Corticocortical communication via the thalamus: Ultrastructural studies of corticothalamic projections from area 17 to the lateral posterior nucleus of the cat and inferior pulvinar

Corticocortical Communication Viathe Thalamus: Ultrastructural Studiesof Corticothalamic Projections From

Area 17 to the Lateral Posterior Nucleusof the Cat and Inferior Pulvinar

Nucleus of the Owl Monkey

SHERRY FEIG* AND JOHN K. HARTING

Department of Anatomy, University of Wisconsin Medical School, Madison, Wisconsin 53706

ABSTRACTElectron microscopic anterograde autoradiography has been used to analyze the morphol-

ogy and postsynaptic relationships of area 17 cortical terminals in the lateral division of thelateral posterior nucleus (LPl) of the cat and medial division of the inferior pulvinar nucleus(IPm) of the owl monkey. Such terminals are thought to arise exclusively from layer 5 in thecat and primate (Lund et al. [1975] J. Comp. Neurol. 164:287–304; Abramson and Chalupa[1985] Neuroscience 15:81–95). All labeled terminals in both nuclei exhibited the morphologyof ascending ‘‘lemniscal’’ afferents. That is, they contained round vesicles, were large, madeasymmetrical synaptic and filamentous nonsynaptic contacts, and were classified as RLs.These cortical RLs also exhibited the postsynaptic relationships of lemniscal afferents. Thus,they were presynaptic to large dendrites within glial encapsulated glomeruli, where amajority was involved in complex synaptic arrangements called triads. They also were foundadjacent to terminal profiles with pleomorphic vesicles but never adjacent to small terminalscontaining round vesicles.

Our results suggest that the layer 5 projection from area 17 provides a functional ‘‘drive’’for some LPl and IPm neurons. Information carried over this ‘‘re-entrant’’ pathway (Guillery[1995] J. Anat. 187:583–592) could be modified within the LPl and IPm by both cortical andsubcortical pathways and subsequently conveyed to higher visual cortical areas, where itcould be integrated with messages carried through the well-documented corticocorticalpathways (Casagrande and Kaas [1994] Cerebral cortex New York: Plenum Press). J. Comp.Neurol. 395:281–295, 1998. r 1998 Wiley-Liss, Inc.

Indexing terms: cortical layer 5; thalamocortical; modulation

Visual, auditory, and somatosensory information reachesthe thalamus via lemniscal afferents. Such afferents,which terminate within primary or first-order (Guillery,1995) thalamic nuclei like the lateral geniculate nucleus(LGN), the medial geniculate nucleus, and the ventrobasalcomplex, form characteristic large synaptic terminals onstem dendrites relatively close to the cell bodies of thalamo-cortical neurons (Guillery, 1967, 1969, 1971; Famigliettiand Peters, 1972; Wong-Riley, 1972; Wilson and Hendrick-son, 1981; Feig and Harting, 1994). These lemniscalterminals, which are large and contain round vesicles(RL), provide the functional ‘‘drive’’ for the sensory evokedproperties of cells within first-order nuclei (Hoffman et al.,1972; So and Shapley, 1979; Sur et al., 1982).

Ascending lemniscal afferents exhibit specific spatialrelationships on thalamocortical neurons with afferentsfrom layer 6 of cortex and the thalamic reticular nucleus(TRN; see Sherman and Guillery, 1996). Terminals associ-ated with the layer 6 cortical projection, which are smalland contain round vesicles (RS), distribute on distal seg-ments of thalamic dendrites (Guillery, 1969; Wilson et al.,

Grant sponsor: National Eye Institute; Grant number: EY-01277.*Correspondence to: Sherry Feig, Department of Anatomy, University of

Wisconsin—Madison, 1300 University Avenue, Madison, WI 53706.E-mail: [email protected]

Received 6 November 1997; Revised 26 January 1998; Accepted 29January 1998

THE JOURNAL OF COMPARATIVE NEUROLOGY 395:281–295 (1998)

r 1998 WILEY-LISS, INC.

1984; Weber and Kalil, 1987; Feig and Harting, 1994).TRN terminals, which are inhibitory and are called Fprofiles because of the flattened shape of their vesicles, lieadjacent to RL profiles on thalamic dendrites (Harting etal., 1991). These latter two pathways are thought tomodulate the transmission of ascending sensory informa-tion through first-order thalamic nuclei (McCormick andBal, 1994; Sherman and Guillery, 1996; Guillery et al.,1998).

In contrast to primary or first-order thalamic nuclei,‘‘association’’ or higher order thalamic nuclei (Guillery,1995) are not targeted by ascending primary afferents.However, a novel concept is beginning to emerge whichsuggests that these nuclei receive descending primary(RL) afferents from the cerebral cortex and in particularlayer 5 (Mathers, 1972b; Robson and Hall, 1977; Guillery,1995; Sherman and Guillery, 1996). That is, whereas RLsin a first-order nucleus such as the LGN provide the mainsource of visual information destined for area 17, RLs inassociation or higher order nuclei would provide detailsregarding the ongoing cortical activity in layer 5 pyrami-dal cells. Such information would then be conveyed tohigher cortical areas via thalamocortical pathways.

The mammalian lateral posterior/pulvinar complex isusually ascribed an associative function because of itsextensive interconnections with visual cortical areas (seeLin and Kaas, 1979; Updyke, 1983). Moreover, the cortico-thalamic projection from area 17 to the lateral posteriornucleus of the cat (LP) and the inferior pulvinar nucleus ofthe primate is thought to arise exclusively from layer 5(Lund et al., 1975; Abramson and Chalupa, 1985). We havetherefore hypothesized that terminals associated witharea 17 projections to both the lateral division of the LP(LPl) of the cat and the medial division of the inferiorpulvinar (IPm) of the owl monkey exhibit the morphologyof RLs. Moreover, we also hypothesize that the spatialrelationships of these cortical RLs with RS and F profiles

on thalamocortical neurons are the same as those exhib-ited by RLs in first-order nuclei. We used electron micro-scopic anterograde autoradiography to support these hy-potheses and discuss the functional implications of thesefindings.

MATERIALS AND METHODS

All animal care and methodological procedures wereapproved by the University of Wisconsin Resources Centerand conformed to the requirements of the NIH Guide forthe Care and Use of Laboratory Animals (Publication85–23) and the US Department of Agriculture AnimalWelfare Act.

Two adult cats and two adult owl monkeys (Aotustrivirgatus) were used in the present analysis. Beforesurgery, each cat was anesthetized with an intramuscularinjection of 0.5 ml ketamine hydrochloride (Ketalar, 100mg/ml), followed by an intravenous injection of sodiumpentobarbital (25–30 mg/kg). Each owl monkey was anes-thetized with intramuscular injections of ketamine (30–40mg/kg) and Xylazine (6 mg/kg). After attaining a state ofdeep anesthesia, which was confirmed by the absence of awithdrawal reflex, the cat or owl monkey was placed in aKopf stereotaxic apparatus. A midline incision was madein the posterior scalp region, after which the scalp andmuscle overlying the occipital lobes were retracted. Asmall hole was made in the skull overlying area 17, and asmall slit was made in the dura matter. A 10-µl Hamiltonsyringe containing an equal mixture of [3H]proline and[3H]leucine in bacteriostatic saline (50 mCi/ml) was thenlowered, via a Kopf manipulator, into primary visualcortex (area 17). In each owl monkey, six small injections(0.5 µl each) were placed adjacent to each other in thecaudal and medial parts of area 17 (Allman and Kaas,1971). In each cat, four small cortical injections (0.25 µleach) were placed into the caudal and medial portions of

Fig. 1. Line drawings show the distribution of transported proteinwithin the lateral division of the lateral posterior nucleus (LPl) of thecat (A) and the medial division of the inferior pulvinar nucleus (IPm) ofthe owl monkey (B) after injections of [3H] amino acids in area 17. D,dorsal; Hab, habenular nucleus; IPc, caudal division of the inferior

pulvinar nucleus; IPp, posterior division of the inferior pulvinarnucleus; L, lateral; LGN, lateral geniculate nucleus; LPm, medialdivision of the lateral posterior nucleus; M, medial; P, pulvinarnucleus; SP, superior pulvinar nucleus; V, ventral. Scale bars 5 1 mm.

282 S. FEIG AND J.K. HARTING

Fig. 2. Electron micrograph shows examples of the various labeledand unlabeled synaptic profiles within the lateral posterior nucleus(LPl) of cat (included in Table 1). There are two terminal classes thatcontain round vesicles and make asymmetrical contacts. Those thatare larger and irregularly shaped with round vesicles are called RL,and the smaller crescent-shaped terminals with round vesicles arecalled RS. The labeled RL profile in the top part of the micrograph ispresynaptic to both a dendrite (D) and a pale vesicle-filled (PVF)

profile. RL terminals frequently make nonsynaptic filamentous con-tacts with dendrites (asterisks) and contain multiple mitochondriaand occasional dense core vesicles (diamonds). RS profiles are presyn-aptic (arrowhead) to dendritic profiles (D). Profiles that containuniformly distributed flat and pleomorphic vesicles and make sym-metrical synaptic contacts are called F. Vesicle-filled boutons that arenonsynaptic are called B. Arrowheads denote synaptic contacts, andsmall arrows denote glial lamellae. Scale bar 5 1 µm.

Fig. 3. Electron micrograph showing examples of the varioussynaptic profiles within the inferior pulvinar nucleus (IPm) of the owlmonkey. Two terminal classes contain round vesicles and makeasymmetrical contacts: One larger and irregularly shaped (RL), andone small and crescent-shaped (RS). The labeled RL profile is presyn-aptic to a dendrite (D) and a pale vesicle-filled (PVF) profile. This RLterminal has an irregular shape and contains mitochondria and

numerous dense core vesicles (diamonds). The profile with uniformlydistributed pleomorphic and flat vesicles is an example of F terminals.The RL, F, and PVF profiles make filamentous contacts (asterisk) on adendrite (D). Nonsynaptic vesicle-filled boutons (B) are in the field.Arrowheads denote synapses, and small arrows denote glial lamellae.Scale bar 5 1 µm.

area 17 (Tusa et al., 1978). After the injections, the syringewas withdrawn and the hole in the skull was sealed withsterile bone wax. Surgical thread was then used to suturethe scalp wound, and a topical pain killer (Zylocaine) wasapplied to the wound area.

All animals were given fluids and antibiotics and weremonitored closely during their recovery from anesthesia.The cats and owl monkeys were observed several timeseach day by the veterinarian staff and appropriate fluids,analgesics, antibiotics, and topical anesthetics were usedas needed. After survival periods of 24 hours, each animalwas given an overdose of sodium pentobarbital (80–100mg/kg Nembutal) and perfused transcardially with 0.9%saline, followed by 1.0% paraformaldehyde/1.0% glutaral-dehyde in 0.1 M sodium phosphate buffer, pH 7.4, and 2.0%paraformaldehyde/2.0% glutaraldehyde in 0.1 M sodiumphosphate buffer.

After each perfusion, the brain was left in the skull for 1hour at 4°C, blocked in a frontal plane, removed, andplaced overnight in the first fixative (1% paraformalde-hyde/1% glutaraldehyde). Blocks of thalamic tissue contain-ing the LPl and IPm were then removed and seriallysectioned on a Vibratome at alternating thicknesses of 50µm (for light microscopy) and 100 µm (for electron micros-copy). The sections were then placed into 0.1 M sodiumphosphate buffer. To define the boundaries of the corticalinjection sites, area 17 was sectioned at 40 µm on afreezing microtome. The 50-µm thalamic sections and thecortical sections were mounted on gelatin-coated slidesand processed according to standard autoradiographicprocedures (Cowan et al., 1972).

The protocol used to process the thalamic (LPl and IPm)tissue for electron microscopic autoradiography has beendescribed in previous studies (Harting et al., 1991, 1997;Feig and Harting, 1992, 1994; Feig et al., 1992). Briefly,the 100-µm sections for electron microscopic autoradiogra-phy were postfixed in buffered 2% osmium tetroxide for 2hours, dehydrated in graded alcohols, and embedded be-tween aclar sheets in Epon-Araldite. The light microscopic(50 µm) sections were examined under darkfield illumina-tion to locate the transported protein within the LPl andIPm. This information allowed us to choose the mostappropriate region of the adjacent 100-µm sections forremounting and thin sectioning. Figure 1 shows the distri-bution of transported protein within the LPl (Fig. 1A) andIPm (Fig. 1B) after injections into all layers of area 17 ofcat 1693 and owl monkey 1710, respectively.

Silver thin sections (70 nm) from the remounted sectionswere then placed on parlodian-coated glass slides anddipped in Ilford L-4 photographic emulsion by using asemiautomatic dipping machine (Bachmann and Salpeter,1965; Kopriwa, 1973). After appropriate exposure periods(16–20 weeks) at 4°C, the slides were developed in D-19 for2 minutes. The parlodian with the sections was thenfloated off the slide, and 200 fine-bar mesh copper gridswere placed over the sections. The sections were thenstained with LKB Ultrastain 1 (uranyl acetate solution)and LKB Ultrastain 2 (lead citrate solution) in the LKBultrastainer.

For the autoradiographic analyses, every grain in 13,000µm2 (for each animal) was photographed at 10,0003 with aJEOL 100CX electron microscope. The negatives wereprinted at 2.53 for analysis. The grain distribution wasdetermined by placing a probability circle, with 50%confidence of including the source of radioactivity, over

each grain and scoring the underlying profile (Bachmannand Salpeter, 1965; Kopriwa, 1967). In most instances, thegrains were clearly over one profile, that is, the circle wascompletely within that profile. In some instances, morethan one profile was within the circle. In these cases, thefraction of the grain that was allotted to each profile wasproportional to the fraction of the circle that was filled bythat profile.

For each electron micrograph, the percentage of totalgrains over the following neural profiles was determined:large irregularly shaped terminal profiles with roundvesicles and making asymmetrical synaptic contacts (RL),small crescent-shaped terminal profiles containing roundvesicles and making asymmetrical synaptic contacts (RS),terminal profiles containing pleomorphic and flat vesiclesand making symmetrical synaptic contacts (F), and palevesicle-filled profiles (PVF) that make symmetrical synap-tic contacts and/or are themselves postsynaptic, nonsynap-tic vesicle-filled boutons (B), and dendrites (D; dendriticspines are included with dendrites). Examples of each ofthese profile categories in the LPl and IPm are shown inFigures 2 and 3, respectively. Myelinated axons, axoncollaterals, glial elements, and all unidentified profileswere placed in the ‘‘other’’ category.

Tissue from two cats and two owl monkeys was analyzedfor each projection. The mean percentage of grain distribu-tions for each projection with standard errors are listed inTable 1. The percentage of the total area occupied by theseprofiles in one LPl and one IPm was then determined in13,000 µm2. Each profile was traced by using a JandelScientific (Corte Madera, CA) bit pad data tablet linked toa computer. Sigma Scan (Jandel Scientific) software wasthen used to calculate the areas of these profiles. Therelative density analysis (Ross and Benditt, 1965; Hen-drickson, 1972) of the distribution of silver grains was thencarried out for the various categories by dividing thepercentage of total grains over each profile for each animalby the percentage of the total area occupied by each profilein the LPl or IPm (Table 1).

The area of labeled corticothalamic terminal profileswas calculated as described above, and the results areshown in Figure 4. The number and type (i.e., dendrite,PVF, unidentified) of individual profiles postsynaptic toeach labeled terminal profile (Figs. 5, 6) were scored. Onlysynaptic contacts that showed clear membrane specializa-tions and a clustering of vesicles at the synaptic zone werescored (e.g., in Fig. 6C the F profile would not be scored,but in Fig. 6D the F profile would be scored as presynap-tic). In the cases where a terminal profile exhibited more

TABLE 1. Distribution of Label in the Lateral Division of the LateralPosterior Nucleus (LPl) and Inferior Pulvinar Nucleus (IPm) After

an Injection of [3H] Amino Acids in Area 171

Profile2

% Grain distribution % Area % Density

LPl IPm LPl IPm LPl IPm

RL 63.0 6 4.0 57.0 6 3.0 4 3 16 6 1.0 19.0 6 1.0RS 3 6 0.2 2 6 0.1 9 7 0.3 6 0.03 0.3 6 0.02B 5 6 0.3 7 6 0.3 7 8 0.7 6 0.01 0.9 6 0.08F 0.4 6 0.1 0.2 6 0.1 1 1 0.4 6 0.05 0.2 6 0.01PVF 3 6 0.6 4 6 0.3 3 4 1 6 0.2 1 6 0.01D 5 6 0.7 5 6 0.4 23 24 0.2 6 0.001 0.2 6 0.001Other 21.6 6 0.2 23.8 6 0.1 44 53 0.5 6 0.03 0.4 6 0.02

1Data taken from two animals (mean 6 S.E.M.).2RL, large terminal with round vesicles; RS, small terminal with round vesicles; B,nonsynaptic vesicle filled bouton; F, terminal with pleomorphic and flat vesicles; PVF,pale vesicle-filled profile; D, dendrite or dendritic spine; other, all other or unidentifiedprofiles.

CORTICOTHALAMIC PROJECTIONS TO THE VISUAL THALAMUS 285

than one synaptic contact on the same profile, the targetwas scored as one. The number of each postsynaptic profiletargeted by the labeled terminals was divided by the totalnumber of labeled terminals analyzed for each animal; themean and standard errors are shown in Figure 7A.

The percentage of RL terminal profiles involved intriadic arrangements was also analyzed (Fig. 7B). Whentwo of the three synaptic relationships were observed inone micrograph and the plasma membranes of all threeprofiles (labeled cortical terminal, dendrite, PVF) weredirectly apposed to one another, they were scored as atriad. Examples are shown in Figures 2, 5B, 6A,B.

We also wanted to compare several features of thelabeled cortical RL profiles (LPl, n 5 98; IPm, n 5 97) withRS profiles within the LPl and IPm. In two animals, theareas of RS profiles in the LPl (n 5 50) and IPm (n 5 50)were calculated (Fig. 4). To compare the synaptic organiza-tion of labeled corticothalamic terminals with that of theRS profile sample, the postsynaptic targets and all presyn-aptic terminal profiles adjacent to RS terminals on thesetargets were analyzed (Fig. 9A,B). The diameter of den-drites postsynaptic to the cortical RL terminal profileswere compared with those postsynaptic to RS terminalprofiles (Table 2). The smallest diameter measurement ofdendritic profiles postsynaptic to these terminal profileswere scored.

RESULTS

Ultrastructural characteristics of synapticprofiles within the LPl and IPm

The ultrastructural features of the various synapticprofiles within the LPl of the cat and the IPm of the owl

monkey were similar and are described together. Fourgeneral categories of presynaptic terminals have beenidentified on the basis of vesicle shape, packing density,postsynaptic specialization, and, in some cases, terminalsize. Figures 2 and 3 show examples of the four categoriesof terminal profiles within the LPl (cat) and IPm (owlmonkey), respectively. Profiles in the first category, desig-nated RL, contained round vesicles and were large andirregularly shaped. RL terminals frequently formed mul-tiple nonsynaptic puncta adhaerentia, termed filamentouscontacts (Guillery, 1967, 1969; indicated by asterisk inFigs. 2, 3), and had a central cluster of mitochondria anddense core vesicles. These terminals were associated withan asymmetric postsynaptic density and were often presyn-aptic to several profiles. Glial lamellae usually surroundedportions of the perimeter of the RL complex. This glial‘‘encapsulated’’ RL complex is termed a glomerulus (Guil-lery, 1969; Guillery and Colonnier, 1969).

The second category of profiles in the LPl and IPm isdesignated RS. Terminals in this category also containedround vesicles, but they were smaller than RLs and

Fig. 4. Histograms show the distributions of sizes for labeled area17 RL profiles from area 17 (lateral posterior nucleus [LPl], n 5 98;inferior pulvinar nucleus [IPm], n 5 97) and a random sample of RSprofiles in the LPl (n 5 50) and the IPm (n 5 50). The mean area and

standard error are given for RL and RS terminal profiles. Note thelarge range in sizes for area 17 terminal profiles (RLs) and thesingular size distribution for RSs.

Fig. 5. Electron micrographs show autoradiographically labeledterminals within the lateral posterior nucleus (LPl) of the cat after aninjection of [3H] amino acids into area 17. A: Labeled cortical terminal(C) that is presynaptic to a dendrite (D), which receives a synapse fromboth a pale vesicle-filled (PVF) and an F profile. B: Labeled corticalterminal (C) that is presynaptic to a dendrite (D), which is alsopostsynaptic to a PVF profile. The plasma lamellae of the cortical andPVF profile are directly apposed to one another. An F profile can beseen making a nonsynaptic filamentous contact (asterisk) with thesame dendrite. Arrowheads indicate synaptic contacts, small arrowsdenote glial lamellae, and diamonds indicate dense core vesicles. Scalebars 5 1 µm.


Figure 5

Figure 6


exhibited a crescent-shaped contour (Figs. 2, 3). RSs wereusually presynaptic to only one profile and exhibitedasymmetrical synaptic specializations. These terminalprofiles rarely contained mitochondria but occasionallyexhibited a dense core vesicle.

In addition to the two categories of terminals thatcontained round vesicles and made asymmetric synapses(RLs and RSs), a third category of profiles containeddensely packed pleomorphic vesicles, some of which weredistinctly flat. These were termed F profiles (Figs. 2, 3).Such profiles were sometimes presynaptic to more thanone profile and were associated with symmetric postsynap-tic specializations. In addition, F profiles sometimes formednonsynaptic filamentous contacts. All three of these termi-

Fig. 6. Examples of labeled corticothalamic terminals within theinferior pulvinar nucleus (IPm) of the owl monkey after an injection of[3H] amino acids into area 17. A: Labeled terminal that has anirregular shape and is presynaptic to two dendritic profiles (D). One ofthese dendrites is also postsynaptic to a pale vesicle-filled (PVF)profile. The cortical terminal contains a dense core vesicle (diamond)and makes a nonsynaptic filamentous contact (asterisk) with thelarger dendrite. The portion of the cortical terminal that is presynaptic(arrowhead with asterisk) to the smaller dendritic profile is similar inappearance to an RS profile but is continuous with the larger portion ofthe terminal profile. There is also a labeled nonsynaptic bouton in thefield (B) that has a dense core vesicle (diamond). B: Labeled corticalprofile that is presynaptic to both a dendrite (D) and a PVF profile andmakes a nonsynaptic filamentous contact (asterisk) with the samedendrite. There is an F profile that is presynaptic to a dendrite, whichis directly apposed, but not synaptic with the labeled cortical profile inthis plane of section. C: An example of a labeled cortical terminal thatmakes both a synaptic and a nonsynaptic filamentous (asterisk)contact with a dendrite (D). An F profile probably makes a synapticcontact with this same dendrite (curved arrow) because there is aclustering of presynaptic vesicles along the darkened region of thetangentially cut plasma membranes. D: An example of a smallerlabeled cortical terminal and an F profile, which are presynaptic to thesame dendritic profile (D) profile. There is a PVF profile adjacent to thelabeled cortical terminal. Arrowheads denote synaptic contacts andsmall arrows denote glial lamellae. Scale bars 5 1 µm.

Fig. 7. A: Histogram showing the percentage of labeled corticalterminal profiles in the lateral posterior nucleus (LPl; two cats) andthe inferior pulvinar nucleus (IPm; two owl monkeys) that arepresynaptic to dendrites and/or pale vesicle-filled (PVF) profiles.B: Histograms showing the percentage of the labeled cortical terminalprofiles that are involved in triad-like synaptic relationships (i.e., two

of the three synaptic contacts are present and the plasma lamellae ofthe three profiles are directly apposed to one another; examples areshown in Figs. 2, 5B, 6A,B). Approximately 75% of the labeledterminal profiles from area 17 in both the LPl and the IPm areinvolved in triad-like arrangements.

TABLE 2. Average Diameter (mean 6 S.E.M.) of DendritesPostsynaptic to Cortical (Area 17) RL Profiles or RS Profiles

Within the Lateral Division of the Lateral Posterior Nucleus (LPl)and Inferior Pulvinar Nucleus (IPm)1

Nucleus RL RS P*

LPl (n) 1.60 6 0.053 (50) 0.77 6 0.077 (26) ,0.001IPm (n) 1.27 6 0.086 (25) 0.65 6 0.041 (26) ,0.001

1Data taken from two animals.*Student’s t-test (two-tailed) comparing the means of these samples.


nal classes in the LPl and the IPm were presynaptic todendrites and dendritic spines. Cortical terminal profileswere never observed presynaptic to cell bodies in either thecat or owl monkey.

The three terminal classes just described (RLs, RSs, andFs) were exclusively presynaptic. The fourth type of synap-tic terminal was both pre- and postsynaptic. This profileexhibited a pale cytoplasmic matrix (when compared withthe other three terminal classes) and contained looselypacked pleomorphic vesicles. This terminal class has beentermed a pale, vesicle-filled profile (PVF; Figs. 2, 3). ThePVF profile was accompanied by a symmetrical postsynap-tic specialization and sometimes formed nonsynaptic fila-mentous contacts. Such profiles, sometimes termed F2 orP, have been shown to be dendritic appendages of thalamicinterneurons (Ogren and Hendrickson, 1979a; Wilson andHendrickson, 1981; Wilson et al., 1984; Hamos et al.,1985).

RL, PVF, and dendritic profiles were frequently involvedin complex synaptic arrangements within the LPl and theIPm (Figs. 2, 5B, 6A,B). The synaptic arrangement of anRL terminal that is presynaptic to both a dendrite and aPVF profile, which is also presynaptic to the same den-drite, has been termed a triad within the pulvinar (Mathers,1972a; Robson and Hall, 1977; Ogren and Hendrickson,1979a). In contrast to RLs, RS profiles were almost exclu-sively presynaptic to one dendrite and did not participatein complex synaptic arrangements.

Distribution and morphology of labeledcortical (area 17) terminals in LPl and IPm

After injections of a mixture of [3H]proline and [3H]leu-cine into area 17 of two cats and two owl monkeys,anterogradely transported protein was observed withinthe LPl and IPm, respectively (Fig. 1A,B). All of ourultrastructural data were taken from these striate (area17) recipient zones of the LPl and IPm. Table 1 shows theresults from the relative density analysis of the ultrastruc-tural distribution of silver grains (Ross and Benditt, 1965;Hendrickson, 1972). In both the LPl and IPm, the large RLprofiles were the most heavily labeled. An analysis of thecortical profiles showed that they had large areas, wereirregularly shaped, and exhibited a wide range of sizes(Figs. 2–6). Figure 6A shows an example of an irregularlyshaped cortical terminal profile with a smaller boutonregion that is presynaptic to a dendrite; another labelednonsynaptic bouton is also in the field.

Synaptic organization of cortical (area 17)terminals in the LPl and IPm

Almost every labeled cortico-LPl and -IPm terminalprofile was presynaptic to a dendrite (Fig. 7A). Approxi-mately 40% of these terminals were also presynaptic toPVF profiles (Fig. 7A). If the PVF profile was in turnpresynaptic to the same dendrite, it was called a triad.Although this complex relationship has been observed inthe pulvinar (Mathers, 1972a; Robson and Hall, 1977;Ogren and Hendrickson, 1979a), in our material it wasrare to see all three synapses in the same section. There-fore, if two of the three contacts were present and theplasma membranes of these three profiles were directlyapposed to one another, they were scored as a triad (Figs.2, 5B, 6A,B). Approximately 70% of cortical terminals inthe LPl and IPm were involved in triads (Fig. 7B).

Sometimes other terminal profiles were presynaptic to adendrite, which was also postsynaptic to a cortical profile.An analysis of these terminal classes showed that labeledcortical terminals and F profiles made synapses on thesame postsynaptic dendrite in the LPl and IPm. Approxi-mately 35% of the labeled RLs exhibited this association(Figs. 5A, 6D, 8). In contrast, cortical RLs were neverobserved adjacent to RS terminal profiles on a postsynap-tic dendrite.

Comparison of the morphology and synapticorganization of RS profiles with cortical RL

(area 17) profiles

RS profiles in the LPl and IPm were unlabeled afterinjections of [3H] amino acids into area 17. Randomsamples of RS profiles in the LPl and IPm were analyzedfor comparison with cortical (RL) profiles. RS profiles wereuniform in shape and showed essentially no variation insize distribution (Figs. 2–4). RS profiles were significantlysmaller than cortical RL profiles (P , 0.0005). Figure 9Ashows that, like cortical RL terminals, the majority of theRS profiles were presynaptic to dendrites. The meandiameters of dendritic profiles postsynaptic to RL (cortical)profiles were significantly larger than those postsynaptic

Fig. 8. Histogram showing the percentage of labeled corticalterminal profiles that are presynaptic to dendrites that are alsopresynaptic to F profiles in lateral posterior nucleus (LPl; two cats)and inferior pulvinar nucleus (IPm; two owl monkeys; for examples,see Figs. 5A, 6D).


to RS profiles in both the LPl and IPm (Table 2). Unlikecortical RL terminals, few RSs were presynaptic to PVFprofiles (Fig. 9A) and none were observed in triadicarrangements in either the LPl or IPm.

The frequency with which other terminal populationswere adjacent to RS profiles on postsynaptic dendrites wasalso examined (Fig. 9B). In contrast to the cortical RLs,which do not lie adjacent to RSs, slightly more than half ofthe RS profiles were adjacent to other RSs on postsynapticdendrites in both the LPl and the IPm. Approximately 10%of the RS terminals were adjacent to F profiles in the LPlor IPm; this is a smaller percentage than that of thecortical profiles exhibited.

DISCUSSION

Our morphological findings suggest that some relay cellsin the LPl and IPm receive their functional drive from area17 and in particular from layer 5 pyramidal cells (Lund etal., 1975; Abramson and Chalupa, 1985). This idea issupported by physiological data showing that the responseproperties and receptive field characteristics of many LPlcells in the cat resemble those of layer 5 pyramidal cells inarea 17 (Casanova et al., 1989; Chalupa and Abramson,1989; Casanova and Savard, 1996) and that inactivation ofarea 17 in the macaque monkey results in the loss of the

receptive field properties of cells within the IP (Bender,1983). Taken together, these morphological and physiologi-cal findings suggest that the synaptic contacts that areestablished in the LPl and IPm by area 17 RLs are criticalin defining the functional role of these association orhigher order thalamic nuclei. Moreover, because the LPland IPm project to several higher cortical visual areas,these corticothalamic synapses may also play a crucial rolein defining the nature of the corticocortical communica-tions that pass through the thalamus.

Comparison of the morphology and spatialrelationships of area 17 cortico-LPl and -IPmterminals and retinogeniculate (lemniscal)

profiles with other terminal classes onthalamic neurons

Our findings show that area 17 cortico-LPl and -IPmterminals exhibit the same morphology and postsynapticrelationships as lemniscal afferents (retinal) within theLGN. Both the corticothalamic and the retinal terminalsare exclusively RLs. Such terminals contain round vesicles,are large, and make asymmetrical synaptic contacts onseveral profiles within glial encapsulated glomeruli; theyalso make nonsynaptic filamentous contacts (Guillery,1969; Mathers, 1972a,b; Robson and Hall, 1977; Ogren

Fig. 9. These two histograms show the synaptic relationships of arandom sample of RS terminals within the lateral posterior nucleus(LPl; n 5 50) and inferior pulvinar nucleus (IPm; n 5 50). A: Thepercentage of RS terminal profiles that are presynaptic to dendrites orpale vesicle-filled (PVF) profiles. Nearly all of the RS profiles are

presynaptic to a dendrite. B: The percentage of RS terminal profilesthat are presynaptic to dendrites that are also presynaptic to RS or Fprofiles. A little more than half of the RS profiles are associated withother RS terminal profiles on the same dendrite, and very few arespatially associated with F profiles.


and Hendrickson, 1979b; Feig and Harting, 1994). Botharea 17 and retinal RLs distribute to proximal (relativelylarge) dendrites and are not associated with RSs on moredistal (relatively small) dendrites. In the LGN, many of theRS profiles arise from layer 6 cells in area 17 (Weber andKalil, 1987; Feig and Harting, 1994). In the LPl and IPm,many of these profiles arise from layer 6 cells withinhigher cortical areas (Mathers, 1972b; Robson and Hall,1977; Abramson and Chalupa, 1985). They frequently lieadjacent to other RSs on the same dendritic profile butwere never observed in glomeruli or in triadic arrange-ments. Collectively, these data suggest that, as in theLGN, RL and RS terminals in the LPl and IPm distributeto different regions of the postsynaptic neuron.

The present findings also show that the relationship ofcortico-LPl and -IPm terminals with inhibitory processesis similar to those of retinal RLs. Thus, both area 17 RLs inthe pulvinar (Mathers, 1972b; Robson and Hall, 1977) andretinal RLs in the LGN (Guillery, 1969; Wilson et al., 1984;Feig and Harting, 1994) participate in a characteristictrisynaptic arrangement with large dendrites and PVFs,which are dendritic appendages of interneurons (Ogrenand Hendrickson, 1979a; Wilson et al., 1984). This trisyn-aptic arrangement has been called a triad (Mathers,1972a; Robson and Hall, 1977; Rapisardi and Miles, 1984;Hamos et al., 1985; Mize et al., 1986; Takacs et al., 1991).The present data show that approximately 70% of the area17 terminals in the LPl and IPm are involved in thesecomplex synaptic arrangements. Although percentagesare not available, X-cell retinal RLs in the cat are morefrequently involved in triads than are Y-cell RLs (Wilson etal., 1984), whereas the opposite is true in the rhesusmonkey (Wilson, 1989).

Our data show that approximately 35% of the labeledarea 17 RLs lie adjacent to F profiles. These findings can becompared with data from the prosimian Galago, where10–20% of the X- and Y-cell retinal RLs lie adjacent to Fprofiles (Feig and Harting, 1994). Such data are notavailable in the cat, but retinal Y-cell RLs lie adjacent to Fprofiles more frequently than do the X-cell RLs (Wilson etal., 1984). The TRN is the primary source of these Fprofiles in the LGN (Harting et al., 1991) and most likelyalso in the LPl and IPm (Fitzgibbon et al., 1995).

These morphological features of area 17 cortico-LPlterminals confirm the observations of Vidnyansky et al.(1996). However, data regarding the spatial relationshipsof these RL terminals to other afferents on LPl cells, whichwould have further confirmed their lemniscal nature, werenot analyzed. Studies in the macaque monkey have shownthat the area 17 projection to the IP is comprised exclu-sively of RLs (Campos-Ortega and Hayhow, 1973) orconsists of a mixture of both RLs and RSs (Ogren andHendrickson, 1979b). The present findings in the owlmonkey confirm those of Campos-Ortega and Hayhow(1973) by showing that, although the cortical profiles varyin size, they are larger than RS profiles and exhibit RLcharacteristics. The smaller profiles (RSs) seen by Ogrenand Hendrickson (1979b) could have been pieces of largerRLs rather than classical RS profiles. However, lightmicroscopic data in the macaque have shown that cortico-pulvinar fibers arising from layer 5 cells in area 17 havetwo distinct morphologies called type I (E) and type II (R;Rockland, 1996). Type II morphologies are the most numer-ous in the IP (Rockland, 1998) and are most likely compa-

rable to our RLs. However, because the ultrastructuralfeatures and synaptic relationships of these terminalshave not been determined, they cannot be compared withRL or RS profiles. The type I (E) profile may be a smallerversion of an RL (there are different sizes of X, Y, and WRLs in the LGN; Mize et al., 1986; Feig and Harting, 1994),or there may be a small population of layer 5 cells thatgives rise to RS profiles. Other alternatives include speciesdifferences and possibly involvement of fibers from othercortical areas in Rockland’s (1998) experiments. In anycase, the present findings in the owl monkey show that allarea 17 profiles in the IPm are RLs. These terminals areinvolved in triadic arrangements and exhibit the connec-tional associations of lemniscal afferents. We emphasizethat neither of the ultrastructural studies in the primate(Campos-Ortega and Hayhow, 1973; Ogren and Hendrick-son, 1979b) analyzed the overall connectional relation-ships of the cortical RLs with other terminal populations.We have done this in both the cat and owl monkey, and ourdata clearly demonstrate that cortical RLs in the LPl andIPm exhibit similar spatial relationships with RS and Fprofiles on thalamic cell dendrites as retinal RLs exhibit inthe LGN.

Functional implications

It is well documented that information reaching primaryvisual cortex from the LGN is conveyed to higher corticalareas via corticocortical pathways (see Casagrande andKaas, 1994). Although considerable attention has beenfocused on the feedforward and feedback connections thatlink the several visual cortical areas to each other (VanEssen and Maunsell, 1983; Zeki and Shipp, 1988; Casa-grande and Kaas, 1994; Salin and Bullier, 1995), thecontribution(s) that thalamic inputs make to these corticalareas is largely undefined. Our findings suggest thatvisual information flowing out of area 17 via the layer 5‘‘re-entrant’’ pathway (Guillery, 1995) provides a secure(‘‘drives’’) and potentially important communication linkbetween primary and higher visual cortical areas via theLPl and IPm (Mathers, 1972b; Ogren and Hendrickson,1976; Trojanowski and Jacobson, 1976; Robson and Hall,1977; Lin and Kaas, 1979; Raczkowski and Diamond,1980; Abramson and Chalupa, 1985; Ojima et al., 1996).

This raises the question of why channel information‘‘back’’ through the thalamus instead of sending it directlyto higher cortical area(s) via corticocortical connections.The layer 5 projection contains information about what ishappening in area 17, and this signal reaches not only‘‘association’’ or higher order thalamic nuclei but alsomidbrain and brainstem areas (Deschenes et al., 1994;Deschenes and Bourassa, 1995). Routing this informationthrough higher order thalamic nuclei provides an opportu-nity for possible modulation and updating by subcorticaland/or higher cortical inputs (see Fig. 10A,B; Mathers,1972b; Ogren and Hendrickson, 1976; Robson and Hall,1977; Graham et al., 1979; Abramson and Chalupa, 1985;Fitzpatrick et al., 1989; Fitzgibbon et al., 1995).

Several different higher cortical areas give rise to layer 6projections to the LPl and IPm (Mathers, 1972b; Ben-evento and Rezak, 1976; Updyke, 1983; Abramson andChalupa, 1985) and could modulate the area 17 driven LPland IPm neurons. We suggest that LPl and IPm cellsdriven by area 17 RLs project to layer 4 (like ascendinglemniscal-driven thalamic cells) of some higher cortical


area(s) (Benevento and Rezak, 1976) and that layer 6 cellsin these same higher cortical areas project back on thearea 17 driven LPl and IPm cells (Fig. 10A). We alsosuggest that this layer 6 projection sends a collateral to theTRN, from which inhibitory fibers reach the LPl and IPm(Fitzgibbon et al., 1995; Ojima et al., 1996). In the LGN,this pattern of connections has been suggested to play animportant role in modulating thalamocortical neurons(Singer, 1977; McCormick, 1992; McCormick and vonKrosigk, 1992; McCormick and Bal, 1994; Guido and Lu,1995; Guido and Weyland, 1995), and there could be a

similar role for these connections within the LPl and IPm(Guillery et al., 1998).

Layer 5 cells in several higher visual cortical areas alsoproject to the LPl and IPm (Raczowski and Diamond, 1980;Abramson and Chalupa, 1985). We propose that all ofthese layer 5 terminals exhibit the morphology and postsyn-aptic relationships of RLs (Fig. 10B). Each RL driven cellin the LPl and IPm would project to layer 4 of a highercortical area and in turn be modulated by a layer 6projection from this same area. Individual cells in the LPland IPm may be driven by layer 5 cells from only onecortical area (e.g., RLs associated with retinal X- andY-cells do not converge on single relay cells in the LGN;Hoffman et al., 1972; So and Shapley, 1979; Sur et al.,1982; Wilson et al., 1984). However, some cells within theLPl have response properties, which suggest that theremay be convergence of layer 5 input from different corticalareas (Chalupa and Abramson, 1989), and it is alsopossible that some cortical areas modulate, via layer 6projections, the activity of LPl and IPm neurons thatproject to layer 4 of an another cortical area (i.e., nonrecip-rocal). These issues need to be explored in further studies.

In conclusion, this proposed pattern of corticothalamic(both drivers and modulators) and thalamocortical connec-tions suggests an important role for the LPl and IPm incorticocortical interactions. Information about ongoingactivity within the various higher visual cortical areascould be reliably relayed between different areas via theLPl and IPm. Moreover, the specific ‘‘driver’’ signals pass-ing between the various higher visual cortical areas couldbe selectively modulated within the LPl and IPm depend-ing on the behavioral needs of the cat or owl monkey.Subcortical inputs (Singer, 1977; Fitzpatrick et al., 1989;Conley and Diamond, 1990; McCormick, 1992; Fitzgibbonet al., 1995) could also work alone or in concert withcortical inputs to carry out such modulation, with the endresult being the generation of visual salience and thegeneration of visuoperceptual and visuomotor states thatare most beneficial to the organism (Robinson and Peter-son, 1992; Robinson, 1993).

Fig. 10. A: A comparison of the connectional relationships oflemniscal afferents in the lateral geniculate nucleus (LGN; retinalRLs) and the lateral posterior nucleus (LPl) and inferior pulvinarnucleus (IPm; area 17 RLs). RLs, in the LGN, are involved in triadicarrangements with the processes of interneurons (I) and the dendritesof relay cells (R). F terminals from thalamic reticular nucleus (TRN)cells are distributed on dendrites near the RL profiles, whereas RSterminals are located on dendritic regions that are distal to RLs. Thepresent findings show that area 17 RLs in the LPl and IPm areinvolved in triadic arrangements and are often accompanied by Fprofiles on the postsynaptic dendrite. We propose that TRN cellsaccount for some of these F profiles (see Guillery et al., 1998). RSprofiles are not spatially associated with RL profiles, and we proposethat these profiles are distributed distally and that some arise fromlayer 6 cells in higher cortical areas. Relay cells within the LGNproject to layer 4 of area 17 and cells in the LPl and IPm project tolayer 4 of higher cortical areas. We propose that some of these relaycells are driven by RL input from area 17. B: An illustration of thewell-documented direct and the presently proposed indirect pathwaysfrom area 17 to higher cortical visual areas. The direct pathwaytravels primarily via supragranular projections to layer 4. The indirectpathway involves the re-entrant layer 5 corticothalamic projection tothe LPl and IPm and thalamocortical projection to layer 4 of highercortical area(s). We propose that layer 5 RL terminals provide afunctional ‘‘drive’’ to cells within the LPl and IPm.


ACKNOWLEDGMENTS

We thank R.W. Guillery for many of the ideas discussedin the paper. We also thank David VanLieshout for help inmany aspects of this study. This study was supported byNational Eye Institute grant EY-01277 to J.K.H.

LITERATURE CITED

Abramson, B.P. and L.M. Chalupa (1985) The laminar distribution ofcortical connections with tecto-and cortico-recipient zones in the cat’slateral posterior nucleus. Neuroscience 15:81–95.

Allman, J.H. and J.H. Kaas (1971) Representation of the visual field instriate and adjoining cortex of the owl monkey (Aotus trivirgatus).Brain Res. 35:89–106.

Bachmann, L. and M.M. Salpeter (1965) Autoradiography with the electronmicroscope. Lab. Invest. 14:303–315.

Bender, D.B. (1983) Visual activation of neurons in the primate pulvinardepends on cortex but not colliculus. Brain Res. 279:258–261.

Benevento, L.A. and M. Rezak (1976) The cortical projections of the inferiorpulvinar and adjacent lateral pulvinar in the rhesus monkey (Macacamulatta): An autoradiographic study. Brain Res. 108:1–24.

Campos-Ortega, J.A. and W.R. Hayhow (1973) The synaptic organization inthe inferior pulvinar of the rhesus monkey (Macaca mulatta). BrainBehav. Evol. 7:203–247.

Casagrande, V.A. and J.H. Kaas (1994) The afferent, intrinsic, and efferentconnections of primary visual cortex in primates. In A. Peters and K.S.Rockland (eds): Cerebral Cortex. New York: Plenum Press, pp. 201–259.

Casanova, C. and T. Savard (1996) Responses to moving texture patterns ofcells in the striate-recipient zone of the cat’s lateral posterior-pulvinarcomplex. Neuroscience 70:439–447.

Casanova, C., R.D. Freeman, and J.P. Nordmann (1989) Monocular andbinocular response properties of cells in the striate-recipient zone of thecat’s lateral posterior-pulvinar complex. J. Neurophysiol. 62:544–557.

Chalupa, L.M. and B.P. Abramson (1989) Visual receptive fields in thestriate-recipient zone of the lateral posterior-pulvinar complex. J.Neurosci. 9:347–357.

Conley, M. and I.T. Diamond (1990) Organization of the visual sector of thethalamic reticular nucleus in Galago. Eur. J. Neurosci. 2:211–226.

Cowan, W.M., D.I. Gottleib, A.E. Hendrickson, J.C. Price, and T.A. Woolsey(1972) The autoradiographic demonstration of axonal connections inthe central nervous system. Brain Res. 37:21–51.

Deschenes, M. and J. Bourassa (1995) Corticothalamic projections from theprimary visual cortex in rats: A single fiber study using biocytin as ananterograde tracer. Neuroscience 66:253–263.

Deschenes, M., J. Bourassa, and D. Pinault (1994) Corticothalamic projec-tions from layer V cells in rat are collaterals of long range corticofugalaxons. Brain Res. 664:215–219.

Famiglietti, E.V. and A. Peters (1972) The synaptic glomerulus and theintrinsic neuron in the dorsal lateral geniculate nucleus of the cat. J.Comp. Neurol. 144:285–334.

Feig, S. and J.K. Harting (1992) Ultrastructural studies of the primateparabigeminal nucleus: Electron microscopic autoradiographic analysisof the tectoparabigeminal projection in Galalgo crassicaudatus. BrainRes. 595:334–338.

Feig, S. and J.K. Harting (1994) Ultrastructural studies of the primatelateral geniculate nucleus: Morphology and spatial relationships ofaxon terminals arising from the retina, visual cortex (area 17), superiorcolliculus, parabigeminal nucleus, and pretectum of Galago crassicau-datus. J. Comp. Neurol. 343:17–34.

Feig, S., D.P. Van Lieshout, and J.K. Harting (1992) Ultrastructural studiesof retinal, visual cortical (area 17), and parabigeminal terminals withinthe superior colliculus of Galago crassicaudatus. J. Comp. Neurol.319:85–99.

Fitzgibbon, T., L.V. Tevah, and A.J. Sefton (1995) Connections between thereticular nucleus of the thalamus and pulvinar-lateralis posteriorcomplex: a WGA-HRP study. J. Comp. Neurol. 363:489–504.

Fitzpatrick, D., I.T. Diamond, and D. Raczkowski (1989) Cholinergic andmonoaminergic innervation of the cat’s thalamus: Comparison of thelateral geniculate nucleus with other principal sensory nuclei. J. Comp.Neurol. 288:647–675.

Graham, J., C.S. Lin, and J.H. Kaas (1979) Subcortical projections of sixvisual cortical areas in the owl monkey, Aotus trivirgatus. J. Comp.Neurol. 187:557–580.

Guido, W. and S.M. Lu (1995) Cellular bases for the control of retino-geniculate signal transmission. Int. J. Neurosci. 80:41–63.

Guido, W. and T. Weyland (1995) Burst responses in thalamic relay cells ofthe awake behaving cat. J. Neurophysiol. 74:1782–1786.

Guillery, R.W. (1967) The organization of synaptic interconnections in thelaminae of the dorsal lateral geniculate nucleus of the cat. Am. J. Anat.120:583–606.

Guillery, R.W. (1969) The organization of synaptic interconnections in thelaminae of the dorsal lateral geniculate nucleus of the cat. Z. Zellforsch.96:1–38.

Guillery, R.W. (1971) Patterns of synaptic interconnections in the dorsallateral geniculate nucleus of cat and monkey: A brief review. Vis. Res.Suppl. 3:211–227.

Guillery, R.W. (1995) Anatomical evidence concerning the role of thethalamus in corticocortical communication: A brief review. J. Anat.187:583–592.

Guillery, R.W. and M. Colonnier (1969) Synaptic patterns in the dorsallateral geniculate nucleus of the monkey. Z. Zellforsch. 103:90–108.

Guillery, R.W., S.L. Feig, and D.A. Lozadi (1998) Paying attention to thethalamic reticular nucleus. Trends Neurosci. 21: 28–32

Hamos, J.E., S.C. VanHorn, D. Raczkowski, D.J. Uhlrich, and S.M.Sherman (1985) Synaptic connectivity of a local circuit neurone inlateral geniculate nucleus of the cat. Nature 317:618–621.

Harting, J.K., D.P. Van Lieshout, and S. Feig (1991) Connectional studies ofthe primate lateral geniculate nucleus: Distribution of axons arisingfrom the thalamic reticular nucleus of Galago crassicaudatus. J. Comp.Neurol. 310:411–427.

Harting, J.K., S. Feig, and D.P. Van Leishout (1997) Cortical somatosensoryand trigeminal inputs of the cat superior colliculus: Light and electronmicroscopic analyses. J. Comp. Neurol. 388:313–326

Hendrickson, A.E. (1972) Electron microscopic distribution of axoplasmictransport. J. Comp. Neurol. 144:381–398.

Hoffman, K.P., J. Stone, and S.M. Sherman (1972) Relay of receptive-fieldproperties in dorsal lateral geniculate nucleus of the cat. J. Neuro-physiol. 35:518–531.

Kopriwa, B.M. (1967) The influence of development on the number andappearance of silver grains in electron microscopic autoradiography. J.Histochem. Cytochem. 15:501–515.

Kopriwa, B.M. (1973) A reliable, standardized method for ultrastructuralelectron microscopic radioautography. Histochemie 37:1–17.

Lin, C.S. and J.H. Kaas (1979) The inferior pulvinar complex in owlmonkeys: Architectonic subdivisions and patterns of input from thesuperior colliculus and subdivisions of visual cortex. J. Comp. Neurol.187:655–678.

Lund, J.S., R.D. Lund, A.E. Hendrickson, A.H. Bunt, and A.F. Fuchs (1975)The origin of efferent pathways from the primary visual cortex, area 17,of the Macaque monkey as shown by retrograde transport of horserad-ish peroxidase. J. Comp. Neurol. 164:287–304.

Mathers, L.H. (1972a) Ultrastructure of the pulvinar of the squirrelmonkey. J. Comp. Neurol. 146: 15–42.

Mathers, L.H. (1972b) The synaptic organization of the cortical projectionto the pulvinar of the squirrel monkey. J. Comp. Neurol. 146:43–60.

McCormick, D.A. (1992) Neurotransmitter actions in the thalamus andcerebral cortex and their role in neuromodulation of thalamocorticalactivity. Prog. Neurobiol. 39:337–388.

McCormick, D.A. and M. von Krosigk (1992) Corticothalamic activationmodulates thalamic firing through glutamate ‘‘metabotropic’’ receptors.Proc. Natl. Acad. Sci. U.S.A. 89:2774–2778.

McCormick, D.A. and T. Bal (1994) Sensory gating mechanisms of thethalamus. Curr. Opin. Neurobiol. 4:550–556.

Mize, R.R., R.F. Spencer, and L.H. Horner (1986) Quantitative comparisonof retinal synapses in the dorsal and ventral (parvicellular) C laminaeof the cat dorsal lateral geniculate nucleus. J. Comp. Neurol. 248:57–73.

Ogren, M. and A.E. Hendrickson (1976) Pathways between striate cortexand subcortical regions in Macaca mulatta and Saimiri sciureus:Evidence for a reciprocal pulvinar connection. J. Comp. Neurol. 53:780–800.

Ogren, M.P. and A.E. Hendrickson (1979a) The structural organization ofthe inferior and lateral subdivisions of the macaca monkey pulvinar. J.Comp. Neurol. 188:147–178.

Ogren, M.P. and A.E. Hendrickson (1979b) The morphology and distribu-tion of striate cortex terminals in the inferior and lateral subdivisions ofthe Macaca monkey pulvinar. J. Comp. Neurol. 188:179–200.


Ojima, H., K. Murakami, and K. Kishi (1996) Dual termination modes ofcorticothalamic fibers originating from pyramids of layers 5 and 6 in catvisual cortical area 17. Neurosci. Lett. 208:57–60.

Raczkowski, D. and I.T. Diamond (1980) Cortical connections of thepulvinar nucleus in Galago. J. Comp. Neurol. 193:1–40.

Rapisardi, S.C. and T.P. Miles (1984) Synaptology of retinal terminals in thedorsal lateral geniculate nucleus of the cat. J. Comp. Neurol. 223:515–534.

Robinson, D.L. (1993) Functional contributions of the primate pulvinar.Prog. Brain Res. 95:371–780.

Robinson, D.L. and S.E. Peterson (1992) The pulvinar and visual salience.Trends Neurosci. 15:127–132.

Robson, J. A. and W.C. Hall (1977) The organization of the pulvinar in thegrey squirrel (Sciurus carolinensis) II. Synaptic organization andcomparisons with the dorsal lateral geniculate nucleus. J. Comp.Neurol. 173:389–416.

Rockland, K.S. (1996) Two types of corticopulvinar terminations: Round(Type 2) and elongate (Type 1). J. Comp. Neurol. 368:57–87.

Rockland, K.S. (1998) Convergence and branching patterns of round, type 2corticopulvinar axons. J. Comp. Neurol. 390:515–536.

Ross, R. and E.P. Benditt (1965) Wound healing and collagen formation V.Quantitative electron microscope radioautographic observations ofproline-3H utilization by fibroblasts. J. Cell Biol. 27:83–106.

Salin, P.-A. and J. Bullier (1995) Corticocortical connections in the visualsystem: Structure and function. Physiol. Rev. 75:107–154.

Sherman, S.M. and R.W. Guillery (1996) Functional organization of thala-mocortical relays. J. Neurophysiol. 76:1367–1395.

Singer, W. (1977) Control of thalamic transmission by corticofugal andascending reticular pathways in the visual system. Physiol. Rev.57:386–420.

So, Y.T. and R. Shapley (1979) Spatial properties of X and Y cells in thelateral geniculate nucleus of the cat and conduction velocities of theirinputs. Exp. Brain Res. 36:533–550.

Sur, M., A.L. Humphrey, and S.M. Sherman (1982) Monocular deprivation

affects X- and Y-cell retinogeniculate terminations in cats. Nature300:183–185.

Takacs, J., J., Hamori, and V. Silakov (1991) GABA-containing neuronalprocesses in normal and cortically deafferented dorsal lateral genicu-late nucleus of the cat: An immunogold and quantitative EM study. Exp.Brain Res. 83:562–574.

Trojanowski, J.Q. and S. Jacobson (1976) Areal and laminar distribution ofsome pulvinar cortical efferents in Rhesus monkey. J. Comp. Neurol.169:371–392.

Tusa, R.J., L.A. Palmer, and A.C. Rosenquist (1978) The retinotopicorganization of area 17 (striate cortex) in the cat. J. Comp. Neurol.177:213–236.

Updyke, B.V. (1983) A reevaluation of the functional organization andcytoarchitecture of the feline lateral posterior complex with observa-tions on adjoining cell groups. J. Comp. Neurol. 219:43–181.

Van Essen, D.C. and J.H.R. Maunsell (1983) Hierarchical organization andfunctional streams in the visual cortex. Trends Neurosci. 6:370–375.

Vidnyanszky, Z., Z. Borostyankoi, T.J. Gorcs, and J. Hamori (1996) Lightand electron microscopic analysis of synaptic input from cortical area 17to the lateral posterior nucleus in cats. Exp. Brain Res. 109:63–70.

Weber, A.J. and R.E. Kalil (1987) Development of corticogeniculate syn-apses in the cat. J. Comp Neurol. 264:171–192.

Wilson, J.R. (1989) Synaptic organization of individual neurons in themacaque lateral geniculate nucleus. J. Neurosci. 9:2931–2953.

Wilson, J.R. and A.E. Hendrickson (1981) Neuronal and synaptic structureof the dorsal lateral geniculate nucleus in normal and monocularlydeprived Macaca monkeys. J. Comp. Neurol. 197:517–539.

Wilson, J.R., M.J. Friedlander, and S.M. Sherman (1984) Fine structuralmorphology of identified X- and Y-cells in the cat’s lateral geniculatenucleus. Proc. R. Soc. Lond. 221:411–436.

Wong-Riley, M.T.T. (1972) Neuronal and synaptic organization of thenormal dorsal lateral geniculate nucleus of the squirrel monkey,Saimiri sciureus. J. Comp. Neurol. 144:25–60.

Zeki, S. and S. Shipp (1988) The functional logic of cortical connections.Nature 335:311–317.


Documents

Corticocortical communication via the thalamus: Ultrastructural studies of corticothalamic projections from area 17 to the lateral posterior nucleus of the cat and inferior pulvinar