Brain Research, 14l (1978) 251 265 19 Elsevier/North-Holland Biomedical Press
251
CORTICAL PROJECTIONS OF THE T H A L A M I C M E D I O D O R S A L NUCLEUS IN T H E RABBIT
ROBERT M. BENJAMIN, JAN
C. J A C K S O N and G R E G O R Y T.
GOLDEN
Department of Neurophysiology, University of Wisconsin Medical Center, Madison, Wisc. 73506 (u.s.A.)
(Accepted May 25th, 1977)
SUMMARY The cortical projection of the thalamic mediodorsal nuclear complex (MD) in the rabbit was mapped with retrograde horseradish peroxidase and anterograde tritiated proline techniques. The projection field occupied the entire medial wall rostral to a mid corpus callosal level, wrapped around the frontal pole onto the lateral convexity and tailed off caudally on the dorsal bank of the rhinal sulcus. The projection of the lateral approximately one-half of MD, the half which does not receive olfactory input, was confined to medial cortex supplying all but the most rostral region. This projection field of lateral MD was precisely organized in two dimensions with the most lateral part projecting most caudally and the most dorsal part projecting most ventrally. A representation for the third, anterior-posterior (A-P), dimension was not evident since any cortical point within the field was supplied by a cylinder of cells extending the entire A-P extent of lateral MD. The medial half of M D, which does receive olfactory input, projected to the remaining rostral medial cortex, the lateral convexity and rhinal sulcal region. The inverse dorsoventral relationship was partially preserved and an overlapping A-P gradient was present with sulcal projections originating more caudally in medial MD and the rostral medial projection originating more rostrally.
INTRODUCTION Two discoveries have forced a reevaluation of long static and firmly entrenched concepts of frontal lobe function. First was the belated appreciation that part of the mediodorsal thalamic complex (MD) which projects to frontal cortex receives a relatively direct olfactory input. Second was the revelation by methods more sensitive than retrograde chromatolysis that the cortical projection of MD, at least in the rat, was both considerably more extensive than was previously recognized and overlapped
252 the projections of a number of other thalamic nuclei. This paper examines the projection of MD in the rabbit with retrograde HRP and anterograde radioactive proline techniques with particular emphasis on the cortical projections from the olfactory vs. non-olfactory parts of MD. METHODS The results of this study are based on brains taken from 33 dark-eyed New Zealand rabbits (Oryctolagus cuniculus) weighing 2.2-3.7 kg. Because initial cases showed only ipsilateral cortical projections from MD, bilateral injections were made in later cases. Twenty-nine proline injections (L-[2,3-aH]protine; specific activity 27 Ci~mmote: New England Nuclear) were made in the thalami of 16 rabbits while t7 animals received bilateral cortical injections of H R P (Sigma, Type VI). The animals were anesthetized with sodium pentobarbital (i.v., to effect) or ketamine hydrochloride (44 mg/kg, i.m.) followed by sodium pentobarbital (i.v.. to effect). One-quarter milliliter of [aH]proline stock solution (1 mCi/mt) was dried by evaporation in a centrifuge and then reconstituted in 5 td of physiological ,saline to a nominal concentration of 50/zCi/#l. Thalamic injections of 0.1-0.4 #1 were accomplished with a one microliter Hamilton syringe and a specially adapted 29 gauge needle. Minimal injection sites were produced by deposition of 0.1 #1 made at a rate of 1/50 #l/min after which the needle was allowed to remain in place for an additional 5 min. Larger injection sites were produced by repeating this procedure at 0.8 mm anterior-posterior intervals. The rabbits survived for 2-8 days after injection and were then perfused through the heart with a 0.9~o saline solution followed by a 10% formalin-saline solution. The brains were placed in 10~o formalin-saline for a minimum of two weeks and then in a 30 To sucrose-I 0 °//oformalin-saline mixture until they sank. Frontal frozen sections were cut at 25 ,urn and every third section was mounted on gelatinized slides. These slides were dipped in a 1:1 mixture of Kodak NTB2 emulsion and 0.1 ~ Dreft solution, placed in light-tight boxes containing Drierite and exposed for 3-4 weeks at 4 °C. The slides were then developed in Kodak D-19 and the sections were stained through the emulsion with cresyl violet. A 10 ~ solution of horseradish peroxidase was injected into the cortex using an air pressure regulated micropipette (40-50 # m tip). The volume of H R P was either 0.03/zl or 0.06/zl injected gradually over a period of 8 or 16 min, respectively, after which the micropipette was left in place for an additional 5 min. In some animals multiple injections were made to produce larger injection sites. Animals were allowed to survive for approximately 34 h. were given an overdose of barbiturate anesthesia and perfused intracardially with one liter of heparinized 0.9)/o saline in 0.125 M phosphate buffer (pH -~ 7.4) at room temperature immediately followed by perfusion with two liters of either 4 ~o paraformaldehyde or 4 ~o glutaraldehyde, both in 0.125 M phosphate buffer (pH 7.4) chilled to 4 °C. All brains were removed promptly and placed in fresh fixative (identical to that used in the perfusion) overnight at 4 °C. and then over the next 24 h either taken through three changes of 3 0 ~ sucrose in Trizma buffer (paraformaldehyde fix) or through one change of 5 ~o sucrose and two changes
253 of 30 °/0 sucrose in 0.125 M phosphate buffer (glutaraldehyde fix). The brains were cut on a freezing microtome into 90 /~m thick coronal sections. Cut sections were immediately placed into Trizma buffer and then transferred to an incubation medium of Trizma buffer, 3',3'-diaminobenzidine, and hydrogen peroxide for 40 rain. Following incubation, the sections were washed for 5 min in three changes of Trizma buffer. The reacted sections were then mounted from Trizma buffer onto gelatinized slides, oven dried, dehydrated and coverslipped. Alternate sections were stained with cresyl violet without prior incubation in the reacting medium. A considerable effort was made to accurately represent the most important basic observations on standard diagrams of the rabbit brain to facilitate accurate topographic comparisons between data from different animals by the experimenters and the reader. The thalamic data were plotted on 4 standard frontal sections illustrated in Fig. 2 : (1) Through the rostral pole of M D. Approximately 15 ~,] of the total A-P extent of MD lies rostral to this level, 85 o~ caudal; (2) Just caudal to the anterodorsal nucleus, 40 jQ,/,level' (3) Halfway between levels 2 and 4, 60'.~/oolevel; (4) Through the initial descent of the thalamointerpeduncular tract, 80 "/ /o level. Throughout the paper, these are referred to as standard levels 1, 2, 3 or 4. In some illustrations only the oval outline of MD is represented (e.g. Fig. 3). The data were transferred from histological to standard sections with the aid of an X-Y plotter driven by potentiometers mounted on the microscope stage. The intensity of the label at the proline injection sites was estimated by a crude, but apparently meaningful process. Slides on a white background were examined without magnification. Easily discernable dark sites were grouped into one category represented by solid black as illustrated in Fig. 2, levels 1,2 and 3. Heavily labeled cell somata were confined to these sites. Moderately dark, but easily discernable sites were depicted by hatching (Fig. 2, level 4). Sites in the third category, which could only be seen under dark field illumination, were outlined. The light halos surrounding the dark and moderate injection sites were also outlined (Fig. 2, levels 1, 2, 3 and 4). Transported label was always associated with the most dark sites (solid black), was variable with those sites judged moderate (hatching) and was never observed with the light category (outlined). Halos which extended into MD on the other side never resulted in cortical labeling. Thus the only thalamic regions unequivocally responsible for observable transported label are indicated by solid black, and those potentially effective, by hatching. Each HRP reactive cell located under dark field was represented as a dot (Figs. 3 and 4). Only soma-like images with at least one extremity were accepted regardless of the intensity. The cortical data were plotted at a standard magnification on medial and lateral outlines averaged from several normal brains. The cortical labeling from thalamic proline injections was judged under dark field as either heavy (solid black), or light (outlined) as in Fig. 2. This unquantified but simple perceptual evaluation produced consistent results for any given field, but not necessarily equivalent results from brain
254
%
I I I 2 rnm
\
B
\ \ ' x ~
\ C
Fig. 1. A: the total cortical projection of M D divided into the field for the medial, olfactory part (hatching) and the lateral non-olfactory part (dots). B: cytoarchitectonic parcetla~ion of rabbit frontal cortex modified from Rose and Woolse> 19 and Rose ~1. The solid area approximates the extent of cortex responsive to electrical stimulation of the chorda tympani nerve and the tinguotonsiUar branch of the IX nerve ~9. All, anterior dorsal agranular insular area; CG, cingular area; t, granular insular area; IL, infralimbic area; LA, anterior limbic area; OF, orbitofrontal area; PC, postcentral area; PI, preinsular area; P R A E C G R , precentral granular area; PRAG,_precentral agranular area; RG, retrosplenial area: TT, taenia tecta. C: sensory and motor areas from Woolsey 2s.
255 to brain. Representation o f the cortical H R P injections was much more difficult and certainly less reliable. In the preparations with paraformaldehyde fixation, a darkly stained core and light surround could be recognized. The judgements were much more insecure with glutaraldehyde fixation (beginning with 76-226). The estimates of core and surround were represented by solid black and outlining respectively (Fig. 4). RESULTS Data from the retrograde H R P and the anterograde proline methods provided complimentary and in all cases compatible information about the total extent and the topographical organization o f the cortical projections of M D. The total extent of the projection is illustrated in Fig. 1A, the dots representing the projection from the lateral half of M D ; the hatching, the projection from the medial half o f MD. The basis for this division of M D was functional, not morphological. Single unit responses from electrical stimulation o f the olfactory bulb in the rabbit were confined to approximately the medial half of the M D nuclear complex TM. The lateral half was unresponsive. The projection field of the lateral, non-olfactory part is restricted to the medial wall extending from a mid corpus callosal level rostrally toward the frontal
A
I
2
3
4
70,1B
74,1B
77,1B
80,2B
75-96
-L
Fig. 2. Reconstruction of a thalamic proline injection in lateral MD. Below: location of the injection site on 4 standard frontal sections through MD, 1 most rostral, 4 most caudal. (See Methods for description.) On each section the oval shaped MD is divided into medial and lateral parts by the dashed line. The second dashed line marks the midline of the thalamus. Three densities of labeling are indicated; black, most intense; hatching, moderate; outlined, light. Above left: projections to medial and lateral cortex with the most intense labeling shown in solid black and light labeling in outl;ne. Above right: laminar distribution of label in the cortex plotted from section A at the location marked by the open rectangle. See methods for details of the reconstruction process. AD, anterodorsal nucleus; HL, lateral habenular nucleus; HM, medial habenular nucleus; MD, mediodorsal nucleus; S, stria medullaris; THP, thalamointerpeduncular tract.
256 pole. The projection of the medial, olfactory half of M D includes the rest of the frontal medial cortex, emerges onto the lateral surface and tails off caudally on the upper bank of the rhinal sulcus. These three regions will be referred to as the rostral medial, the convexity and the sulcal projection fields, respectively. The total projection field encompasses, but rarely respects the boundaries of a number of cytoarchitectonic (Fig. 1B) and functionally defined fields (Fig. 1B and C). Notice particularly that the sulcal projection lies ventral to and does not overlap the cortical taste area defined by mapping both potentials evoked by electrical stimulation of taste nerves as well as single unit responses to adequate taste stimulation 29. The projection of the lateral half of MD. the dotted area in Fig. I A. will be considered first. None of the thalamic proline injections was sufficiently large nor appropriately situated to exactly encompass the entire lateral half. One attempt is illustrated in Fig. 2. The injection site is plotted on the 4 standard frontal sections of MD. The most dense part of the injection site (solid black) is largely confined to lateral MD throughout most of its A-P extent. The most dense cortical label (solid black) was largely confined to the medial wall with only minor invasion of the rostral part. the province of medial MD. The total extent and the topographical organization of the 76 - 1 2 8 - R - 1 7 2
76-269-R-182
76-2t4-
L-163
76- 269-L-t82
;
76 ~146_L_177
J
76 - 2 2 6 - R- 2 0 2
~
•
7 6 =121 - R - 1 ~ 5
/
7 6 - 1 2 8 - L - t7;"
7 6 - 261 - L - 178
76-26i-R-178
Fig. 3. Reconstruction of the results of 10 cortical HRP injections. Center: location of the injection sites on the medial wall. Only the darkly stained core Of the injection sites is outlined (see meth0ds). Hatching locates the projection of the medial, olfactory half of MD. Surround: outlir~s of MD, one for each case, all at standard level 2~ Each dot locates a reactive cell. Hatching approximates the medial, olfactory part of MD.
257 lateral MD projection were more precisely defined with cortical HRP injections. Ten injection sites are outlined on the medial wall in Fig. 3. For each injection reactive thalamic cells were plotted only at standard level 2 (see Fig. 2) although they were distributed throughout the entire A-P extent of MD. Only the oval outline of MD at this level was drawn on the figure. Beginning with the most caudal injection site (76-146-L) and proceeding ventrally and rostrally across the cortex, the zone of stained thalamic cells initially in a far lateral position shifts in a dorsal and medial direction toward, and finally into, the medial olfactory part of MD (hatched area). Repeating the sequence, but this time proceeding dorsally and rostrally on the cortex,
76-121-R
I
2
5
4
143
p45
15~
154
S
B
76 -288-R
I
2
3
lao
~82
186
4 ~81a
I
76 - 2 6 6 - R
D
76 - 1 4 6 - R
I
2
3
4
1187
191
197
201
I
2
5
4
~74
IZe
JeZ
le6
Fig. 4. Left: location of 4 cortical H R P injections. Both the darkly stained central core (black) a n d the lightly stained s u r r o u n d (outlined) are indicated. T h e location of the approximately oval cortical taste area ~ is outlined by dashes in B, C a n d D. Right: location of reactive cells in M D on 4 s t a n d a r d sections. In case D no cells in M D were reactive, but were evident in the ventromedial nucleus.
258 a ventral and medial shift of thalamic cells occurs. The affected cells rarely occupy the most ventral part of MD, presumably because the injection sites do not usually reach the most dorsal border of the projection field. Only one thalamic section was required to represent the location of reactive cells because, in general, one cortical point within the field receives projections from a cylinder of thalamic cells extending the entire A-P dimension of MD. Thus. the three-dimensional lateral half of M D seems to have been
~r~
i
\--_ ..
~
_.-
~7
1
i//' ~:' i
4
j
't
'
f"
y
L
¸
/,"~i ~ a:
4 i
C~
259
i
J
i
Fig. 5. Reconstruction of 7 cases with proline injections in MD. See Fig. 2 for description.
260 transformed into a two-dimensional representation on the cortex by compressing the A-P dimension. The resulting half pancake is inverted top for bottom and laid on the cortex with its lateral edge posterior. The salient features of the lateral projection are summarized in Fig. 6. Notice that the lateral half of MD and its projection field are represented by circles, the D-V dimension by solid vs. open circles, the M-L dimension by the size of the circles, and that the thalamic A-P dimension is not differentially represented at the cortex. The topographical relationships of the projection field of the medial half of M D are more complicated for two reasons. The A-P dimension is represented on the cortex and the D-V representation becomes blurred especially on the lateral convexity. Three H R P cases (Fig. 4) illustrate the main relationships. The rostrat medial cortex (Fig. 4A) receives projections from the entire A-P extent of M D. but only sparsely from the caudal regions. The lateral convexity (Fig. 4B) also receives projections from the entire A-P extent, but only sparsely from the anterior regions. Finally, the sulcal cortex (Fig. 4C) receives projections primarily from the caudal 40 °Jo/of MD. Thus, as one moves from rostral medial cortex to the lateral convexity to the posterior sulcal region, the concentration of stained thalamic cells gradually shifts towards the posterior regions of MD. The proline injections plotted in Fig. 5 support these relationships and add some detail. First. the anterior medial cortex receives the heaviest projections from the rostral, approximately, three-quarters of medial M D. The smallest thalamic injection to affect only this cortical region is plotted in Fig. 5B and an injection which did not
2
4
Fig. 6. Schematic representation of the topographical relationships between MD and Cortex. Below: outlines of MD taken from frontal sections at rostral (standard level 2) and Caudal (standard level :4) levels. Above: medial (left) and lateral (right) cortex. The lateral, non,olfactory, half of MD projects only to the medial wall exclusive of the rostral part. Notice that ventral MD (soiid circles) projects to dorsal cortex and most lateral MD (large circles) projects to most posterior cortex. Hatching and solid black mark the medial, olfactory half of MD and its projections. The rostra! medial cortex receives projections primarily from rostrat MD: (small and medium triangles)and the lateral :convexity primarily from caudal MD (large triangles). Sulcal cortex (solid black) receives projections only from caudal and medial MD. For the sake of clarity the overlap inherent in the system has been eliminated.
261 affect the region, in 5E. The region was densely labeled in 5A, C and F and lightly labeled in 5G. Secondly, the lateral convexity exclusive of posterior sulcal cortex receives the heaviest projections from the caudal approximately three-quarters of medial MD, especially from its more lateral and ventral part. Case 5E illustrates a minimally sufficient injection. Cases 5B, C, D, F and G were not sufficient while 5A was adequate. Finally, the sulcal cortex receives the heaviest input from the caudal half of medial MD, particularly from the more medial and dorsal part. Case 5G was sufficient, but 5B and E were not. Labeling was heavy in 5D and F but weak in 5C. The main features of this rather complex presentation have been simplified in Fig. 6 by ignoring the overlap inherent in the system. The projections of the rostral medial wall are shown to originate only from rostral MD (small and medium triangles) and the convexity from caudal MD (large triangles). In both cases the inverse D-V relationship is appropriately coded. The sulcal system in solid black originates from caudal and medial MD. The thalamocortical fibers terminated in two distinct zones of cortical lamina (Figs. 2 and 5). With rare exceptions the outer portion of layer I was heavily labeled always in conjunction with a deep termination usually just above layer V in layer II1. Layer IV, where discernable, was usually labeled and the superficial part of layer V was at times included particularly in sulcal cortex where all deep layers were labeled. The HRP injections generated some information about the projections of other thalamic nuclei to MD fields. Fig. 7 summarizes the data which is not detailed nor complete. Crosshatching approximates the heaviest projections, parallel lines the regions which produced few labeled cells and the question marks indicate either probable projection areas which were not injected or parts of large injected areas which may have no projections from the nucleus in question. In brief: VA (ventroanterior). The major focus is situated on supracallosal coltex. AM (anteromedial). The projection as shown is coextensive with the projection of lateral M D. A topographical pattern was evident, but a dense projection for the ventral and posteromedial parts was not demonstrated. HRP injections rostral to the region labeled 'P,M,V' resulted in inconsistent reactions always in only a few ceils. One proline injection that strayed rostral to MD and was confined almost exclusively to the posterior half of AM, produced no labeling rostral to the level indicated by 'P,M,V' but dense labeling caudally in cingulate and retrosplenial cortex. One explanation for
AM
VA A,M,D
PT
Fig. 7. Plots of the cortical projections of the ventroanterior nucleus (VA), anteromedial nucleus (AM) and parataenial nucleus (PT). Crosshatching indicates the most dense projection area. See results section for further description and limitations.
262 the minimal H R P effects from injections rostral to 'P,M,V' may be that injections in this region do not affect axon terminals, but rather supply damaged fibers destined for more posterior cortex. Thus, this region may not be part of the cortical A M field at all. PT (parataenialis). The focus is situated on the rostral medial wall within the projection of the medial half of MD. A very few cells were labeled f r o m injections outside this focus. Proline injections of MD revealed strong connections to caudate-putamen, olfactory tubercle and other regions. They will not be described further here. DISCUSSION The division of the rabbit M D complex into medial and lateral parts was based on a functional rather than morphological criterion. Single unit responses to electrical stimulation of the olfactory bulb were confined to the medial, approximately one half 10. In squirrel monkey the equivalent responsive area could be readily identified on cytoarchitectonic grounds as the magnoceUular division ~. Responsive units never violated the border between medial magnocellular and more lateral parvocetlular territory. In the opossum a clear correspondence was also demonstrated between functional and morphological characteristics. Essentially all the opossum MD is magnocellular7, ~6 and essentially all was responsive to olfactory tract stimulation 1~. Unfortunately, the rabbit M D complex is cyt oarchitectonically complex ~2 and no simple correlation could be made between a morphologically distinct area and olfactory responses. The rat M D has been described as totally magnocellular 14 and olfactory responses have been recorded 12,27. but their distribution has not been mapped m detail. It seems unlikely that the whole rat MD will prove responsive. Olfactory responses in the cat were described only as being mediaP °,16 and the relationships to morphology not reported. Because of the only occasional correlation with cytoarchitectonic fields, the functional division of M D into olfactory vs. non-olfactory zones seems to be the most satisfactory, if not the only, division tbr comparative purposes at present. Undoubtedly, more meaningful functional and anatomical parcellations will evolve. The first demonstration of anatomical connections between olfactory structures and M D by Powell et at. in t96517 has been confirmed in several subsequent studies, all on the rat 9,1z,~5,2z. Prepyriform cortex and olfactory tubercle distribute afferents to the central part of rat MD. not to the lateral segment, but also not to the most medial segment. Whether this most medial segment of rat M D would exhibit olfactory responses like the rabbit and squirrel monkey is not known. If so, they must be mediated byother olfactory structures or perhaps f r o m unexplored parts of prepyriform cortex or tubercle. If this medial segment is not responsive, that rat M D must be constructed from a different basic plan than that for the rabbit and squirrel monkey. Previous studies limited the cortical projection of M D in the rabbit to a cone of cortex capping the frontal pole 20. a disturbingly small fraction of the total projection established in this study. The explanation for this considerable discrepancy undoubtedly lies in the relative insensitivity of the retrograde chromatolysis method, the
263 only method used previously. The discrepancy is even more striking in the rat where frontal pole lesions failed to elicit any detectable thalamic chromatolysis 14, yet Leonard's analysis of Fink-Heimer material following electrolytic lesions of MD demonstrated projections to the entire medial wall rostral to the genu of the corpus callosum as well as to cortex on the dorsal lip of the rhinal sulcusH, l'~. Other studies using Fink-Heimer s, H R P 4 and autoradiographic tracer techniques 13 have confirmed these dual projections, have extended the caudal boundary on the medial wall to the rostral border of cingulate and retrosplenial cortex ~,s and have established that the sulcal projection does not invade the cortical taste area is as previously reported 1~. Just how closely the topographical organization of the rabbit thalamocortical projections corresponds to the organization in the rat is difficult to assess. In one respect, at least, they fit quite well. The lateral half of rabbit MD projects only to the medial wall exclusive of the most rostral part which along with sulcal cortex is territory belonging to the medial half of MD (Fig. 1A). In the rat Beckstead 4 noted that H R P injections in medial cortex resulted in stained cells confined to the lateral one-third or one-half of M D. Since his injections did not invade the most rostral medial cortex which in the rabbit receives projections from medial MD, the results are quite compatible. Beckstead also stated that sulcal injections were correlated with medial M D effects and Leonard 14 mentioned that one posterior lesion limited to the dorsomedial portion of MD produced degeneration limited to sulcal cortex. These statements indicate that the M-L dimension of MD is equivalently represented in the cortex of rat and rabbit. Much of the rabbit M D cortical projection on the medial wall is shared by other thalamic nuclei (Fig. 7). The entire field of lateral MD may be overlapped by AM projections as in the rat 4,s. Heavy projections were apparent in H R P material from all except the ventral and posteromedial portions of AM which project only sparsely, if at all, to rostral and dorsal medial cortex as discussed in the results section. The supracallosal portion of the lateral MD field, the region of the supplementary motor area (Fig. I B), is supplied by a moderate number of VA cells. The field of the medial, olfactory half of MD on the rostral medial wall receives strong input from all parts of PT. The paradigm for the MD projection system is the macaque system analyzed by countless retrograde chromatolysis studies 1. Three divisions of MD are generally recognized, each cytoarchitectonically distinct, each with its own cortical projection and each related to different functions. The medial magnocellular division projects to the orbital surface and is related to olfaction2,3,6,z4, 25. The more lateral parvocellular division projects to the frontolateral convexity and has always been overburdened with delayed response type behavior. The most lateral paralamellar division projects to the frontal eye fields and is involved with visual function. At present it is reasonable to equate the medial, magnocellular division of the monkey M D with the medial, olfactory part of the rabbit M D and, therefore, the frontal orbital cortex of the monkey with the sulcal cortex of the rabbit. Further comparisons might well be postponed until additional data is available on both species.
264
ACKNOWLEDGEMENTS S u p p o r t e d by G r a n t s NS-12721, NS-00782 a n d NS-07026. We wish to express a p p r e c i a t i o n to m e m b e r s o f o u r h i st o l o g y l a b o r a t o r y ( J o A n n Ekleberry, C a t h y G o o c h , J o a n Meister, A n n Switzky, a n d especially Inge S i g g e l k o w ) for their usual excellent product. C a r o l D i z a c k t r a n s f o r m e d o u r crude drawings into the final illustrations which were p h o t o g r a p h e d by Mr. T. P. Stewart,
REFERENCES 1 Akert, K., Comparative anatomy of frontal cortex and thalamofrontal connections. In J. M. Warren and K. Akert (Eds.), The Frontal Granular Cortex and Behavior, McGraw-Hill, New York. 1964, pp. 372-396. 2 Allen, W. F., Effect of ablating the frontal lobes, hippocampi, and occipoto-parieto-temporal (excepting pyriform areas) lobes on positive and negative olfactory conditioned reflexes. Amer. J. PhysioL, 128 (1940) 754-771. 3 Allen, W. F., Results of prefrontal lobectomy on acquired and acquiring correct conditioned differential responses with auditory, general cutaneous and optic stimuli, Amer..I. PhysioL, 139 (1943) 525-531. 4 Beckstead, R. M., Convergent thalamic and mesencephalic projections to the anterior medial cortex in the rat. J. comp. NeuroL, 166 (1976) 403~,t6. 5 Benjamin, R. M. and Akert, K., Cortical and thalamic areas involved in taste discrimination in albino rat, J. comp. NeuroL, 111 (1959) 231-260. 6 Benjamin, R. M and Jackson, J. C., Units discharges in the mediodorsal nucleus of the squirrel monkey evoked by electrical stimulation of the olfactory bulb, Brain Research. 75 (1974) 181-191. 7 Bodian, D., Studies on the diencephalon of the Virginia opossum. Part I: the nuclear pattern in the adult, J. comp. Neurol., 71 (1939) 259-323. 8 Domesick, V. B., Thalamic relationships of the medial cortex in the rat, Brain Behav. Evol.. 6 (1972) 457483. 9 Heimer, L., The olfactory connections of the diencephalon in the rat. Brain Behav. Evol.. 6 (1972) 484-523. 10 Jackson, J. C. and Benjamin, R. M., Unit discharges in the mediodorsal nucleus of the rabbit evoked by electrical stimulation of the olfactory bulb, Brain Research, 75 (1974) 193-201. 11 Jackson, J. C., Golden, G. T. and Benjamin, R. M.. The distribution of olfactory input in the opossum mediodorsal nucleus, Brain Research, submitted for publication. 12 Komisaruk, B. R. and Beyer, C., Responses of diencephalic neurons to olfactory bulb stimulation. odor and arousal, Brain Research, 36 (1972) 153-170. 13 Krettek, J. E. and Price, J. L.. A direct input from the amygdala to the thalamus and the cerebral cortex, Brain Research. 67 (1974) 269-274. 14 Leonard, C. M., The prefrontal cortex of the rat. 1. Cortical projection of the mediodorsal nucleus. II. Efferent connections, Brain Research, 12 (1969) 321-343. 15 Leonard, C. M., The connections of the dorsomedial nuclei, Brain Behav. Evol., 6 (1972) 524-54l. 16 Motokizawa, F., Olfactory input to the thalamus: electrophyslological evidence, Brain Research. 67 (1974) 334-337. 17 Powell, T. P. S., Cowan, W. M. and Raisman, G., The central olfactory connections, J. Anat. (Lond.), 99 (1965) 791-813. 18 Price, J. L., Personal communication. 19 Rose, J. E. and Woolsey, C. N., Structure and relations of limbic cortex and anterior thalamic nuclei in rabbit and cat, J. comp. Neurol., 89 (1948) 279-340. 20 Rose, J. E. and Woolsey, C. N., The orbitofrontal cortex and its connections with the mediodorsal nucleus in rabbit, sheep and cat, Res. Publ. Ass. herr. merit. Dis., 27 (1948) 210-232. 21 Rose, M., Zytoarchitekt. Atlas der Grosshirnrinde des Kaninchens. J. Psychol. Neurol. (Lpz.), 43 (1931) 353-340.
265 22 Rose, M., Das Zwischenhirn des Kaninchens, Extrait des M~m. Acad. pol. Sci., Serie B, 1935, pp. 1 108. 23 Scott, J. W. and Leonard, C. M., The olfactory connections of the lateral hypothalamus in the rat, mouse and hamster, J. comp. Neurol., 141 (1971) 331-344. 24 Tanabe, T., Ilno, M. and Takogi, S. F., Distribution of odors in the olfactory bulb, pyrifGrmamygdaloid areas, and orbitofrontal cortex of the monkey, J. Neurophysiol., 38 (1975) 1284-1296. 25 Tanabe, T., Yarita, H., lino, M., Ooshima, Y. and Takogi, S. F., An olfactory projection area in orbitofrontal cortex of the monkey, J. Neurophysiol., 38 (1975) 1283-1296. 26 Tobias, T. J. and Ebner, F. F., Thalamocortical projections from the mediodorsal nucleus in the Virginia opossum, Brain Research, 52 (1973) 79-96. 27 Wedgwood, M., Connexions between olfactory bulb and habenula and the dorsomedial thalamic nuclei, J. Physiol. (Lond.), 239 (1974) 88-89. 28 Woolsey, C. N., Organization of somatic sensory and motor areas in the cortex. In H. F. Harlow and C. N. Woolsey (Eds.), Biological and Biochemical Bases t~/"Behavior, Univ. of Wisconsin Press, Madison, 1958, pp. 63 81. 29 Yamamoto, T. and Kawamura, Y., Cortical responses to electrical and gustatory stimuli in the rabbit, Brain Research, 94 (1975) 447-463.
251
CORTICAL PROJECTIONS OF THE T H A L A M I C M E D I O D O R S A L NUCLEUS IN T H E RABBIT
ROBERT M. BENJAMIN, JAN
C. J A C K S O N and G R E G O R Y T.
GOLDEN
Department of Neurophysiology, University of Wisconsin Medical Center, Madison, Wisc. 73506 (u.s.A.)
(Accepted May 25th, 1977)
SUMMARY The cortical projection of the thalamic mediodorsal nuclear complex (MD) in the rabbit was mapped with retrograde horseradish peroxidase and anterograde tritiated proline techniques. The projection field occupied the entire medial wall rostral to a mid corpus callosal level, wrapped around the frontal pole onto the lateral convexity and tailed off caudally on the dorsal bank of the rhinal sulcus. The projection of the lateral approximately one-half of MD, the half which does not receive olfactory input, was confined to medial cortex supplying all but the most rostral region. This projection field of lateral MD was precisely organized in two dimensions with the most lateral part projecting most caudally and the most dorsal part projecting most ventrally. A representation for the third, anterior-posterior (A-P), dimension was not evident since any cortical point within the field was supplied by a cylinder of cells extending the entire A-P extent of lateral MD. The medial half of M D, which does receive olfactory input, projected to the remaining rostral medial cortex, the lateral convexity and rhinal sulcal region. The inverse dorsoventral relationship was partially preserved and an overlapping A-P gradient was present with sulcal projections originating more caudally in medial MD and the rostral medial projection originating more rostrally.
INTRODUCTION Two discoveries have forced a reevaluation of long static and firmly entrenched concepts of frontal lobe function. First was the belated appreciation that part of the mediodorsal thalamic complex (MD) which projects to frontal cortex receives a relatively direct olfactory input. Second was the revelation by methods more sensitive than retrograde chromatolysis that the cortical projection of MD, at least in the rat, was both considerably more extensive than was previously recognized and overlapped
252 the projections of a number of other thalamic nuclei. This paper examines the projection of MD in the rabbit with retrograde HRP and anterograde radioactive proline techniques with particular emphasis on the cortical projections from the olfactory vs. non-olfactory parts of MD. METHODS The results of this study are based on brains taken from 33 dark-eyed New Zealand rabbits (Oryctolagus cuniculus) weighing 2.2-3.7 kg. Because initial cases showed only ipsilateral cortical projections from MD, bilateral injections were made in later cases. Twenty-nine proline injections (L-[2,3-aH]protine; specific activity 27 Ci~mmote: New England Nuclear) were made in the thalami of 16 rabbits while t7 animals received bilateral cortical injections of H R P (Sigma, Type VI). The animals were anesthetized with sodium pentobarbital (i.v., to effect) or ketamine hydrochloride (44 mg/kg, i.m.) followed by sodium pentobarbital (i.v.. to effect). One-quarter milliliter of [aH]proline stock solution (1 mCi/mt) was dried by evaporation in a centrifuge and then reconstituted in 5 td of physiological ,saline to a nominal concentration of 50/zCi/#l. Thalamic injections of 0.1-0.4 #1 were accomplished with a one microliter Hamilton syringe and a specially adapted 29 gauge needle. Minimal injection sites were produced by deposition of 0.1 #1 made at a rate of 1/50 #l/min after which the needle was allowed to remain in place for an additional 5 min. Larger injection sites were produced by repeating this procedure at 0.8 mm anterior-posterior intervals. The rabbits survived for 2-8 days after injection and were then perfused through the heart with a 0.9~o saline solution followed by a 10% formalin-saline solution. The brains were placed in 10~o formalin-saline for a minimum of two weeks and then in a 30 To sucrose-I 0 °//oformalin-saline mixture until they sank. Frontal frozen sections were cut at 25 ,urn and every third section was mounted on gelatinized slides. These slides were dipped in a 1:1 mixture of Kodak NTB2 emulsion and 0.1 ~ Dreft solution, placed in light-tight boxes containing Drierite and exposed for 3-4 weeks at 4 °C. The slides were then developed in Kodak D-19 and the sections were stained through the emulsion with cresyl violet. A 10 ~ solution of horseradish peroxidase was injected into the cortex using an air pressure regulated micropipette (40-50 # m tip). The volume of H R P was either 0.03/zl or 0.06/zl injected gradually over a period of 8 or 16 min, respectively, after which the micropipette was left in place for an additional 5 min. In some animals multiple injections were made to produce larger injection sites. Animals were allowed to survive for approximately 34 h. were given an overdose of barbiturate anesthesia and perfused intracardially with one liter of heparinized 0.9)/o saline in 0.125 M phosphate buffer (pH -~ 7.4) at room temperature immediately followed by perfusion with two liters of either 4 ~o paraformaldehyde or 4 ~o glutaraldehyde, both in 0.125 M phosphate buffer (pH 7.4) chilled to 4 °C. All brains were removed promptly and placed in fresh fixative (identical to that used in the perfusion) overnight at 4 °C. and then over the next 24 h either taken through three changes of 3 0 ~ sucrose in Trizma buffer (paraformaldehyde fix) or through one change of 5 ~o sucrose and two changes
253 of 30 °/0 sucrose in 0.125 M phosphate buffer (glutaraldehyde fix). The brains were cut on a freezing microtome into 90 /~m thick coronal sections. Cut sections were immediately placed into Trizma buffer and then transferred to an incubation medium of Trizma buffer, 3',3'-diaminobenzidine, and hydrogen peroxide for 40 rain. Following incubation, the sections were washed for 5 min in three changes of Trizma buffer. The reacted sections were then mounted from Trizma buffer onto gelatinized slides, oven dried, dehydrated and coverslipped. Alternate sections were stained with cresyl violet without prior incubation in the reacting medium. A considerable effort was made to accurately represent the most important basic observations on standard diagrams of the rabbit brain to facilitate accurate topographic comparisons between data from different animals by the experimenters and the reader. The thalamic data were plotted on 4 standard frontal sections illustrated in Fig. 2 : (1) Through the rostral pole of M D. Approximately 15 ~,] of the total A-P extent of MD lies rostral to this level, 85 o~ caudal; (2) Just caudal to the anterodorsal nucleus, 40 jQ,/,level' (3) Halfway between levels 2 and 4, 60'.~/oolevel; (4) Through the initial descent of the thalamointerpeduncular tract, 80 "/ /o level. Throughout the paper, these are referred to as standard levels 1, 2, 3 or 4. In some illustrations only the oval outline of MD is represented (e.g. Fig. 3). The data were transferred from histological to standard sections with the aid of an X-Y plotter driven by potentiometers mounted on the microscope stage. The intensity of the label at the proline injection sites was estimated by a crude, but apparently meaningful process. Slides on a white background were examined without magnification. Easily discernable dark sites were grouped into one category represented by solid black as illustrated in Fig. 2, levels 1,2 and 3. Heavily labeled cell somata were confined to these sites. Moderately dark, but easily discernable sites were depicted by hatching (Fig. 2, level 4). Sites in the third category, which could only be seen under dark field illumination, were outlined. The light halos surrounding the dark and moderate injection sites were also outlined (Fig. 2, levels 1, 2, 3 and 4). Transported label was always associated with the most dark sites (solid black), was variable with those sites judged moderate (hatching) and was never observed with the light category (outlined). Halos which extended into MD on the other side never resulted in cortical labeling. Thus the only thalamic regions unequivocally responsible for observable transported label are indicated by solid black, and those potentially effective, by hatching. Each HRP reactive cell located under dark field was represented as a dot (Figs. 3 and 4). Only soma-like images with at least one extremity were accepted regardless of the intensity. The cortical data were plotted at a standard magnification on medial and lateral outlines averaged from several normal brains. The cortical labeling from thalamic proline injections was judged under dark field as either heavy (solid black), or light (outlined) as in Fig. 2. This unquantified but simple perceptual evaluation produced consistent results for any given field, but not necessarily equivalent results from brain
254
%
I I I 2 rnm
\
B
\ \ ' x ~
\ C
Fig. 1. A: the total cortical projection of M D divided into the field for the medial, olfactory part (hatching) and the lateral non-olfactory part (dots). B: cytoarchitectonic parcetla~ion of rabbit frontal cortex modified from Rose and Woolse> 19 and Rose ~1. The solid area approximates the extent of cortex responsive to electrical stimulation of the chorda tympani nerve and the tinguotonsiUar branch of the IX nerve ~9. All, anterior dorsal agranular insular area; CG, cingular area; t, granular insular area; IL, infralimbic area; LA, anterior limbic area; OF, orbitofrontal area; PC, postcentral area; PI, preinsular area; P R A E C G R , precentral granular area; PRAG,_precentral agranular area; RG, retrosplenial area: TT, taenia tecta. C: sensory and motor areas from Woolsey 2s.
255 to brain. Representation o f the cortical H R P injections was much more difficult and certainly less reliable. In the preparations with paraformaldehyde fixation, a darkly stained core and light surround could be recognized. The judgements were much more insecure with glutaraldehyde fixation (beginning with 76-226). The estimates of core and surround were represented by solid black and outlining respectively (Fig. 4). RESULTS Data from the retrograde H R P and the anterograde proline methods provided complimentary and in all cases compatible information about the total extent and the topographical organization o f the cortical projections of M D. The total extent of the projection is illustrated in Fig. 1A, the dots representing the projection from the lateral half of M D ; the hatching, the projection from the medial half o f MD. The basis for this division of M D was functional, not morphological. Single unit responses from electrical stimulation o f the olfactory bulb in the rabbit were confined to approximately the medial half of the M D nuclear complex TM. The lateral half was unresponsive. The projection field of the lateral, non-olfactory part is restricted to the medial wall extending from a mid corpus callosal level rostrally toward the frontal
A
I
2
3
4
70,1B
74,1B
77,1B
80,2B
75-96
-L
Fig. 2. Reconstruction of a thalamic proline injection in lateral MD. Below: location of the injection site on 4 standard frontal sections through MD, 1 most rostral, 4 most caudal. (See Methods for description.) On each section the oval shaped MD is divided into medial and lateral parts by the dashed line. The second dashed line marks the midline of the thalamus. Three densities of labeling are indicated; black, most intense; hatching, moderate; outlined, light. Above left: projections to medial and lateral cortex with the most intense labeling shown in solid black and light labeling in outl;ne. Above right: laminar distribution of label in the cortex plotted from section A at the location marked by the open rectangle. See methods for details of the reconstruction process. AD, anterodorsal nucleus; HL, lateral habenular nucleus; HM, medial habenular nucleus; MD, mediodorsal nucleus; S, stria medullaris; THP, thalamointerpeduncular tract.
256 pole. The projection of the medial, olfactory half of M D includes the rest of the frontal medial cortex, emerges onto the lateral surface and tails off caudally on the upper bank of the rhinal sulcus. These three regions will be referred to as the rostral medial, the convexity and the sulcal projection fields, respectively. The total projection field encompasses, but rarely respects the boundaries of a number of cytoarchitectonic (Fig. 1B) and functionally defined fields (Fig. 1B and C). Notice particularly that the sulcal projection lies ventral to and does not overlap the cortical taste area defined by mapping both potentials evoked by electrical stimulation of taste nerves as well as single unit responses to adequate taste stimulation 29. The projection of the lateral half of MD. the dotted area in Fig. I A. will be considered first. None of the thalamic proline injections was sufficiently large nor appropriately situated to exactly encompass the entire lateral half. One attempt is illustrated in Fig. 2. The injection site is plotted on the 4 standard frontal sections of MD. The most dense part of the injection site (solid black) is largely confined to lateral MD throughout most of its A-P extent. The most dense cortical label (solid black) was largely confined to the medial wall with only minor invasion of the rostral part. the province of medial MD. The total extent and the topographical organization of the 76 - 1 2 8 - R - 1 7 2
76-269-R-182
76-2t4-
L-163
76- 269-L-t82
;
76 ~146_L_177
J
76 - 2 2 6 - R- 2 0 2
~
•
7 6 =121 - R - 1 ~ 5
/
7 6 - 1 2 8 - L - t7;"
7 6 - 261 - L - 178
76-26i-R-178
Fig. 3. Reconstruction of the results of 10 cortical HRP injections. Center: location of the injection sites on the medial wall. Only the darkly stained core Of the injection sites is outlined (see meth0ds). Hatching locates the projection of the medial, olfactory half of MD. Surround: outlir~s of MD, one for each case, all at standard level 2~ Each dot locates a reactive cell. Hatching approximates the medial, olfactory part of MD.
257 lateral MD projection were more precisely defined with cortical HRP injections. Ten injection sites are outlined on the medial wall in Fig. 3. For each injection reactive thalamic cells were plotted only at standard level 2 (see Fig. 2) although they were distributed throughout the entire A-P extent of MD. Only the oval outline of MD at this level was drawn on the figure. Beginning with the most caudal injection site (76-146-L) and proceeding ventrally and rostrally across the cortex, the zone of stained thalamic cells initially in a far lateral position shifts in a dorsal and medial direction toward, and finally into, the medial olfactory part of MD (hatched area). Repeating the sequence, but this time proceeding dorsally and rostrally on the cortex,
76-121-R
I
2
5
4
143
p45
15~
154
S
B
76 -288-R
I
2
3
lao
~82
186
4 ~81a
I
76 - 2 6 6 - R
D
76 - 1 4 6 - R
I
2
3
4
1187
191
197
201
I
2
5
4
~74
IZe
JeZ
le6
Fig. 4. Left: location of 4 cortical H R P injections. Both the darkly stained central core (black) a n d the lightly stained s u r r o u n d (outlined) are indicated. T h e location of the approximately oval cortical taste area ~ is outlined by dashes in B, C a n d D. Right: location of reactive cells in M D on 4 s t a n d a r d sections. In case D no cells in M D were reactive, but were evident in the ventromedial nucleus.
258 a ventral and medial shift of thalamic cells occurs. The affected cells rarely occupy the most ventral part of MD, presumably because the injection sites do not usually reach the most dorsal border of the projection field. Only one thalamic section was required to represent the location of reactive cells because, in general, one cortical point within the field receives projections from a cylinder of thalamic cells extending the entire A-P dimension of MD. Thus. the three-dimensional lateral half of M D seems to have been
~r~
i
\--_ ..
~
_.-
~7
1
i//' ~:' i
4
j
't
'
f"
y
L
¸
/,"~i ~ a:
4 i
C~
259
i
J
i
Fig. 5. Reconstruction of 7 cases with proline injections in MD. See Fig. 2 for description.
260 transformed into a two-dimensional representation on the cortex by compressing the A-P dimension. The resulting half pancake is inverted top for bottom and laid on the cortex with its lateral edge posterior. The salient features of the lateral projection are summarized in Fig. 6. Notice that the lateral half of MD and its projection field are represented by circles, the D-V dimension by solid vs. open circles, the M-L dimension by the size of the circles, and that the thalamic A-P dimension is not differentially represented at the cortex. The topographical relationships of the projection field of the medial half of M D are more complicated for two reasons. The A-P dimension is represented on the cortex and the D-V representation becomes blurred especially on the lateral convexity. Three H R P cases (Fig. 4) illustrate the main relationships. The rostrat medial cortex (Fig. 4A) receives projections from the entire A-P extent of M D. but only sparsely from the caudal regions. The lateral convexity (Fig. 4B) also receives projections from the entire A-P extent, but only sparsely from the anterior regions. Finally, the sulcal cortex (Fig. 4C) receives projections primarily from the caudal 40 °Jo/of MD. Thus, as one moves from rostral medial cortex to the lateral convexity to the posterior sulcal region, the concentration of stained thalamic cells gradually shifts towards the posterior regions of MD. The proline injections plotted in Fig. 5 support these relationships and add some detail. First. the anterior medial cortex receives the heaviest projections from the rostral, approximately, three-quarters of medial M D. The smallest thalamic injection to affect only this cortical region is plotted in Fig. 5B and an injection which did not
2
4
Fig. 6. Schematic representation of the topographical relationships between MD and Cortex. Below: outlines of MD taken from frontal sections at rostral (standard level 2) and Caudal (standard level :4) levels. Above: medial (left) and lateral (right) cortex. The lateral, non,olfactory, half of MD projects only to the medial wall exclusive of the rostral part. Notice that ventral MD (soiid circles) projects to dorsal cortex and most lateral MD (large circles) projects to most posterior cortex. Hatching and solid black mark the medial, olfactory half of MD and its projections. The rostra! medial cortex receives projections primarily from rostrat MD: (small and medium triangles)and the lateral :convexity primarily from caudal MD (large triangles). Sulcal cortex (solid black) receives projections only from caudal and medial MD. For the sake of clarity the overlap inherent in the system has been eliminated.
261 affect the region, in 5E. The region was densely labeled in 5A, C and F and lightly labeled in 5G. Secondly, the lateral convexity exclusive of posterior sulcal cortex receives the heaviest projections from the caudal approximately three-quarters of medial MD, especially from its more lateral and ventral part. Case 5E illustrates a minimally sufficient injection. Cases 5B, C, D, F and G were not sufficient while 5A was adequate. Finally, the sulcal cortex receives the heaviest input from the caudal half of medial MD, particularly from the more medial and dorsal part. Case 5G was sufficient, but 5B and E were not. Labeling was heavy in 5D and F but weak in 5C. The main features of this rather complex presentation have been simplified in Fig. 6 by ignoring the overlap inherent in the system. The projections of the rostral medial wall are shown to originate only from rostral MD (small and medium triangles) and the convexity from caudal MD (large triangles). In both cases the inverse D-V relationship is appropriately coded. The sulcal system in solid black originates from caudal and medial MD. The thalamocortical fibers terminated in two distinct zones of cortical lamina (Figs. 2 and 5). With rare exceptions the outer portion of layer I was heavily labeled always in conjunction with a deep termination usually just above layer V in layer II1. Layer IV, where discernable, was usually labeled and the superficial part of layer V was at times included particularly in sulcal cortex where all deep layers were labeled. The HRP injections generated some information about the projections of other thalamic nuclei to MD fields. Fig. 7 summarizes the data which is not detailed nor complete. Crosshatching approximates the heaviest projections, parallel lines the regions which produced few labeled cells and the question marks indicate either probable projection areas which were not injected or parts of large injected areas which may have no projections from the nucleus in question. In brief: VA (ventroanterior). The major focus is situated on supracallosal coltex. AM (anteromedial). The projection as shown is coextensive with the projection of lateral M D. A topographical pattern was evident, but a dense projection for the ventral and posteromedial parts was not demonstrated. HRP injections rostral to the region labeled 'P,M,V' resulted in inconsistent reactions always in only a few ceils. One proline injection that strayed rostral to MD and was confined almost exclusively to the posterior half of AM, produced no labeling rostral to the level indicated by 'P,M,V' but dense labeling caudally in cingulate and retrosplenial cortex. One explanation for
AM
VA A,M,D
PT
Fig. 7. Plots of the cortical projections of the ventroanterior nucleus (VA), anteromedial nucleus (AM) and parataenial nucleus (PT). Crosshatching indicates the most dense projection area. See results section for further description and limitations.
262 the minimal H R P effects from injections rostral to 'P,M,V' may be that injections in this region do not affect axon terminals, but rather supply damaged fibers destined for more posterior cortex. Thus, this region may not be part of the cortical A M field at all. PT (parataenialis). The focus is situated on the rostral medial wall within the projection of the medial half of MD. A very few cells were labeled f r o m injections outside this focus. Proline injections of MD revealed strong connections to caudate-putamen, olfactory tubercle and other regions. They will not be described further here. DISCUSSION The division of the rabbit M D complex into medial and lateral parts was based on a functional rather than morphological criterion. Single unit responses to electrical stimulation of the olfactory bulb were confined to the medial, approximately one half 10. In squirrel monkey the equivalent responsive area could be readily identified on cytoarchitectonic grounds as the magnoceUular division ~. Responsive units never violated the border between medial magnocellular and more lateral parvocetlular territory. In the opossum a clear correspondence was also demonstrated between functional and morphological characteristics. Essentially all the opossum MD is magnocellular7, ~6 and essentially all was responsive to olfactory tract stimulation 1~. Unfortunately, the rabbit M D complex is cyt oarchitectonically complex ~2 and no simple correlation could be made between a morphologically distinct area and olfactory responses. The rat M D has been described as totally magnocellular 14 and olfactory responses have been recorded 12,27. but their distribution has not been mapped m detail. It seems unlikely that the whole rat MD will prove responsive. Olfactory responses in the cat were described only as being mediaP °,16 and the relationships to morphology not reported. Because of the only occasional correlation with cytoarchitectonic fields, the functional division of M D into olfactory vs. non-olfactory zones seems to be the most satisfactory, if not the only, division tbr comparative purposes at present. Undoubtedly, more meaningful functional and anatomical parcellations will evolve. The first demonstration of anatomical connections between olfactory structures and M D by Powell et at. in t96517 has been confirmed in several subsequent studies, all on the rat 9,1z,~5,2z. Prepyriform cortex and olfactory tubercle distribute afferents to the central part of rat MD. not to the lateral segment, but also not to the most medial segment. Whether this most medial segment of rat M D would exhibit olfactory responses like the rabbit and squirrel monkey is not known. If so, they must be mediated byother olfactory structures or perhaps f r o m unexplored parts of prepyriform cortex or tubercle. If this medial segment is not responsive, that rat M D must be constructed from a different basic plan than that for the rabbit and squirrel monkey. Previous studies limited the cortical projection of M D in the rabbit to a cone of cortex capping the frontal pole 20. a disturbingly small fraction of the total projection established in this study. The explanation for this considerable discrepancy undoubtedly lies in the relative insensitivity of the retrograde chromatolysis method, the
263 only method used previously. The discrepancy is even more striking in the rat where frontal pole lesions failed to elicit any detectable thalamic chromatolysis 14, yet Leonard's analysis of Fink-Heimer material following electrolytic lesions of MD demonstrated projections to the entire medial wall rostral to the genu of the corpus callosum as well as to cortex on the dorsal lip of the rhinal sulcusH, l'~. Other studies using Fink-Heimer s, H R P 4 and autoradiographic tracer techniques 13 have confirmed these dual projections, have extended the caudal boundary on the medial wall to the rostral border of cingulate and retrosplenial cortex ~,s and have established that the sulcal projection does not invade the cortical taste area is as previously reported 1~. Just how closely the topographical organization of the rabbit thalamocortical projections corresponds to the organization in the rat is difficult to assess. In one respect, at least, they fit quite well. The lateral half of rabbit MD projects only to the medial wall exclusive of the most rostral part which along with sulcal cortex is territory belonging to the medial half of MD (Fig. 1A). In the rat Beckstead 4 noted that H R P injections in medial cortex resulted in stained cells confined to the lateral one-third or one-half of M D. Since his injections did not invade the most rostral medial cortex which in the rabbit receives projections from medial MD, the results are quite compatible. Beckstead also stated that sulcal injections were correlated with medial M D effects and Leonard 14 mentioned that one posterior lesion limited to the dorsomedial portion of MD produced degeneration limited to sulcal cortex. These statements indicate that the M-L dimension of MD is equivalently represented in the cortex of rat and rabbit. Much of the rabbit M D cortical projection on the medial wall is shared by other thalamic nuclei (Fig. 7). The entire field of lateral MD may be overlapped by AM projections as in the rat 4,s. Heavy projections were apparent in H R P material from all except the ventral and posteromedial portions of AM which project only sparsely, if at all, to rostral and dorsal medial cortex as discussed in the results section. The supracallosal portion of the lateral MD field, the region of the supplementary motor area (Fig. I B), is supplied by a moderate number of VA cells. The field of the medial, olfactory half of MD on the rostral medial wall receives strong input from all parts of PT. The paradigm for the MD projection system is the macaque system analyzed by countless retrograde chromatolysis studies 1. Three divisions of MD are generally recognized, each cytoarchitectonically distinct, each with its own cortical projection and each related to different functions. The medial magnocellular division projects to the orbital surface and is related to olfaction2,3,6,z4, 25. The more lateral parvocellular division projects to the frontolateral convexity and has always been overburdened with delayed response type behavior. The most lateral paralamellar division projects to the frontal eye fields and is involved with visual function. At present it is reasonable to equate the medial, magnocellular division of the monkey M D with the medial, olfactory part of the rabbit M D and, therefore, the frontal orbital cortex of the monkey with the sulcal cortex of the rabbit. Further comparisons might well be postponed until additional data is available on both species.
264
ACKNOWLEDGEMENTS S u p p o r t e d by G r a n t s NS-12721, NS-00782 a n d NS-07026. We wish to express a p p r e c i a t i o n to m e m b e r s o f o u r h i st o l o g y l a b o r a t o r y ( J o A n n Ekleberry, C a t h y G o o c h , J o a n Meister, A n n Switzky, a n d especially Inge S i g g e l k o w ) for their usual excellent product. C a r o l D i z a c k t r a n s f o r m e d o u r crude drawings into the final illustrations which were p h o t o g r a p h e d by Mr. T. P. Stewart,
REFERENCES 1 Akert, K., Comparative anatomy of frontal cortex and thalamofrontal connections. In J. M. Warren and K. Akert (Eds.), The Frontal Granular Cortex and Behavior, McGraw-Hill, New York. 1964, pp. 372-396. 2 Allen, W. F., Effect of ablating the frontal lobes, hippocampi, and occipoto-parieto-temporal (excepting pyriform areas) lobes on positive and negative olfactory conditioned reflexes. Amer. J. PhysioL, 128 (1940) 754-771. 3 Allen, W. F., Results of prefrontal lobectomy on acquired and acquiring correct conditioned differential responses with auditory, general cutaneous and optic stimuli, Amer..I. PhysioL, 139 (1943) 525-531. 4 Beckstead, R. M., Convergent thalamic and mesencephalic projections to the anterior medial cortex in the rat. J. comp. NeuroL, 166 (1976) 403~,t6. 5 Benjamin, R. M. and Akert, K., Cortical and thalamic areas involved in taste discrimination in albino rat, J. comp. NeuroL, 111 (1959) 231-260. 6 Benjamin, R. M and Jackson, J. C., Units discharges in the mediodorsal nucleus of the squirrel monkey evoked by electrical stimulation of the olfactory bulb, Brain Research. 75 (1974) 181-191. 7 Bodian, D., Studies on the diencephalon of the Virginia opossum. Part I: the nuclear pattern in the adult, J. comp. Neurol., 71 (1939) 259-323. 8 Domesick, V. B., Thalamic relationships of the medial cortex in the rat, Brain Behav. Evol.. 6 (1972) 457483. 9 Heimer, L., The olfactory connections of the diencephalon in the rat. Brain Behav. Evol.. 6 (1972) 484-523. 10 Jackson, J. C. and Benjamin, R. M., Unit discharges in the mediodorsal nucleus of the rabbit evoked by electrical stimulation of the olfactory bulb, Brain Research, 75 (1974) 193-201. 11 Jackson, J. C., Golden, G. T. and Benjamin, R. M.. The distribution of olfactory input in the opossum mediodorsal nucleus, Brain Research, submitted for publication. 12 Komisaruk, B. R. and Beyer, C., Responses of diencephalic neurons to olfactory bulb stimulation. odor and arousal, Brain Research, 36 (1972) 153-170. 13 Krettek, J. E. and Price, J. L.. A direct input from the amygdala to the thalamus and the cerebral cortex, Brain Research. 67 (1974) 269-274. 14 Leonard, C. M., The prefrontal cortex of the rat. 1. Cortical projection of the mediodorsal nucleus. II. Efferent connections, Brain Research, 12 (1969) 321-343. 15 Leonard, C. M., The connections of the dorsomedial nuclei, Brain Behav. Evol., 6 (1972) 524-54l. 16 Motokizawa, F., Olfactory input to the thalamus: electrophyslological evidence, Brain Research. 67 (1974) 334-337. 17 Powell, T. P. S., Cowan, W. M. and Raisman, G., The central olfactory connections, J. Anat. (Lond.), 99 (1965) 791-813. 18 Price, J. L., Personal communication. 19 Rose, J. E. and Woolsey, C. N., Structure and relations of limbic cortex and anterior thalamic nuclei in rabbit and cat, J. comp. Neurol., 89 (1948) 279-340. 20 Rose, J. E. and Woolsey, C. N., The orbitofrontal cortex and its connections with the mediodorsal nucleus in rabbit, sheep and cat, Res. Publ. Ass. herr. merit. Dis., 27 (1948) 210-232. 21 Rose, M., Zytoarchitekt. Atlas der Grosshirnrinde des Kaninchens. J. Psychol. Neurol. (Lpz.), 43 (1931) 353-340.
265 22 Rose, M., Das Zwischenhirn des Kaninchens, Extrait des M~m. Acad. pol. Sci., Serie B, 1935, pp. 1 108. 23 Scott, J. W. and Leonard, C. M., The olfactory connections of the lateral hypothalamus in the rat, mouse and hamster, J. comp. Neurol., 141 (1971) 331-344. 24 Tanabe, T., Ilno, M. and Takogi, S. F., Distribution of odors in the olfactory bulb, pyrifGrmamygdaloid areas, and orbitofrontal cortex of the monkey, J. Neurophysiol., 38 (1975) 1284-1296. 25 Tanabe, T., Yarita, H., lino, M., Ooshima, Y. and Takogi, S. F., An olfactory projection area in orbitofrontal cortex of the monkey, J. Neurophysiol., 38 (1975) 1283-1296. 26 Tobias, T. J. and Ebner, F. F., Thalamocortical projections from the mediodorsal nucleus in the Virginia opossum, Brain Research, 52 (1973) 79-96. 27 Wedgwood, M., Connexions between olfactory bulb and habenula and the dorsomedial thalamic nuclei, J. Physiol. (Lond.), 239 (1974) 88-89. 28 Woolsey, C. N., Organization of somatic sensory and motor areas in the cortex. In H. F. Harlow and C. N. Woolsey (Eds.), Biological and Biochemical Bases t~/"Behavior, Univ. of Wisconsin Press, Madison, 1958, pp. 63 81. 29 Yamamoto, T. and Kawamura, Y., Cortical responses to electrical and gustatory stimuli in the rabbit, Brain Research, 94 (1975) 447-463.
