus the rostrally-placed injection in experiment R 104 resulted in the labeling of fibers in the medial part of the fimbria, and these fibers appeared to terminate bilaterally in the dorsal part of the septum; conversely, the more temporally-located injections in experiments R 104 and R 105, labeled fibers in the middle third of the fimbria and in more ventral sectors of the two lateral septal complexes. In the above account of the septal projections of fields CA, and “CA2,”we have made no mention of the medial septal nucleus. The reason for this is that in all the experiments available to us with injections involving these two fields only two showed

any indication of silver grains over the region containing the large neurons of the medial septal and diagonal band nuclei. In both these cases the labeling over the medial septum was extremely light and nothing comparable to it was found in the other experiments with similar injections. We can only conclude from this that if fields CA, and “,A2” do in fact project directly to the medial septal-diagonal band complex in the rat, this connection must be sparse and at the limits of detection by the autoradiographic method. Of course, this does not rule out the possibility that fibers from CA, and “CA,” may establish axodendritic contacts with neurons in the medial septal or diagonal band nuclei since, as our Golgi-impregnated material indicates, the dendrites of the cells in these two nuclei project for considerable distances into the lateral septal complex. C. The subcortical projection of Field CA, As in the case of fields CA, and “CA,” the only descending projection we have


61

62


observed from field CAI is to the lateral septal complex, with the possible exception of that part of the field which adjoins the ventral subiculum, which may contribute, with the latter to the hippocampal projection to the nucleus accumbens, the taenia tecta, the infralimbic cortex (of Rose and Woolsey, ’481, and the medial and posterior parts of the anterior olfactory nucleus. As this possibility is discussed below, with the account of the connections of the subiculum, we shall limit ourselves here to a consideration of the septal projection of field CAI. In a series of eight experiments with injections confined to field CAI, and in several additional brains with heavily labeled neurons in this and the adjacent hippocampal fields, we have found a distinct, topographically-organized projection from field CAI to the lateral septal complex. A typical experiment from the latter group is experiment R 102. Figure 8 is a photomicrograph of the injection site in this brain. The focus of the injection is within the septal one-third of the hippocampus and here the injected isotope has labeled the pyramidal cells of field CAI over a limited mediolateral extent. There has been some spread of label across the hippocampal fissure into the dorsal blade of the dentate gyrus, but fields “CA,,” CA, and (2%have not been encroached upon. The distribution of the transported label in this brain is shown in figure 9. It is evident from this that field CAI at this level projects rostrally through both the dorsal fornix and the medial part of the fimbria. The projection to the septum is strictly limited to the side of the injection, and involves principally the dorsal part of the septofimbrial and triangular septal nuclei, and the dorsomedial part of the lateral septal complex. Whether in fact CA, fibers terminate upon cells in the septofimbrial or triangular septal nuclei cannot be determined from this kind of material, but the localized projection within the lateral septal complex is clearly a terminal one. From this, and the other experiments involving field CA, ,it is evident that the pro-

jection of this field upon the septum resembles that of fields CA, and “CA,” (with the notable exception that it is strictly unilateral). There is also a similar ordering of the fibers within the fimbria (fig. 10D) and a similar topographic arrangement of the terminal projection within the lateral septal complex. Thus, a medial to lateral series of injections in the septal part of field CAI in different experiments, results in labeling within a series of overlapping mediolateral zones within the lateral septum (c.f. figs. 10B and F).A comparable septo-temporal organization is also evident. For example, the temporally placed injection in experiment R 104 (fig. 10E) gave labeling in the ventral part of the lateral septal complex close to the medial septal nucleus (fig.10B) whereas the more septally placed injection in R 101 labeled only the dorsal part of the lateral septal complex. It is important to add that the septal projection from field CAI differs in two further respects from that of fields CA, and “CA,.” (1) As we have pointed out in the account of experiment R 102, the septal portion of CAI projects by way of the dorsal fornix as well as through the fimbria, whereas fields CA, and “CA,” project only through the fimbria; and (2) the projection from field CAI tends to be distributed further rostrally within the lateral septal complex, than that from fields CA, and “CA,.” As in the case of field CA, there appears to be a very limited projection from CAI to the medial septal/diagonal band complex. In only three cases with injections involving CAI were silver grains seen in this complex, and in several others, with similar injections, there was no evidence of such a projection. Again we conclude from this that if field CAI projects directly to the medial Fig. 10 A schematic representation of the injection sites in a number of experiments involving field CA, (drawings E and F) and their projections through the dorsal fornix and the fimbria (drawing D),to their terminal fields in the septum (drawings A, B and C). Note the topographic ordering of the fibers in the dorsal fornix and the fimbria from medial to lateral, and the organization of the terminations within the ipsilateral septal complex in the dorsoventral and mediolateral dimensions.

63


a

64


Fig. 11 A series of drawings to show the site of a 3H-proline injection in the dorsal part of the subiculum, and the distribution of the transported label seen in experiment R 176.

septal and diagonal band nuclei, the projection must be either very sparse, or largely to the peripheral dendrites of the large neurons in these nuclei.

D. The subcortical projections of the subiculum All parts of the subiculum6 have been found to project ipsilaterally to the lateral septal complex. The most dorsal part of this field (which has, on occasion, been regarded as a caudal extension of Ammon’s horn) almost certainly gives rise to a significant projection to the mammillary complex, by way of the fimbria and postcommissural fornix, while its ventral-most part appears to give rise to the medial corticohypothalamic tract which is known to terminate in relation to the ventromedial nucleus of the hypothalamus. For convenience we shall refer to these two Darts as the ‘‘dorsal” and ‘‘Ventral SUbicUlUW’’ respectively.

(i) The dorsal subiculum Two experiments with injections involving the dorsal subiculum are particularly useful. In experiment R 176 heavily labeled cells were strictly confined to the dorsal subiculum (fig. 11). In experiment R 2 3 the adjoining part of the pre- and parasubiculum were also labeled to some extent, as were the pyramidal cells in the subjacent field (2%. Since as we have seen, field CA, does not project into either the pre- or postcommissural fornix, its incidental involvement in the latter case does not complicate the interpretation of the findings. The difficulty presented by the involvement of the pre- and parasubiculum will be considered later. As figure 11 shows the dorsal subiculum projects to: (i) the lateral septal complex (fig. 12)in much the same way as the adjacent part of field CA, ; 6 We have followed Blackstad (‘56)in not recognizing a separate prosubiculum k.f. Lorente de N6, ’34).


65

Fig. 12 Three photomicrographs to show the distribution of transported label in the septum following an injection of 3H-proline into the dorsal part of the subiculum (experiment R 176).Figure (A) is a low-power, bright field photomicrograph of a thionin-stained frontal section through the mid-portion of the septum. The appearance of the transported label in the medial part of the ipsilateral lateral septal nucleus is shown in the higher-power dark field photomicrograph of the corresponding autoradiogram (B); for comparison, the background level of radioactivity in these autoradiographs is shown in (C) which is taken from the corresponding part of the contralateral lateral septal nucleus. Scale: 250 pm.

and (ii) through the descending column of the fornix to the mammillary complex where its efferents end bilaterally in the pars medianus of the medial mammillary nucleus, and in the dorsal halves of the

pars medialis and the pars posterior. No projection could be followed caudal to the mammillary complex. The findings in experiment R 23 were similar in most respects, but in addition, there is evidence of

66


Fig. 13 Drawings of two experiments with similarly placed injections which involve the ventral part of the subiculum (experiments R 135 and F EI 9). Although in neither case was the injection confined to the ventral subiculum, in both brains the projection into the pre- and postcommissural fornix was quite distinctive, as shown in tracings A through G . (Brain R 135 was sectioned horizontally, but for the sake of clarity the projections have been plotted as they might appear in the frontally sectioned brain in RH 91. Injections which labeled fields CA,, CA,, or the pre- and parasubiculum, but not the ventral subiculum, did not show these projections.

a massive input to the anterior thalamic nuclei in this brain. This thalamic projection is probably attributable to the involvemeht of the pre- and/or parasubiculum by the isotope injection as we shall discuss below. Since we have only a small number of idjections which involve the dorsal subiculum it has not been possible to determine the total extent of the subicular input to the mammillary complex; however, in no instance was transported label observed over the lateral or supramammillary nuclei. (ii) The ventral subiculum

Our autoradiographic experiments have

enabled us to confirm, with some precision, the earlier suggestion of Raisman et al. ('66) that the medial cortico-hypothalamic tract has its origin within the ventral part of the subiculum. In every experiment in which this tract was labeled there was some involvement of the ventral part of the subiculum by the injection. Although we have no injection wholly confined to the ventral subiculum, two experiments (R 135 and RH 9) in which it was significantly labeled are depicted in figure 13. In experiment R 135 the pre- and parasubiculum were also labeled by the injection and in RH 9 parts of fields CA, and (2% were im-


67

Fig. 14 “ A ’ is a bright field photomicrograph of a frontal section through the olfactory peduncle in experiment FUI 9, with an injection involving the ventral part of the subiculum (see fg.13).The distribution of transported label in the medial part of the ipsilateral anterior olfactory nucleus is shown in the higher-power dark field photomicrograph in “B”. The background level of radioactivity in the anterior olfactory nucleus of the opposite side is shown in “ C ’ . Scale: 1 mm; thionin preparations.

plicated. In both experiments the fibers of the medial cortico-hypothalamic tract were clearly labeled as they run in the lateral tip of the h b r i a , and could be followed from there into the postcommis-

sural fornix to their termination in the cellfree, capsular zone surrounding the ventromedial nucleus of the hypothalamus. Since we have never observed labeling of the medial cortico-hypothalamic tract after in-

68


jections involving fields CA, and C&, or the pre- and parasubiculum, but sparing the ventral subiculum, we feel reasonably confident in attributing the heavy labeling of this tract in R 135 and RH 9 to the labeling of the cells in the ventral subiculum. The precommissural projections of the ventral subiculum were clearly labeled in nine experiments with injection sites involving this area. As the crucial findings in all these experiments are essentially similar, R 135 and RH 9 can again be used to illustrate the relevant points. The rostra1 projections pass through the septofimbrial nucleus, and it seems probable from the pattern of grain distribution in this nucleus and in the adjoining lateral septal complex, that many fibers terminate in this region. Other fibers continue into the bed nucleus of the stria terminalis, and further ventrally into the medial parts of the nucleus accumbens, and to the deeper layers of the infralimbic cortex, to the posterior and medial parts of the anterior olfactory nucleus (fig. 141, and to the molecular layer of the taenia tecta. In five of the nine injections which gave rise to this pattern of labeling in the precommissural fornix the medial cortico-hypothalamic tract was not labeled. Whether this is because separate groups of neurons in the ventral subiculum give rise to these two pathways, or because different proportions of the adjoining cortical fields were labeled, is not known. In at least some cases, the latter explanation seems likely. Thus in four experiments (R 104, R 106, R 107, and R 178) the projection to the anterior olfactory nucleus and adjacent areas was labeled by injections which appeared to include the temporal part of field CA,, but not the ventral subiculum. In a number of experiments the injections involved a region of the subiculum which, on topographic grounds, could not be described as either “dorsal” or “ventral” subiculum,but rather an intermediate zone. These cases, when considered together with the most dorsal and ventral subicular injections, serve to establish that within the precommissural projection of

the subiculum to the lateral septal complex, there is a distinct dorsal-to-ventral organization. By this we mean that injections involving the dorsal subiculum always give rise to labeling over the dorsal part of the lateral septum, while intermediate and ventrally placed injections respectively labeled intermediate and ventral parts of the lateral septal nucleus.

E. The subcortical projections of the pre- and parasubiculum Our material on the connections of these two cortical fields is rather limited and, unfortunately, we do not have injections strictly confined to either the pre- or the parasubiculum. Therefore, in what follows we shall make no distinction between the two fields with respect to their subcortical efferent projections. However, from one experiment in which the isotope injection was confined entirely to the area of the pre- and parasubiculum (R 133) and six additional experiments in which these fields were involved together with one or more adjacent cortical areas, we have been able to identify two major subcortical projections arising in this region. As shown in figure 15, the injection in experiment R 133 was confined entirely to the ventral portions of the pre- and parasubiculum, heavily labeling neurons in all layers of both fields. Following this, transported label could be followed to: (i) the mammillary nuclei, and (ii)the anterior thalamic and the associated lateral dorsal nucleus, and to the lateral posterior nucleus of the thalamus. The thalamic projections were by way of fibers which run both in the fornix and in the posterior thalamic radiation. The fibers in the fornix appeared to terminate principally in the anterior thalamic nuclei, while those in the posterior thalamic radiation seemed to end for the most part in the lateral dorsal and lateral posterior nuclei (fig. 15). Taken in conjunction with the findings in the other brains with injections involving the pre- and parasubiculum the following generalizations about the projections of this cortical region may be made.


69

Fig. 15 To show the site of the injection and the distribution of transported label to the anterior thalamic and rnammillary nuclei, and to the contralateral parasubiculuin and entorhinal area, in experiment R 133 with an injection involving the pre- and parasubiculum.

First, of the anterior thalamic nuclei the anteroventral nucleus receives by far the heaviest input, and this is bilateral; significantly fewer fibers appear to reach the ipsilateral anteromedial and anterodorsal nuclei. Second, different injections resulted in somewhat different patterns of transported label in the lateral dorsal and lateral posterior nuclei. Although this suggests that there may be a topographical

relationship between the pre- and parasubiculum and the lateral thalamus, we could not establish the details of this pattern in the limited material available to us. However, it is clear that injections in both the dorsal part (as in experiment R 2 3 described above) and in the ventral half of these cortical fields (e.g., R 133) gave rise to these thalamic projections. Third, the appearance of labeled fibers in the lateral

Fig. 16 To show the site of the injection in the postsubiculurn,and the distribution of the transported label, in experiment R 162.


posterior nucleus does not appear to be due to incidental involvement of the visual cortex since the lateral geniculate nucleus and the superior colliculus were unlabeled in all these experiments. Finally, unlike the subiculum, the pre- and/or parasubicular input to the mammillary complex terminates bilaterally in both the medial and lateral mammillary nuclei. Many more silver grains are seen over the ipsilateral lateral mammillary nucleus than over the contralateral nucleus; the pars medianus of the medial mammillary nucleus does not seem to receive an input from this source. Shipley (’75) has recently described the projection of the presubiculum in the guinea pig in some detail, but did not report evidence for a thalamic or mammillary component. From this it might be inferred that the projections we have described arise exclusively in the parasubiculum, but we have no direct evidence for this, nor can we rule out that the organization of these connections might be quite different in the rat and guinea pig. F. The subcortical projections of the ‘postsubicular’’ cortex 7 The subcortical projections of the postsubiculum, a narrow cortical field adjacent to the parasubiculum (see Rose and Woolsey, ’481, were studied in two experiments with injections confined to this area. The projections seen in one of these experiments (R 162) are illustrated in figure 16. The primary descending connections of this field are by way of the posterior thalamic radiation to a terminal field within a restricted region of the lateral dorsal and lateral posterior thalamic nuclei. A few fibers could be traced rostrally in the ipsilateral dorsal fornix into the ipsilateral fornix column, but could not be followed much beyond the level of the anterior commissure. Essentially the same observations were made in experiment R161 in which, interestingly, only the cells in the deepest third of the postsubiculum were labeled by the injection.

G . The entorhinal cortex We have seven injections confined to the

71

entorhinal cortex, but since this field appears to project exclusively to Ammon’s horn and the dentate gyrus (via the alveus and the perforant path) as previously reported by Raisman et al. (’65)we shall not comment further upon these experiments here. 111. Cortico-cortical connections within the hippocampal formation In this section we shall consider some of the more prominent cortico-cortical interconnections between the hippocampus and the adjacent cortical fields, which have emerged from the analysis of our autoradiographic material. We shall not discuss either the commissural connections of the hippocampus and dentate gyrus (as these have been described previously - Gottlieb and Cowan, ’73) or the intrahippocampal association pathways, which will be the subject of a separate study.

A. The dentate gyrus and field CA As we have pointed out above, no extrahippocampal projections have been found either from the dentate gyrus or from field CA, of Ammon’s horn. Their efferent connections are entirely intrahippocampal .

B. The cortico-cortical connections of field CA, and “CA,” All injections involving fields CA, and “CA,” give rise to transported label within the subiculum. An example of this is illustrated in figure 5 (experimentRH 38). Following the labeling of pyramidal cells at the septa1 pole of field CA, and “CA,” large numbers of silver grains can be seen over the deep half of the stratum moleculare of the subiculum and to a lesser extent scattered among the pyramidal cell somata themselves (fig. 17). The projection is es7 The term “postsubiculum”has been adopted from Rose and Woolsey (‘48). The region in questioncorresponds to area 48 Ithe retrosubicular area) of Brodmann (‘09).As Blackstad has pointed out, histochemically it is difficult to distinguish this field from the presubiculum, and for this reason he, and his colleagues, have regarded this area simply as the “dorsal part of the presubiculum” (personal communication, ’76). However, we believe that on cytoarchitectonic grounds and, as we shall see, on the basis of their connections, the two are distinguishable, and for these reasons we have followed Rose and Woolsey.

72


Fig. 17 Bright field (A) and dark field (B) photomicrographs to show the distribution of transported label to the subiculum following an injection of field CA, in experiment RH 24 (fig. 6). Note that most of the silver grains lie over the inner half of the molecular layer of the subiculum, but a few are also found over the pyramidal cell layer. Scale: 500 pm; thionin preparation; frontal sections.

sentially an uncrossed one, although there may be some transported label over the contralateral subiculum at its interface with field CA, (which receives a distinct crossed projection from field CA,). The projection of field CA, to the subiculum shows a topographic ordering such that after injections of the septa1 part of CA, the more dorsal parts of the subiculum are labeled, whereas after injections of the temporal portion of CA, the transported

label is confined to more ventral parts of the subiculum. In addition to its projection to the subiculum, field CA, appears also to give rise to a more limited projection to the presubiculum (fig. SE), to the parasubiculum,to the deeper layers (especially layer IV) of the entorhinal cortex, and to the cingulate area (of Rose and Woolsey, '48). The precise organization of these various projections is difficult to determine, and


73

area it is found primarily over layer I. These observations have been confirmed in the second experiment, RH5, with an injection involving the septal part of field CA,, just caudal to that in RH 38. On the other hand, after injections involving field CA, at about the middle of its septo-temporal extent (experiments R 104 and R 141: fig. 18)there is clear evidence for a projection to a localized part of the parasubiculum (ending in layers I and 11) and an equally marked projection to layer IV of the entorhinal cortex (fig. 191. Of 15 injections involving various parts of field CA,, in only these two was the projection to the entorhinal cortex clearly labeled. HjorthSimonsen (’71) has described a projection from the “temporal third of CA,” to layer IV of the medial entorhinal area, which in most respects resembles the findings in R 104 and R 141 of our series; however, in our material this hippocampo-entorhinal projection appears to be distributed to both the medial and lateral parts of the entorhinal cortex, and not simply to the former.

C. Cortico-cortical connections of field CAI Field CA, appears on connectional grounds to be divisible into septal and temporal subfields, the former projecting through (and perhaps to) the subiculum, Fig. 18 A low-power bright field photomicro- the parasubiculum and the deeper layers graph to show the site of the injection (marked by arrows) in field CA, and the adjoining part of the den- of the perirhinal area of Rose (’29) and tate gyrus in experiment R 141. Scale:500pm; thionin Krieg (’461,while the latter projects only to preparation,horizontal section. the subiculum. We have examined eight experiments the evidence we have at present bearing with injections involving the septal portion on this issue is far from complete. How- of field CAI.Figure 9 is taken from one of ever, some indication of the possible re- these cases (R 102).The caudally directed gional differences in the projections to fibers which are labeled after this injection these areas can be given by reference to a pass through, and possibly terminate withfew individual experiments. As we have in, the subiculum and the presubiculum en seen in experiment RH 38, an injection in route to the perirhinal area. The terminal the septal portion of field CA, and possibly labeling of this cortical field is restricted to “CA,” results in heavy labeling of the its deepest cellular layer (fig. 9) in which presubiculum and a limited part of the the cells are very closely packed together. cingulate area (fig. 5). In the presubiculum We have several experiments with injecthe transported label is heaviest over tions involving only the temporal half of layers I and 111, whereas in the cingulate field CAI (e.g., experiments R 171 and

74


Fig. 19 Tracings to show the site of the injection and the resulting transported label in the adjacent cortical fields in experiment R 141. Note the laminar distribution of the labeling over the subiculum, parasubiculum, and entorhinal cortex in this brain. The tracings are from horizontal sections, and are arranged from dorsal (A) to ventral (D);the number of the serial sections is shown to the lower left of each tracing.

HIPPOCAMPAL EFFERENI'S

R 178). In these cases a caudally-directed projection to a limited part of the SUbiculum is clearly labeled. Unlike the findings in the cases with more septallyplaced injections, the subicular projection from the temporal part of CA, does not continue beyond the subiculum. Most of the labeling over the subiculum is found over the deeper part of the stratum moleculare, but some is seen over the pyramidal cells, and there is a good deal between these cells and the subjacent white matter. D. The cortico-cortical connections of the subiculum The only cortical area caudal to the subiculum which appears to receive a projection from this part of the hippocampal formation in the rat is the perirhinal area, although, because of the possible complication due to the diffusion of label, we cannot rule out shorter connections to the adjoining parts of the pre- and/or parasubiculum. The projection to the perirhinal area was only seen after injections into the dorsal part of the subiculum, and again it appeared to terminate within the deeper layers of the perirhinal cortex (fig. 11).

E. The cortico-cortical connections of the pre- and parasubiculum

75

biculum cross in the ventral hippocampal commissure, and then course back through the alveus, to the angular bundle. These projections from the pre- and parasubiculum have been confirmed in each of three additional experiments, but unfortunately in no case was the injection confined to either field individually.

F. Cortico-cortical connections of the postsubiculum The efferent connections of the postsubiculum are illustrated in figure 16 which is based on findings in experiment R 162 After this injection there is distinct labeling of a projection to Rose and Woolsey's ('48) retrosplenial and cingulate areas (where it is distributed mainly to the molecular layer). In addition, the postsubiculum projects bilaterally to the entorhinal cortex and to the parasubiculum; the inputs to the entorhinal and parasubicular cortices again appear to be distributed predominantly to layer I. Whether this region has additional projections, and if so, whether they are topographically organized has not been determined, since we have only two other experiments with similar injections.

G. Cortico-cortical connections of the entorhinal area The entorhinal cortex has been found to project only to Ammon's horn and to the dentate gyrus as described above. However, it is worth emphasizing that there is a well-defined crossed tempero-ammonic tract which arises within the entorhinal area, crosses in the dorsal hippocampal commissure, as described by Blackstad ('56) and Goldowitz et al. ('75) and terminates within the contralateral field CA, and the outer half or two-thirds of the molecular layer of the dentate gyrus of that side.

Following an injection of 3H-proline confined to the pre- and parasubiculum (R 133: fig. 151, in addition to the thalamic and hypothalamic projections already described (p. 68), there is a light but distinct projection to the ipsilateral entorhinal cortex, and a crossed projection to the entorhinal area and parasubiculum of the opposite side. Since both the pre- and parasubiculum were involved in the injection, it is not clear which of these two fields gives rise to the various cortical projections and conceivably both may do so. In addition to the crossed projection to DISCUSSION the parasubiculum there is extensive labelIt is convenient to begin the discussion of ing throughout the parasubiculum of the same side which is especially obvious in the our observations by reference to table 1 in molecular layer. The fibers to the con- which the major subcortical and cortical tralateral entorhinal cortex and parasu- projections of each of the principal cortical

76


TABLE 1

Summary of hippocampal efferent projections shown autoradiographically in the rat Subcortical ~

Cortico-cortical

~~

1 Ammon’s horn and the dentate gyrus a Dentate gyrus: none b C&: probably none c CA, (+“CA,”): lateral septal complex (bilaterally)

d CA, : lateral septal complex 2 Subicular complex a Subiculum: lateral septal complex (i) Subiculum, dorsal part: also to mammillary complex (ii) Subiculum, ventral part: also to ventromedial nucleus; bed nucleus of stria terminalis, n. accumbens, anterior olfactory nucleus, taenia tecta, infralimbic area h Pre-/parasubiculum: anterior and lateral thalamic nuclei, mammillary complex

c. Postsubiculum: lateral thalamic nuclei

3 Entorhinal cortex a Ehtorhinal area: none

fields of the hippocampal formation have been tabulated, and to figure 20 which is an attempt to summarize diagrammatically the major findings of this study. Before considering the projections of individual fields in detail it is perhaps worthwhile pointing out that virtually all of the connections which we have been able to demonstrate autoradiographically have, at one time or another, been reported in either normal material or in tissue stained by either the original Nauta-Gygax method or one of its more recent variants, such as the Fink-Heimer technique (for detailed reviews of this literature see especially Powell and Cowan, ’55; Nauta, ’56, ’58; Guillery, ’56; Blackstad, ’56; Raisman et al., ’66; Hjorth-Simonsen, ’71; ’73). However, it is clear that many of the earlier accounts of the origin, and in some cases the course, of these efferent hippocampal pathways have been misinterpreted. In large measure this has been due to the in-

a Dentate gyrus: CA,, CA, b C&: dentate gyrus (bilateral) c CA, (+“CA,”): subiculum (i) CA,, septal part: also to cingulate area, presubiculum (ii) CA,, intermediate part: also to parasubiculum, entorhinal area d CA, : subiculum 6) CA,, septal part: also to perirhinal area a Subiculum, dorsal part: perirhinal area

b Pre-/parasubiculum: entorhinal area (bilaterally), parasubiculum (bilaterally) c Postsubiculum: cingulate and retrosplenial areas, entorhinal area (bilaterally), parasubiculum (bilaterally) a Ehtorhinal area: dentate gyrus (bilaterally), Ammon’s horn (bilaterallv)

evitable interruption of “fibers of passage” by even the smallest lesions. This is particularly true of lesions involving fields such as the subiculum, or such fiber bundles as the alveus, through which many fibers of diverse origin pass. There is perhaps no more striking example of this than the repeated descriptions of axonal and terminal degeneration in the postcommissural fornix after lesions of Ammon’s horn, the dorsal fornix, or the fimbria, whereas it is now apparent that in the rat, at least, the fibers of the post-commissural fornix arise principally, if not exclusively, from the retrohippocampal region including the subiculum and the pre- and/or parasubiculum (Swanson and Cowan, ’75). It is for this reason that a re-examination of the entire efferent projection of the hippocampal for mation using the autoradiographic method seemed justified. As has been frequently pointed out one of the major advantages of this method, in addition to its increased

Fig. 20 Three schematic drawings which summarize the major efferent connections of the hippocampal region as determined in this study. Note that Ammon’s horn (drawing A1 projects rostrally, only through the precommissural fornix to the septum, that the subiculum (drawing B) projects into both the pre- and postcommissurd fornix, while the pre-/parasubiculum (drawing C) projects solely through the postcommissural fornix. The efferent connections of the dentate gyrus and the entorhinal area which are confined to the hippocampal region are not shown.

4 4

78


sensitivity, is that it is not complicated by the fiber of passage problem (Cowan and Cuenod, ’75). Perhaps the major generalization which can be drawn from this study is that each cortical area within the hippocampal formation is characterized by a distinctive set of efferent connections. And from what we know of their afferent connections it is evident that the distinctive cytoarchitecture of this region is paralleled by an equally distinctive constellation of connections. It is important to point out, however, that the projections that we and others have described represent only a first step towards a complete understanding of the wiring of these various fields. In a few cases it has been possible to make reasonable assumptions about the cells which give rise to the various efferent connections. To cite an obvious example, no one would question that the efferents from field CA, arise from cells whose somata lie within the stratum pyramidale. But in other cases -especially in fields with more complex cellular layering - any one of several classes of cells might contribute to the long projections we have studied. In general the autoradiographic method, as we have used it, is not well suited for determining the precise cells of origin of fiber systems. For this, methods based on retrograde axonal transport are more useful, and now that the terminal projection sites of the various fields are known in broad outline, it should be possible to address the problem of their precise origins with the horseradish peroxidase technique (LaVail and LaVail, ’72; Lavail, ’75). This problem is particularly acute for those fields which give rise to two or more efferent projections. At present we cannot say whether these arise from different subpopulations of cells or from collateral branches of neurons of the same type. We shall return to this problem later when we consider the caudally directed outflow from Ammon’s horn but here we may note that except in those instances in which the different subpopulations of cells are topographically segregated (for instance in different cell layers) a new

methodology will be needed which will permit the labeling of terminals in two different areas simultaneously by distinctive markers. Similar difficulties arise with respect to the precise termination of the efferent pathways we have described. In some few instances the pattern of termination of fibers as seen in the autoradiographs is sufficiently distinctive to permit a reasonable inference as to the identity of the postsynaptic targets. Thus, even if there were not collateral evidence from Golgi and electron microscopical material, it could hardly be doubted that the very dense labeling seen in the stratum Zucidum of field CA, after an injection involving the dentate gyrus is indicative of the labeling of mossy fiber terminals which synapse upon the large spines on the proximal portions of the dendrites of the pyramidal cells. But in most cases the distribution of silver grains within a projection field lacks this distinctive character, and it is therefore most difficult to determine whether the fibers end upon cell somata or dendrites. In addition, it is usually not possible to determine upon which class of cell in the projection field they synapse. Until information about this important facet of the connectivity of the hippocampus is forthcoming, our understanding of the organization of its connections must remain at a rather elementary level. Moreover, without a knowledge of the excitatory or inhibitory nature of the different pathways, and whether they terminate upon cells which themselves are excitatory or inhibitory (as well as where upon these cells the synapses are distributed), all attempts to model what the different fields of the hippocampal formation may do, are bound to be frustrated.

I. The efferent connections of Ammon’s horn and the dentate gyrus The principal conclusion to be drawn from our findings on the rostrally directed efferent projections of the various subfields of Ammon’s horn and the dentate gyrus is that they have a form of hierarchical


organization. Thus, at the lowest or simplest level, the dentate gyrus has no extrahippocampal projection, and indeed its outflow is confined entirely to the largecelled regio inferior of the hippocampus of the same side. At the next level, field C& again appears to have no extrahippocampal projection, but it does give rise to an undoubtedly important associational system of connections within the hippocampus of both sides. The efferents from field CAI occupy a somewhat higher level in this scheme, in that this field has a substantial extrahippocampal projection although this appears to be confined to the lateral septa1 complex of the same side. Moreover field CA, appears to be the most limited in terms of its intrahippocampal projection since it does not give rise to either associational projections to other fields of Ammon’s horn or to the dentate gyrus, nor does it project to the hippocampal formation of the opposite side. Finally, in this scheme, field CA, presents the most complex pattern of efferent connections: not only does it give rise to extensive intrahippocampal association and commissural projections (Gottlieb and Cowan, ’731, but its extrahippocampal projection is distributed bilaterally to the septum. A fuller understanding of the significance of these differences must await further information about the organization of the afferents to these various fields, and a more complete understanding of their associational connections. It is also evident that this hierarchical pattern has to be fitted into the context of what is already known of the sequential processing of information within the hippocampal formation as a whole. Here the dentate gyrus occupies a pivotal position in the processing sequence since it is the recipient of the major extrinsic input to the hippocampus from the entorhinal cortex (Hjorth-Simonsen, ’72a,b) and it projects upon fields C& and CA, by way of the mossy fiber system (Blackstad et al., ’70).Field C& feeds back upon the dentate gyrus of both sides while field CA, not only provides a major hippocampal outflow to the septum but also pro-

79

jects massively to field CAI through its Schaffer collaterals, as well as to fields CA, and CA, of the contralateral side. From this point of view field CA, appears to be the principal focus of integration of all intr ahippocampal activity. That the hippocampus proper contributes only to the precommissural fornix, and within the precommissural component only to the septum, is one of the most unexpected findings in our study. There appear to be approximately 300,000 cells in the pyramidal layer of field CA,, and about 150,000 such cells in field CA, (including “CA,”) - (Banker, Schlessinger, and Cowan, unpublished observations). Assuming that the number of fibers projecting into the septum in the rat represents a similar proportion of the number of fibers in the fimbria and dorsal fornix as it does in other mammals in which this has been directly established (Powell et al., ’57) it would seem likely that at most between 50,000 and 100,000 fibers enter the septum. If this calculation is even approximately correct, and if each of these precommissural fibers arose from a single pyramidal cell, it would appear merely on numerical grounds that the rostrally directed projection of Ammon’s horn comprises only a relatively minor part of its total outflow; and since the subiculum also contributes to the precommissural fornix the actual figure for the number of hippocampal pyramidal cells contributing to the precommissural fornix may be even smaller. It follows from this that a substantial proportion of the cells in Ammon’s horn must send their axons elsewhere. Since some of the fibers in the alveus and fimbria were described by Cajal (’111 and Lorente de N6 (’34)as giving off collateral branches it has generally been assumed that most of the commissural and associational connections of the hippocampus were formed by collaterals of axons which project rostrally into the fornix. We have no direct evidence bearing on this point, but in studies of the commissural connections of Ammon’s horn and the dentate gyrus using the retrograde transport of

80


horseradish peroxidase, we have been impressed b y the constant finding that only an occasional cell in fields CA, and C& can be labeled from the contralateral side. It is possible of course that this simply reflects the fact that the concentration of the injected marker was such that the terminals of only a small number of commissural fibers were adequately labeled. But on the whole this seems unlikely, since it might be supposed that if only a small number of terminals were labeled, one would find a restricted focus on the contralateral side in which the cells were labeled, rather than a small number of cells scattered over a considerable area in the contralateral hippocampus. The alternate possibility is that the population of pyramidal cells is a heterogeneous one with, for example, some cells sending their axons to the hippocampus of the opposite side, others giving rise to association fibers, and still others projecting to the septum. If this is the case we have no means at present of estimating the relative numbers of cells in each of these different categories, nor can we determine how many cells contribute to the caudally directed projections to the subiculum, parasubiculum, entorhinal cortex, or cingulate area. By any measure the caudally directed projection from Ammon's horn must be regarded as one bf the major features of this system. It is all the more surprising, therefore, that this component of the hippocampal formation has not been adequately recognized until quite recently (Votaw, '60; Hjorth-Simonsen, '71, '73). The only systematic analysis of these retrohippocampal projections is that of Hjorth-Simonsen ('71,'73).In 1971 this author identified a hippocampo-entorhinalprojection, which was said to arise in the temporal one-third of the regio inferior (principally field CA,) and to terminate in lamina IV of the medial entorhinal area. In their course through the hippocampus these fibers were described as passing through the stratum radiatum and stratum lacunosum-moleculareof field CAI. In a general sense the pathway ap-

peared to be topographically organized such that different levels in the septo-temporal axis of the hippocampus were found to project to different levels within the medial entorhinal area. However, although only the temporal one-third of field CA, contributed to the pathway the entire dorsoventral extent of the medial entorhinal area was said to receive a hippocampal input. In 1973 Hjorth-Simonsen identified a second caudally directed projection from field CAI to the deep part of the plexiform layer of the adjacent subiculum. This projection appeared to course through the alveus and again showed evidence of being topographically organized. We have been able to confirm the existence of both of these projections, but in addition, it would appear from our autoradiographic material that the caudally directed projections of Ammon's horn are considerably more extensive than has previously been described. Thus it would appear that field CA, not only projects to both the medial and lateral entorhinal areas, but also that it has a significant projection to the subiculum, to the pre- and parasubiculum, and to the cingulate area. Furthermore, our evidence suggests that these pathways are topographically organized in the sense that the septal portion of field CA, projects preferentially to the cingulate area and the presubiculum, whereas the projections to the parasubiculum and to the entorhinal area derive mainly from the intermediate (and perhaps temporal) regions of this field as HjorthSimonsen has suggested. Similarly, our material indicates that field CAI as a whole projects in an orderly manner to the subiculum, but the septal part of this field seems to have an additional projection to the perirhinal area. At present it is difficult to even speculate about the possible significance of these caudally directed projections, or to comment on their topographic organization. It is, however, worth noting in view of our findings on the rostrally directed projections of the hippocampus proper, that its


fields may still exert a significant influence on hypothalamic activity through two quite separate routes. The first is through the well-known projection to the lateral septa1 complex -which in turn projects into the medial forebrain bundle, a pathway we have dealt with elsewhere (Swanson and Cowan, '76). The second is through the caudally directed inputs to the subiculum, and to the pre- and parasubiculum, which as we shall see constitute the major if not the sole source of fibers in the postcommissural fornix.

11. The efferent connections of the subicular complex Undoubtedly the most significant finding in this study is the discovery that, in the rat, the entire postcommissural fornix, including the projections to the anterior thalamic and mammillary nuclei, have their origin in the subiculum and/or the pre- and parasubiculum, rather than in Ammon's horn as has been believed for close to a century. That some components of the postcommissural fornix arise from the subiculum was first suggested by Raisman et al. ('66) who observed degeneration in the socalled medial cortico-hypothalamic tract following lesions of the ventral part of the subicular complex. However, since the rest of the postcommissural system showed degeneration after lesions involving t h e various fields of Ammon's horn, they concluded (as had everyone before them) that the projections to the anterior thalamic and mammillary nuclei had their origin in the pyramidal cells of fields CA, through C&. It is now clear that lesions of Ammon's horn inevitably interrupt the rostrally directed fibers of the subicular complex which run in the alveus to the dorsal fornix and fimbria and that only a non-destructive method such as the autoradiographic technique could resolve this issue. It is worthwhile emphasizing, before dealing with the outflow from each of the fields of the subicular complex, that our findings pertain only to the rat. We have some reason to believe that a similar pat-

81

tern of projections exists in the brain of the mouse, but a comprehensive comparative study of the region is needed before we can confidently conclude that all fibers in the post-commissural fornix arise within the subicular complex. It is of particular importance to establish this point in the primate brain before extrapolations are made from our findings to the organization of the limbic system in man. We emphasize this because it is well-known that there are other differences in the organization of the hippocampal projections between species (Valenstein and Nauta, '59); and our own unpublished observations on the organization of the hippocampal connections in a variety of common laboratory mammals indicate that such differences may be more significant than has hitherto been suspected. Our evidence regarding the efferent connections of the subicular complex indicates that the rostrally directed efferents from t h e subiculum itself a r e topographically organized such that the dorsal part of the subiculum projects predominantly upon the mammillary complex, whereas the ventral part of the subiculum is responsible for most of the hippocampal input to the rostral hypothalamus (by way of the medial cortico-hypothalamic tract) and to a number of basal forebrain structures including the bed nucleus of the stria terminalis, the nucleus accumbens, the anterior olfactory nucleus, the taenia tecta, and the infralimbic area. Space does not permit a detailed consideration of these various projections, except to note that through its links with the various basal forebrain and hypothalamic areas the subicular complex is in a position to modulate the activity of most of the structures contributing to the rostral part of the medial forebrain bundle, and beyond them to such important brainstem structures as the locus coeruleus (for example, it is known that the bed nucleus of the stria terminalis projects to the locus coeruleus, see Swanson and Cowan, '76). Finally, for the sake of completeness, it should be mentioned that all

82


parts of the subiculum project in an organized manner to the lateral septal nucleus on the same side, overlapping extensively with the similar projections from Ammon’s horn. The finding that the efferent connections of the dorsal and ventral parts of the subiculum are quite different is of considerable interest since the area as a whole appears, cytoarchitectonically, to be a homogeneous field. However, it is becoming increasingly clear that despite the apparent uniformity of the field, in terms of its connections it really is comprised of two quite distinct subfields. Thus the ventral subiculum is now known to receive a substantial projection from the amygdala (Krettek and Price, ’74) as well as the topographically ordered input from the hippocampus that we and others have described. In view of the very clear-cut difference in the outflow from the two parts of the subiculum it would appear that this division into a dorsal and ventral subfield is one of the major determinants of the organization of the efferent projections of the hippocampal formation as a whole. It has been known for some time that field CAI can be divided on the basis of its connections into an anterior, or septal, and a caudal, or temporal, portion. The finding of an ordered projection from the hippocampus onto the two parts of the subiculumprovides further grounds for considering this division along the longitudinal axis of the hippocampus as one of the major organizational principles of the hippocampus as a whole. A second projection to the mammillary nuclei is derived from the pre- and parasubiculum and there is some suggestion in our material that these inputs are spatially organized in terms of their fields of origin. Additionally, the pre- and parasubiculum, together with what we have called the postsubiculum (which, as we have pointed out above, corresponds in part at least to the dorsal part of the presubiculum as defined by Blackstad, ’56) also gives rise to the “hippocampal” input to the anterior thalamic nuclei. It is presumably through these and the mammillary projections that

the “return loop” of the so-called Papez circuit is completed. Our evidence on the caudally directed connections from the subicular complex is less complete. In brief, this may be summarized by saying that the most extensive connections are derived from the pre- and parasubiculum which project to the entorhinal area (as Shipley has found, ’75) and to the parasubiculum (both projections being bilateral), and that the postsubiculum is connected with most of the retrohippocampal fields, including the cingulate and retrosplenial areas, the entorhinal area and the parasubiculum, the last two projections again being bilateral. It is also noteworthy that the subiculum itself (like field CAI) is one of the few areas in the hippocampal region which does not contribute a crossed projection to the hippocampal formation of the other side.

111. The efferent connections of the entorhinal cortex We have not attempted a systematic analysis of the efferent projections of the entorhinal area in this study, but so far as it goes our evidence has confirmed the general organization of the projections from the medial and lateral entorhinal areas upon the dentate gyrus and Ammon’s horn described by others (Raisman et al., ’65; Hjorth-Simonsen, ’72a,b), and has also established that these projections are indeed bilateral. The latter finding is of some significance since it had been suggested that the crossed temporo-ammonic pathway is not substantial in the brains of normal rats, and that it only becomes evident in a restricted portion of the hippocampus following early removal of the ipsilateral entorhinal cortex (Steward et al., ’74). To this extent our findings are in agreement with Stewards later work (’761, and are also confirmatory of the topographic organization of the entorhinal inputs to the dentate gyrus that he and Hjorth-Simonsen (’72a,b)have described. The only point to which we wish to draw attention regarding the entorhinal cortex is that while its projection by way of the per-


forant path to the dentate gyrus is undoubtedly its major outflow, the entorhinal area as a whole also provides inputs to both the regio inferior and the regio superior of Ammon’s horn. In our material it is clear that the latter connections are partly by way of the alveus. This pathway, which corresponds to the temporo-alvearpath of Lorente de N6 (’34) has been called into question by Hjorth-Simonsen (“72a,b) on the basis of degeneration studies, although it had been formerly demonstrated experimentally by Raisman et al. (’65) and has recently been confirmed in an autoradiographic study by Steward (’76).Since the entorhinal and perirhinal areas appear to be the major focus of convergence of neocortical inputs to the hippocampal formation (Cragg, ’65; Jones and Powell, ’70; Van Hoesen and Pandya, ’75; Van Hoesen et al., ’75) the further elucidation of this important region remains one of the most pressing problems in the morphology of the hippocampal formation. ACKNOWLEDGMENTS

We should like to thank Mrs. LueVerne Bell, Mrs. Pamela Seyer and Miss Lynn Rogers for their excellent technical assistance, and Mrs. Mary Murphy for typing the manuscript.Figures 1 and 2 were drawn by Mr. R. A. Filbey, Jr. LITERATURE CITED Blackstad, T. W. 1956 Commissural connections of the hippocampal region in the rat with special reference to their mode of termination. J. Comp. Neur., 105: 417-538. Blackstad, T. W., K. Brink, J. Hem and B. Jeune 1970 Distribution of hippocampal mossy fibers in the rat. An experimental study with silver impregnation methods. J. Comp. Neur., 138; 433-450. Brodmann, K. 1909 Vergleichende Lokalisationslehre der Grosshirnrhinde in ihren Prinzipien dargestellt auf Grund des Zellenbaues.T. A. Barth, Leipzig. Cajal, S. Ram6n y 1911 Histologie du Systbme Nerv e w de 1’Homme e t des Vertebrbs, Vol. 11. A. Maloine, Paris. Cowan, W. M., and M. Cubnod 1975 The use of axonal transport for the study of neural connections: A retrospective survey. In: The Use of Axonal Transport for Studies of Neuronal Connectivity. M.

83

Cuenod and W. M. Cowan, eds. Elsevier, Amsterdam, pp. 1-24. Cowan, W. M., D. I. Gottlieb, A. E. Hendrickson, J. L. Price and T. A. Woolsey 1972 The autoradiographic demonstration of axonal connections in the central nervous system. Brain Res., 37: 21-51. Cragg, B. G. 1965 Afferent connexionsof the allocortex. J. Anat., 99: 339-357. Goldowitz, D., W. I. White, 0.Steward, G. Lynch and C. Cotman 1975 Anatomical evidence for a projection from the entorhinal cortex to the contralateral dentate gyrus of the rat. Exp. Neurol., 47: 433-441. Cottlieb, D. I., and W. M. Cowan 1973 Autoradiographic studies of the commissural and ipsilateral association connections of the hippocampus and dentate gyrus of the rat. I. The commissural connections. ]. Comp. Neur., 149; 393-422. Guillery, R. W. 1956 Degeneration in the post-commissural fornix and the mammillary peduncle of the rat. J. Anat., 91: 350-370. Hjorth-Simonsen, A. 1971 Hippocampal efferents to the ipsilateral entorhinal area: An experimental study in the rat. J. Comp. Neur., 142: 417-438. 1972 Projection of the lateral part of the entorhinal area to the hippocampus and fascia dentata. J. Comp. Neur., 146: 219-232. 1973 Some intrinsic connnections of the hippocampus in the rat: An experimental analysis. J. Comp. Neur., 147: 145-162. Hjorth-Simonsen, A,, and B. Jeune 1972 Origin and termination of the hippocampal perforant path in the rat studied by silver impregnation. J. a m p . Neur., 144: 215-232. Jones, E. G., and T. P. S. Powell 1970 An anatomical study of converging sensory pathways within the cerebral cortex of the monkey. Brain, 93: 793-820. Krettek, J . E., and J. L. Price 1974 Projections from the amygdala to the perirhinal and entorhinal cortices and the subiculum. Brain Res., 71: 150-154. Krieg, W. J. S. 1946 Connections of the cerebral cortex. I. The albino rat. B. Structure of the cortical areas. J. Comp. Neur., 84: 277-323. LaVail, J. H. 1975 The retrograde transport method. Fed. Proc., 34: 1618-1624. LaVail, J. H., and M. M. LaVail 1972 Retrograde axonal transport in the central nervous system. Science, 176: 1416-1417. Lorente de N6, R. 1934 Studies on the structure of the cerebral cortex. 11. Continuation of the study of the ammonic system. J. Psychol. Neur., 46: 113-177. Nauta, W. J. H. 1956 An experimental study of the fornix system in the rat. J. Comp. Neur., 104: 247- 27 1. 1958 Hippocampal projections and related neural pathways to the mid-brain in the cat. Brain, 81: 319-340. Powell, T. P. S., and W. M. Cowan 1955 An experimental study of the efferent connections of the hippocampus. Brain, 78: 115-135. Powell, T. P. S., R. W. Guillery and W. M. Cowan 1957 A quantitative study of the fornix-mammillothalamic system. J. Anat., 91: 419-437.

-

-

84


Raisman, G., W. M. Cowan and T. P. S. Powell 1965 The extrinsic afferent, commissural and association fibres of the hippocampus. Brain, 88: 963-996. 1966 An experimental analysis of the efferent projection of the hippocampus. Brain, 89; 83-108. Rose, J. E., and C. N. Woolsey 1948 Structure and relations of limbic cortex and anterior thalamic nuclei in rabbit and cat. J. Comp. Neur., 89: 279-347. Rose, M. 1927 Gyrus limbicus anterior und regio retrosplenialis (cortex holoprotoptychos quinquestratificatus). Vergleichende Architektonik bei Tier und Mensch. J. Psychol. Neur., 35: 65-173. 1929 Cytoarchitektonischer Atlas der grosshimrinde der Maus. J. Psychol. Neur., 40; 1-51. Schlessinger, A. R., W. M. Cowan and D. I. Gottlieb 1975 An autoradiographic study of the time of origin and the pattern of granule cell migration in the dentate gyrus of the rat. J. G m p . Neur., 159: 149-176. Shipley, M. T. 1975 The topographical and laminar organization of the presubiculum’s projection to the ipsi- and contralateral entorhinal cortex in the guinea pig. J. Comp. Neur., 160: 127-146. Siegel, A., H. Edinger and S. Ohgami 1974 The topological organization of the hippocampal projection to the septal area: A comparative neuroanatomical analysis in the gerbil, rat, rabbit and cat. J. Comp. Neur., 157: 359-378. Steward, 0. 1976 Topographic organization of the projections from the entorhinal area to the hippocampal formation of the rat. J. Comp. Neur., 167: 285-314. Steward, O., C. W. Cotman and G. L. Lynch 1974

-

Growth of a new fiber projection in the brain of adult rats: Re-innervation of the dentate gyrus by the contralateral entorhinal cortex following ipsilateral entorhinal lesions. Exp. Brain Res., 30: 45-66. Swanson, L. W., and W. M. Cowan 1975 Hippocampo-hypothalamic connections: Origin in subicular cortex, not Ammon’s horn. Science, 189: 303-304. 1976 Autoradiographic studies of the development and connections of the septal area in the rat. In: The Septa1Nuclei. J. DeFrance, ed. In press. Valenstein, E. S., and W. J. H. Nauta 1959 A comparison of the distribution of the fornix system in the rat, guinea pig, cat, and monkey. J. Comp. Neur., 11 3: 337-363. Van Hoesen, G. W., and D. N. Pandya 1975 Some connections of the entorhinal (area 28) and perirhinal (area 35) cortices of the rhesus monkey. I. Temporal lobe afferents. Brain Res., 95; 1-24. Van Hoesen, G. W., D. N. Pandya and N. Butters 1975 Some connections of the entorhinal (area 28) and perirhinal (area 35) cortices of the rhesus monkey. 11. Frontal lobe afferents. Brain Res., 95: 25-38. Votaw, C. L. 1960 Certain functional and anatomical relations of the comu ammonis of the macaque monkey. J. Comp. Neur., 114: 283-293. White, L. E. 1959 Ipsilateral afferents to the hippocampal formation in the albino rat. 1. Cingulum projections. J. Comp. Neur., 113: 1-32. Zimmer, J. 1971 Ipsilateral afferents to the commissural zone of the fascia dentata, demonstrated in decommissurated rats by silver impregnation. J. Comp. Neur., 142: 393-416.

An autoradiographic study of the efferent connections of the lateral hypothalamic area in the rat.

An autoradiographic study of the efferent connections of the superior colliculus in the cat.

Efferent connections of the habenular nuclei in the rat.

An autoradiographic study of the organization of intrahippocampal association pathways in the rat.

Topographic organization of the cerebellar nucleocortical projection in the albino rat: an autoradiographic orthograde study.

An autoradiographic study of the efferent projections of the ventral lateral geniculate nucleus of the hooded rat.

Olfactory relationships of the telencephalon and diencephalon in the rabbit. II. An autoradiographic and horseradish peroxidase study of the efferent connections of the anterior olfactory nucleus.

Olfactory relationships of the telencephalon and diencephalon in the rabbit. I. An autoradiographic study of the efferent connections of the main and accessory olfactory bulbs.

Topographic organization of the projections from the entorhinal area to the hippocampal formation of the rat.

Connections between the retrosplenial cortex and the hippocampal formation in the rat: a review.

The efferent connections of the suprachiasmatic nucleus of the hypothalamus.

The afferent and efferent connections of the ventromedial thalamic nucleus in the rat.

Efferent connections of the caudate nucleus in the Virginia opossum.

Asymmetric cross-hemispheric connections link the rat anterior thalamic nuclei with the cortex and hippocampal formation.

Efferent connections of the substantia nigra and ventral tegmental area in the rat.

Afferent and efferent connections of the laterodorsal tegmental nucleus in the rat.

The efferent connections of the feline nucleus cuneatus.

Afferent connections of the parvocellular reticular formation: a horseradish peroxidase study in the rat.

Organization of cortical afferent and efferent pathways in the white matter of the rat visual system.

An autoradiographic study of the projections of the mammillothalamic tract in the rat.

Dompamine receptors in the rat frontal cortex: an autoradiographic study.

The laminar organization of certain afferent and efferent fiber systems in the rat somatosensory cortex.

Efferent projections from the flocculus in the albino rat as revealed by an autoradiographic orthograde tracing method.

An autoradiographic study of the projections from the lateral geniculate body of the rat.