Tectal Mosaic: Organization of the Descending Tectal ... JCN 2004.pdf · Descending Tectal...

Tectal Mosaic: Organization of theDescending Tectal Projections in

Comparison to the Ascending TectofugalPathway in the Pigeon

BURKHARD HELLMANN, ONUR GUNTURKUN, AND MARTINA MANNS*

Abteilung Biopsychologie, Institut fur Kognitive Neurowissenschaft, Fakultat furPsychologie, Ruhr-Universitat Bochum, 44780 Bochum, Germany

ABSTRACTThe optic tectum of vertebrates is an essential relay station for visuomotor behavior and

is characterized by a set of connections that comprises topographically ordered input from theeyes and an output that reaches premotor hindbrain regions. In the avian tectofugal system,different ascending cell classes have recently been identified based on their dendritic andaxonal projection patterns, although comparable information about the descending cells ismissing. By means of retrograde tracing, the present study describes the detailed morphologyof tectal output neurons that constitute the descending tectobulbar and tectopontine path-ways in pigeons. Descending cells were more numerous in the dorsal tectum and differed interms of 1) their relative amount of ipsi- vs. contralateral projections, 2) the location of theefferent cell bodies within different tectal layers, and 3) their differential access to visualinput via dendritic ramifications within the outer retinorecipient laminae. Thus, the descend-ing tectal system is constituted by different cell classes presumably processing diverseaspects of the visual environment in a visual field-dependent manner. We demonstrate, basedon a careful morphological analysis and on double-labeling experiments, that the descendingpathways are largely separated from the ascending projections even when they arise from thesame layers. These data support the concept that the tectum is arranged as a mosaic ofmultiple cell types with diverse input functions at the same location of the tectal map. Suchan arrangement would enable the tectal projections onto diverse areas to be both retinotopi-cally organized and functionally specific. J. Comp. Neurol. 472:395–410, 2004.© 2004 Wiley-Liss, Inc.

Indexing terms: tectum opticum; superior colliculus; tectobulbar pathway; tectopontine pathway;

extrageniculocortical pathway; birds

In vertebrates, the optic tectum is an essential relaystation for visuomotor transformation. In accordance withthis function, the tectum is characterized by an evolution-ally conserved set of projections conveying ascending vi-sual output to the forebrain and descending projections topremotor hindbrain regions. Despite this similar assemblyof efferent neurons, the size and laminar organization ofthe tectum vary considerably among vertebrate species(Reiner, 1994).

Two distinct tectal pathways convey visual informationto the forebrain in amniotes. One of them arises fromneurons characterized by wide dendritic fields projectingto the nucleus rotundus in sauropsids (Karten et al., 1997;Luksch et al., 1998; Hellmann and Gunturkun, 2001;Davila et al., 2002; Belekhova et al., 2003; Marın et al.,

2003) or the lateral posterior/pulvinar nucleus in mam-mals (Major et al., 2000, and references therein). Thesediencephalic structures are then the source of projections

Grant sponsor: SFB Neurovision of the Deutsche Forschungsgemein-schaft; Grant sponsor: Alfried Krupp-Stiftung.

*Correspondence to: Martina Manns, Biopsychologie, Institut fur Kog-nitive Neurowissenschaft, Fakultat fur Psychologie, Ruhr-Universitat Bo-chum, 44780 Bochum, Germany.E-mail: [email protected]

Received 10 October 2003; Revised 16 December 2003; Accepted 22December 2003

DOI 10.1002/cne.20056Published online the week of March 29, 2004 in Wiley InterScience

(www.interscience.wiley.com).

THE JOURNAL OF COMPARATIVE NEUROLOGY 472:395–410 (2004)

© 2004 WILEY-LISS, INC.

to diverse telencephalic structures (sauropsids: Benowitzand Karten, 1976; Nixdorf and Bischof, 1982; Guirado etal., 2000; Laverghetta and Shimizu, 2003; mammals: Linand Kaas, 1980; Beck and Kaas, 1998). A second tectalefferent system arises from radial neurons with small-field dendrites projecting to the dorsolateral thalamus,from which projections lead to the forebrain (Gamlin andCohen, 1986; Wild, 1989; Harting et al., 1991; Martinez-Marcos et al., 1998; Belekhova et al., 2003). In amphibi-ans, similar cell types innervate dorsal and ventral tha-lamic areas (Lazar et al., 1983; Dicke and Roth, 1996;Dicke, 1999; Roth et al., 1999). Descending tectal informa-tion is conveyed by ipsi- and contralateral projections topremotor hindbrain regions. In amniotes, an ipsilateraltectopontine pathway (ITP) can be distinguished from acrossed tectobulbar pathway (CTB; or predorsal bundle inmammals; for review see Reiner, 1994). Comparably to theascending projections, the descending pathways are con-stituted of radial and multipolar cell populations. Thesimilarity of ascending and descending efferent cell typesin tetrapodes is considered as important evidence for thecommon origin of these cells (Reiner, 1994). The morpho-logical and electrophysiological similarities of the charac-teristic wide-field neurons in birds and mammals are sostriking that Major et al. (2000) termed this a “cellularhomology.”

Despite these cellular similarities between tetrapods,the ascending and descending efferent systems displayconsiderable cytoarchitectonic variations concerning thelaminar localization of their cell bodies, their axonal pro-jection patterns, and their dendritic lamination and,hence, their access to retinal input. In amphibians, as-cending and descending cells are intermingled, and singlecells eventually give rise to axons with collaterals project-ing to both thalamus and hindbrain (Dicke, 1999; Roth etal., 1999). In amniotes, ascending and descending cellsmigrate during development to a differential degree intothe superficial laminae (Reiner, 1994). In the mammaliansuperior colliculus, most tectal neurons with descendingprojections lie deeper than those with ascending projec-tions (for review see Reiner, 1994). In contrast, sauropsidsdisplay a substantial overlap in the localization of ascend-ing and descending cell bodies, although morphologicaldata suggest that both populations represent separatecells types (Reiner and Karten, 1982; Reiner, 1994; Kartenet al., 1997). A detailed analysis of ascending and descend-ing tectal cell types in sauropsids is partly hampered bythe insufficient definition of “superficial” and “deep” withrespect to the location of cells along the tectal z-axis.Although the stratum opticum of the mammalian superiorcolliculus provides a clear means to discriminate cells inthe stratum griseum superficiale with monosynaptic reti-nal input as “superficial” and those with indirect polysyn-aptic retinal input in the stratum griseum intermediumand below as nonretinorecipient and “deep,” such a simpledifferentiation is not feasible in sauropsids. In birds, forexample, the majority of cells in the deep layer 13 shouldin fact be classified as “superficial” because of their den-dritic ramifications within the superficial retinorecipientlayers (Karten et al., 1997) and their monosynaptic retinalinput (Hardy et al., 1984). Subpopulations of efferent tec-tal cells display unique dendritic arborization patternswithin distinct tectal laminae and are innervated by dif-ferent retinal ganglion cells types (mammals: Harrell etal., 1982; Ogawa et al., 1985; Abramson and Chalupa,

1988; Mooney et al., 1988a; Major et al., 2000; birds:Karten et al., 1990; Hellmann and Gunturkun, 1999,2001; Luksch et al., 2001). Thus, they probably processqualitatively divergent visual information. These func-tionally divergent ascending tectal cell types converge indifferent thalamic substructures, so a transformationfrom a retinotopic to a functionotopic organization is re-quired (Wang et al., 1993; Hellmann and Gunturkun,2001; Marın et al., 2003). Neurophysiological studies sug-gest comparably separate descending channels from thetectum to premotor hindbrain areas that encode distinctparameters of stimulus location (Masino and Grobstein,1989; Masino and Knudsen, 1990). However, the subserv-ing tectal cell types are not well characterized, and it isunclear how the retinotopic coding of the tectum is trans-formed into visuomotor coordinates.

Birds display the highest degree of tectal differentia-tion, with at least 15 distinguishable laminae (accordingto Cajal, 1911). Neurons ascending to the dorsal and ven-tral thalamus (Gamlin and Cohen, 1986; Wild, 1989) anda subpopulation of neurons descending within the tec-topontine tract arise from cells whose somata are locatedwithin intermediate layers 8–10 (Reiner and Karten,1982). Cells establishing projections to the rotundus and alarge portion of cells descending into the brain stem arelocated in efferent layer 13 (Benowitz and Karten, 1976;Reiner and Karten, 1982; Hellmann and Gunturkun,2001). Recently, the neuronal elements of the ascendingtectorotundal pathway, which connects the tectum via thediencephalic nucleus rotundus with the telencephalic ect-ostriatum, have been studied in detail. These cells aredifferentiated into at least five subtypes, demonstratingthe high degree of complexity of the tectorotundal system(Karten et al., 1997; Luksch et al., 1998; Hellmann andGunturkun, 2001; Marın et al., 2003). This constitutes anexcellent basis for comparative description of the descend-ing neuronal elements. The clarification of how far theascending information is separated from the descendinginformation in such a highly differentiated visual systemas the avian tectofugal pathway might help to elucidatethe functional organization principles of tectal visuomotorprocessing.

The present work, based on a detailed morphologicalcomparison of ascending and descending neuronal popu-lations as well as on double-labeling experiments, demon-strates that both pathways emerge from largely separatedtectal cell groups. Additionally, our study suggests that,similarly to the ascending tectal projections, the descend-ing tectal efferents also constitute a mosaic of diverse celltypes that transform a retinotopic code into segregatedfunctional domains.

MATERIALS AND METHODS

Seventeen adult pigeons (Columba livia) of both sexesfrom local breeding stocks were used in this study. Toidentify the localization of the descending cell populations,choleratoxin subunit B (CtB; Sigma, Deisenhofen, Ger-many) was injected unilaterally either into the lateraltegmentum to label the tectopontine fiber tract (n � 5; Fig.1A) or into the medial tegmentum to label the tectobulbarfiber tracts (n � 8; Fig. 1B) at rostrocaudal levels A1.25–A3.00 (Karten and Hodos, 1967). Additionally, two ani-mals received bilateral injections of CtB into the rostro-medial mesencephalic tegmentum at the level of nucleus

396 B. HELLMANN ET AL.

Fig. 1. Photomicrographs of tracer injection sites. A: To label thetectopontine system (ITP), tracer injections were placed in the lateraltegmental region of the FRL. B: To label the tectobulbar system(CTB), tracers were applied to the medial tegmental region of BCDand DBC. Note that the PL as the target of the ITP was not labeled,indicating the specificity of the medial injection. C: To label tectalefferents projecting to the rostral tegmentum, tracer was injected intothe rostromedial tegmentum, including Ru/IS. D: To label the ascend-

ing tectal cell populations, TDA was injected into the RT and GLd.BCD, brachium conjunctivum descendens; DBC, decussatio brachi-orum conjunctivorum; FRL, formation reticularis lateralis mesen-cephali; FRM, formation reticularis medialis mesencephali; GLd, nu-cleus geniculatus lateralis, pars dorsalis; IS, nucleus interstitialis(Cajal); PL, nucleus pontis lateralis; RT, nucleus rotundus; Ru, nu-cleus ruber. Scale bars � 1,000 �m in A–C, 500 �m in D.

397DESCENDING TECTAL PROJECTIONS

ruber and nucleus interstitialis (Cajal; Fig. 1C). All CtB-injected animals also received bilateral injections of Texasred dextran amine (TDA; 10,000 MW, lysine fixable; Mo-lecular Probes, Leiden, The Netherlands) into the centraland dorsal diencephalon (Fig. 1D). All experiments werecarried out according to the specifications of the Germanlaw for the prevention of cruelty to animals.

Prior to surgery, the pigeons were anesthetized withequithesin (0.31–0.33 ml/100 g body weight), and the an-imals were placed into a stereotaxic apparatus (Kartenand Hodos, 1967). The scalps were infiltrated with xylo-caine and incised dorsally. Next, the skull was openedwith a dental drill, and a glass micropipette (outer tipdiameter 20 �m) mounted to a mechanic pressure device(WPI nanoliterinjector) was inserted into varying sites ofthe brain (diencephalic injections: anterior 5.3–7.0, dorsal5.0–8.5, lateral 1.8–3.8; rostromedial tegmentum: ante-rior 3.5–4.5, dorsal 5.0–7.0, lateral 0.5–0.9; lateral teg-mentum: anterior 1.5–2.5, dorsal 3.0–4.0, lateral –1, with15° dorsolateral slant; medial tegmentum: anterior 1.5–2.5, dorsal 2.0–4.0, lateral 0.5) according to the stereo-taxic coordinates of the pigeon brain atlas by Karten andHodos (1967). Injection depths ranged from 1.0 to 2.8 mm.In cases of tegmental injections, approximately 50 nl CtB[1% (w/v) in distilled water] were injected in steps of 2 nlover a period of 15–20 minutes. Diencephalic injectionswere performed with 100 nl TDA [10% (w/v) in 2% di-methyl sulfoxide]. Subsequently, the pipette was removed,and the skin was infiltrated again with xylocaine andsutured.

After 2 days of survival, animals received an injection of200 U sodium heparin and were then deeply anesthetizedwith an overdose of equithesin (0.55 ml/100 g bodyweight). The pigeons were perfused through the heartwith 100 ml 0.9% (w/v) sodium chloride and 800 ml ice-cold 4% paraformaldehyde in 0.12 M phosphate buffer(PB), pH 7.4. The brains were removed and stored for 2hours in fixative with a supplement of 15% sucrose (w/v).Subsequently, the brains were stored overnight in a solu-tion of 30% sucrose in 0.12 M PB. On the following day, thebrains were cut in the frontal plane at 35 �m on a freezingmicrotome, and the slices were collected in PB containing0.1% sodium azide (w/v).

Brain slices were reacted freely floating according to theimmuno-ABC technique. The sections were placed for 35minutes in 0.5% hydrogen peroxidase in distilled water toreduce endogenous peroxidase activity. In cases of soleCtB detection, sections were incubated overnight at 4°C inthe primary antibody [rabbit anticholeragenoid; SigmaGermany; 1/20,000 in 0.12 M PB with the addition of 2%NaCl (w/v), 0.3% Triton X-100 (v/v), and 5% normal goatserum]. After being rinsed, the sections were incubated for60 minutes at room temperature in the biotinylated sec-ondary antibody (goat anti-rabbit; Vectastain; VectorCamon, Wiesbaden, Germany; 1/250 in 0.12 M PB � 2%NaCl � 0.3% Triton X-100). After additional rinsing, thesections were incubated for 60 minutes in avidin-biotin-peroxidase solution (Vectastain ABC-Elite Kit, VectorCamon; 1/100 in 0.12 M PB � 2% NaCl � 0.3% TritonX-100). After washing, the peroxidase activity was de-tected by using a heavy-metal-intensified 3�3-diaminobenzidine (DAB; Sigma) reaction (Adams, 1981),modified by the use of �-d-glucose/glucose oxidase (Sigma)instead of hydrogen peroxidase (Shu et al., 1988). Thesections were mounted on gelatin-coated slides, dehy-

drated, and coverslipped with Permount (Fisher Scien-tific, Fair Lawn, NJ). Some sections were counterstainedwith cresyl violet. In one series of sections (every tenthsection), primary CtB antibody was detected with a sec-ondary fluorescein-labeled antibody (goat anti-rabbit; Vec-tor), which was incubated for 1 hour at room temperature(1/75 in 0.12 M PB � 2% NaCl � 0.3% Triton X-100).Subsequently, sections were rinsed, mounted on gelati-nized slides, and coverslipped with the ProLong AntifadeKit (Molecular Probes, Eugene, OR). Sections were ob-served with an Olympus BH2 epifluorescence microscopewith the following filter settings: for fluorescein, OlympusIB set with additional shortpass emission filter G520; forTDA, Chroma filters (Brattleboro, VT) with excitation fil-ter HQ-577/10, dichroic mirror Q-585LP, and emissionfilter HQ-645/75.

The rotundal tracer injection sites and the resultingretrograde CtB labeling within the optic tectum were an-alyzed with an Olympus BH2 microscope. Qualitative re-constructions of the rotundal/triangular CtB diffusionarea and of the position of labeled somata as well as theirperipheral processes were made in Nissl-counterstainedsections. Drawings were made using digitized microscopicimages (JVC-TK C1381 and GrabitPCI grabber; SIS,Munster, Germany) in the PC software Designer 3.1 (Mi-crografx). Quantitative determinations of soma size, num-ber, and distribution of the labeled cells were performedon digitized images with the help of an image-analysissystem (analySIS 3.0 Doku; SIS). Statistical analysis wasperformed with the help of the PC-based statistical pro-gram Statistica (StatSoft, Tulsa, OK).

The number of retrogradely labeled layer 13 somata wasestimated within the ipsilateral as well as contralateraltectum along a rostrocaudal extent of 2.8 mm (A1.5–A4.3;Karten and Hodos, 1967) by counting fluorescein-labeledCtB-filled cells in every tenth section at �450 magnifica-tion. The amount of fluorescein/TDA double-labeled cellswas determined by manual marking fluorescein-labeledcells in digital images and by a projection of these marksonto the corresponding digital image with Texas red label.We were interested only in an estimation of the number ofdouble-labeled cells relative to the number of cells withsole fluorescein label. Therefore, no correction procedureswere used. Photographic documentation was carried outwith a digital camera system (Zeiss Axiocam) attached tothe microscope. Images were processed with Zeiss Axiovi-sion 3.0 and Photoshop 5.5 software. Color balance, con-trast, and brightness were adjusted to a variable extent toachieve satisfying output results with the Fuji “Media-Lab” printer device.

RESULTS

Tectopontine system

CtB injections into the lateral tegmentum (n � 5; Fig.1A) labeling the tectopontine pathway resulted in tracerspread along rostrocaudal levels between A1.50 andA3.75. Overall, the number of labeled cells, counted inevery tenth section, varied between 249 and 3,387 (mean1,788, SD 1,389). Despite these large quantitative differ-ences, the tectal labeling pattern was very similar in allcases and differed substantially from the pattern of me-dial tegmental injections. Retrogradely labeled neuronswere located primarily within the ipsilateral tectum, with


a mean of 77% (SD 22.75%) of all labeled cells located onthe ipsilateral side. Within the ipsilateral tectum, cellbodies formed two rows. The outer row consisted of cellslocated within layers 6–10 (Figs. 2, 3B). Most of theirsomata were located within the outer part of layer 10.Most of these cells, located exclusively within the ipsilat-eral tectum (Fig. 3), exhibited neurites of probably den-dritic nature, which entered the outer retinorecipient lam-inae and bifurcated extensively within layers 2–5a (Fig.4). Many superficial layer 10 cells as well as most cellslocated within layers 6–9 exhibited a single or two maindendrites, which were directed toward the outer retinore-cipient layers. Ramification widths of single radial den-drites ranged between 50 and 100 �m within layers 3–4.In the case of two radially oriented processes, interden-drite distances within layer 5b were between 20 and 100�m (Fig. 4B). Radial cells also exhibited smaller dendritesthat originated at the medial or lateral soma and took alateral to medial course. These processes could be followedto tectal layers 9–11. In contrast to radial cells, somatalocated in central to deep layer 10 often exhibited thickhorizontal dendrites, which could be followed for up to 300�m in lateral extent. Horizontal layer 10 dendrites irreg-

ularly gave rise to radial processes that could be followedto retinorecipient layer 3 (Fig. 4B).

Although only few labeled cells were visible in layers 11and 12, high numbers of layer 13 neurons were labeled(Figs. 2, 3). These multipolar cells were located through-out peripheral to deep sublayers. Horizontally orienteddendritic ramifications of these cells could be followed forup to 500 �m within layer 13. High densities of CtB-filledprocesses were also present in layer 12. Within layer 11,only few, thin fibers could be detected, and these processescould often be identified as dendritic ramifications of layer10 cells, suggesting that dendritic processes of multipolarlayer 13 cells rarely reached higher than tectal layer 12.Within the dorsal tectum, a considerable number of mul-tipolar cells also extended into layer 14. From these large(up to 55 �m diameter), fusiform to round cells, dendriticprocesses were directed in both horizontal and radial di-rections (Fig. 5). Smaller cells with ovoid somata (diame-ter around 20 �m) were also labeled in periventriculartectal layer 15. Both layer 14 and layer 15 cells wererestricted to tectal regions that overlaid the tectal ventri-cle. In the ventrolateral tectum, the inner band of tectalsomata was restricted to layer 13 cells. Dorsally, layer 14

Fig. 2. Photomicrographs of both tectal hemispheres after CtBinjections into the right lateral tegmentum, labeling the tectopontinepathway (ITP). The ipsilateral tectum (A) contained the majority oflabeled cells roughly clustered within an inner and an outer row of

neurons. Within the contralateral tectum (B), labeled cells were con-fined to deep layers 12 and 13. Ico, nucleus intercollicularis; MLd,nucleus mesencephalis lateralis, pars dorsalis; MLv, nucleus mesen-cephalis lateralis, pars ventralis. Scale bar � 1,000 �m.


and 15 cells passed over into the nucleus intercollicularis,whereas, ventrally, cells contributed to nucleus mesen-cephalis lateralis, pars ventralis (Fig. 2).

A minority of labeled cells (22.9%, SD 22.7) was locatedwithin the contralateral tectum (Figs. 2, 3A). The labelingpattern differed from that of the ipsilateral side, in that allcell bodies were confined to layers 12–15. Cells exhibitedthe same features (distribution, shape, extent) as neuronswithin the inner row of the ipsilateral tectum (Fig. 3A,B).Possibly because of the fact that an outer row of cells wasmissing, labeled processes were absent superficially tolayer 11. In the ipsilateral as well as in the contralateraltectum, the number of back-filled somata was highest inthe dorsal one-third of the optic tectum and lowest withinthe ventral segment (Friedman ANOVA: df 8.00, P � .05;Fig. 6A).

Tectobulbar system

CtB injections into the ventromedial tegmentum (n � 8;Fig. 1B) labeling the tectobulbar pathway resulted intracer spread along rostrocaudal levels between A1.75 andA4.00. The tectobulbar fibers crossed within the decussa-tio brachiorum conjunctivorum (DBC; Fig. 1B), but virtu-ally no fibers could be detected descending within thecontralateral tegmentum, supporting the specificity of ourmedial injections. The number of labeled cells, counted inevery tenth section, varied between 300 and 5,818 (mean2,317, SD 2,559). Although the amount of retrogradelyfilled cells exhibited considerable variation, the resultingqualitative labeling pattern did not. Medial tegmentalinjections labeled cells in deep tectal layers 12–15 within

both hemispheres. An outer row of cells, as labeled afterlateral tegmental CtB injections, was virtually absent. Incontrast to the case with lateral tegmental injections, thecontralateral tectum contained the majority of labeledcells, with a mean of 59.4% (SD 22.7%) of all labeled cellslocated on the contralateral side; Fig. 7). Within the ven-tral tectum, most cells were present within layer 13 (Figs.7, 8). Additionally, few cells were labeled in layer 12.These cells exhibited the same morphological features aslayer 13 cells, with multipolar fusiform to round somata(diameter 22.8 �m, SD 6.0). Processes of labeled somataoften ramified in horizontal to lateral directions. Thesepresumably dendritic processes could be followed for up to350 �m in lateral/horizontal directions, and they typicallydid not ascend more superficially than to tectal layer 11(Fig. 8A,C). Only five cells were detected to exhibit thickradial processes reaching into tectal layer 2 (Fig. 9). So-mata located in layer 12 exhibited roughly the same mor-phological features as layer 13 neurons and may, there-fore, be regarded as displaced layer 13 cells. Within layer14, labeled cells of both hemispheres were often smaller(19.9 �m diameter, SD 9.4) and showed stronger morpho-logical variation, with multipolar round to extremely fusi-form shapes with the long axis oriented both radially andhorizontally (Fig. 8D). Additionally, few cells were labeledin periventricular layer 15. As with lateral tegmental CtBinjections, the number of retrogradely labeled somata washighest in the dorsal one-third, significantly decreasingfrom the dorsal to ventral one-thirds of the ipsi- as well asthe contralateral tectum (Friedman ANOVA: df 9.750, P �.01; Fig. 6B).

Fig. 3. Photomicrographs of the contralateral (A) and ipsilateral(B) tectum after labeling of the tectopontine pathway. Within theipsilateral tectum, cells of the outer row (layers 7–10) frequentlyexhibited dendrites ascending into the outer retinorecipient layers. In

contrast, processes of neurons located in deeper layers 12–15 did notreach higher than layer 11. This was also apparent for all neuronslocated in the contralateral tectum. Scale bar � 200 �m.


In two animals, injections into the rostromedial tegmen-tum at the level of the nucleus ruber and nucleus inter-stitialis (Cajal) were performed (Figs. 1C, 10). Both injec-tions labeled high numbers of tectal cells (1,497 and 2,097

somata counted in every tenth section). In both animals,neurons were only faintly labeled; the CtB reaction prod-uct was limited to the cell bodies and thick proximalprocesses. As with more caudally situated medial tegmen-tal injections, the overwhelming majority of cells was lo-cated in inner tectal layers 12–15. Compared with ventro-medial tracer injections, two major variations wereobserved: 1) Most CtB filled cells were located in theipsilateral (88.8% and 93.2%) and not the contralateralhemisphere and 2) in the lateral and especially the ventraltectum, minor numbers of cell bodies (5.9% and 9.1% of allneurons) were also filled in layer 10 or 11. No labeling wasobserved in the outer retinorecipient layers. However, be-cause of the poor intensity of the labeling, we cannotexclude dendritic ramifications within these layers. Infact, the fusiform and radially oriented shape of somelayer 10/11 cell bodies points to this possibility. As in allother cases, cells of inner layers 14 and 15 were nearlyrestricted to the dorsal tectum, with these cell bands lo-cated in a periventricular position and passed over tonuclei intercollicularis and mesencephalis lateralis, parsventralis.

Double-labeling experiments

To examine whether tectal output neurons contribute toboth the ascending and the descending tectofugal path-ways, a second fluorescent tracer (TDA) was injected intothe central thalamus to label the nucleus rotundus (Fig.1D). In nearly all animals, TDA labeled bilaterally highnumbers of tectal cells within tectal layer 13 (Fig. 11).Additionally, in some cases, a band of cells within deeptectal layer 10 was filled within the ipsilateral tectum.These cells exhibited radial processes that typically ex-tended to the superficial retinorecipient laminae. Insofaras labeling of these cells was always associated with sub-stantial TDA spread into the dorsal thalamus, these cellspresumably represent the ascending output that inner-vates the nucleus geniculatus lateralis, pars dorsalis(GLd; Wild, 1989).

In all animals that exhibited substantial labeling ofboth the ascending (TDA) and the descending (CtB) tectal

Fig. 4. Higher magnification of cells labeled after lateral tegmen-tal CtB injections labeling the tectopontine pathway. A: CtB filledcells of the outer row located in layers 8–10. Whereas the moresuperficially situated neurons often formed radial processes, primarydendrites of cells located in layer 10 were often horizontally oriented.Secondary, radially to tangentially oriented ramifications were ob-served to arise from these horizontal arbors (arrow). B: Higher mag-nification of the outer tectal layers. Dendritic ramifications with prob-ably postsynaptic swellings were located in layers 4 and 3 (arrow). Acorresponding, cresyl violet-counterstained section is attached to al-low better visualization of tectal lamination (bottom numbers). Scalebars � 100 �m in A, 50 �m in B.

Fig. 5. Higher magnification of layers 12–15 of the ipsilateraltectum after lateral tegmental CtB injections labeling the tectopon-tine system. Scale bar � 100 �m.


projections, the amount of double-labeled cells was deter-mined. After all tegmental injections, the proportion ofdouble-labeled cells (relative to the number of CtB-filledcells) was extremely low, 0.38% (n � 7, SD 0.55) for medialtegmental and 0.26% (n � 4, SD 0.12) for lateral tegmen-tal injections. The only animal with a rostral tegmentalCtB injection combined with thalamic TDA injections ex-hibited 0.33% double-labeled cells. Thus, the data stronglysuggest that the tectal cell populations that contribute tothe ascending thalamic and descending tegmental projec-tions are nearly completely segregated, although the cellbodies are located at least to some degree in the sametectal layers.

DISCUSSION

The present paper describes the detailed morphology oftectal output neurons that establish the descending con-tralateral tectobulbar (CTB) and the ipsilateral tectopon-tine (ITP) pathways in pigeons. These systems differ interms of 1) their relative amounts of ipsilateral vs. con-tralateral projections, 2) the location of the efferent cellbodies within different tectal layers, and 3) their differen-tial access to visual input via dendritic ramificationswithin the outer retinorecipient tectal layers. Despite anoverlap in the laminar location with ascending cells, dou-ble labeling showed a virtually complete separation ofascending and descending projection neurons.

The use of the highly sensitive tracer choleratoxin dem-onstrated a higher portion of bilateral projections thanhad been previously demonstrated (Karten and Revzin,1966; Reiner and Karten, 1982) and revealed a detailedpicture of the axonal and dendritic branching pattern.This allows us to compare the morphology of descendingcells with that of the well-described ascending neurons(Karten et al., 1997; Luksch et al., 1998; Hellmann andGunturkun, 2001). In principle, two main groups of de-scending tectal efferents can be distinguished based onthe laminar localization of their cell bodies. Whereas thetectopontine pathway emerges from two cellular popula-tions located within layers 6–12 and layers 13–15, thetectobulbar tract arises virtually exclusively from layers13–15. According to the dendritic arborization pattern, thetwo descending pathways receive variable visual input,which differs from the input conveyed to ascending cells.The differential access to retinal input points to the pro-cessing of different aspects of the visual environment andhence corresponds to the differential role in visuomotorbehavior. Studies in a wide range of species suggest thatthe crossed tectobulbar tract is involved in approach andorientation toward a novel, desired object, whereas theipsilateral tectopontine pathway participates in move-ments away from obstacles or aversive stimuli (Ingle,1983; Dean et al., 1988; Ellard and Goodale, 1988). Asoutlined in detail below, the cells of the descending sys-tems are not overlapping with those constituting othertectal projection streams.

The layer 6–10 system

After lateral tegmental CtB injections labeling the tec-topontine pathway, an outer band of cells was detected,which combines somata located within ipsilateral layers6–10. Only in these cases was a labeling of distal cellprocesses within retinorecipient layers 2–5a detected.Most of these processes could be directly verified as den-drites of pyramidal cells of layers 8–10 or multipolar cellswithin layers 9 and 10. Based on the extent of their den-dritic ramifications over the tectal surface, these twogroups could be distinguished as “widefield” multipolarlayer 10 cells vs. “narrowfield” pyramidal layer 8–10 cells.Individual ramifications of the “widefield” neurons typi-cally could be followed for up to 300 �m in the lateraldirection. This dendritic pattern is comparable to that ofascending multipolar “type I” layer 13 neurons labeledafter CtB injections into the rotundus (Karten et al., 1997;Luksch et al., 1998; Hellmann and Gunturkun, 2001).However, whereas the typical bottlebrush endings of thesecells arborize within layer 5b, the dendrites of descendingcells tend to branch in layers 2–5a.

Fig. 6. Mean percentage of retrogradely labeled cells in the dorsal,lateral, and ventral tectal segments after lateral (A) or medial (B)tegmental CtB injections. In both experimental groups, the proportionof labeled cells differed significantly between tectal segments. Thefrontal plane of the tectal sections was subdivided into three segmentsaccording to Manns and Gunturkun (1997).


Although the multipolar layer 10 neurons share somedendritic features of tectal neurons that establish theascending tectorotundal pathway in layer 13, the pyrami-dal cells of layers 8–10 do definitely not. Instead, theydisplay characteristics comparable to those of radial effer-ent neurons in layers 8–11 (Hunt and Brecha, 1984).These latter cells project onto the isthmic nuclei nucleusisthmoopticus (ION; Crossland and Hughes, 1978; Uchi-yama and Watanabe, 1985; Woodson et al., 1991); nucleusisthmi, pars parvocellularis (Ipc; Hunt et al., 1977); andnucleus isthmi, pars semilunaris (SLu; Hellmann et al.,2001) and onto the thalamic nucleus geniculatus lateralis,pars ventralis (GLv; Hunt and Kunzle, 1976; Crosslandand Uchwat, 1979) and GLd (Gamlin and Cohen, 1986;Wild, 1989). A morphological comparison reveals that allthese cells represent cell classes that differ from the de-scending tectopontine neurons.

Ipc-projecting neurons exhibit a single spiny apical den-drite arborizing within layers 3, 5, and 9. These cells arecharacterized by an axon arising from the apical processas a shepherd’s crook (Hunt et al., 1977; Woodson et al.,1991). Thus, these cells differ from the near-field ITP-projecting cells displaying dendritic arborizations in lay-ers 2–5a.

In contrast, the apical dendritic arborizations and thelocation of the somata at the border between layers 9 and 10suggest a good match between ITP- and ION-projecting cells(Uchiyama and Watanabe, 1985; Woodson et al., 1991; Uch-iyama et al., 1996). However, because the ION-projectingcells do not possess ascending dendrites extending into su-perficial layers 2–7 and because ITP near-field neurons lackthe extensive basal dendritic ramifications within tectal lay-ers 12–14, which are typical for ION afferents (Uchiyama etal., 1996), ITP- and ION-projecting neurons probably like-wise represent different cell populations.

As with the ITP-projecting cells, GLd afferents wereshown to be located primarily within layers 6–10 of thedorsal tectum (Gamlin and Cohen, 1986; Wild, 1989). Ourrotundal TDA injections often resulted in considerabletracer spread within the dorsally located GLd and henceretrogradely labeled its tectal input. Compared with theITP-projecting cells, these radial neurons were locatedslightly more at the outer surface of layer 10. Additionally,double labeling of ascending and descending projectionsrevealed virtually no double-labeled cells. Thus, we candefinitely exclude a common origin for tectal ITP and GLdafferents. Comparable assertions regarding whether GLvand ITP projections result from axon collaterals of a com-

Fig. 7. Photomicrographs of the contralateral (A) and ipsilateral (B) tectal halves after CtB injectionsinto the right ventromedial tegmentum, labeling the tectobulbar pathway (CTB). Neuronal somata werelocated predominantly in the contralateral hemisphere. On both sides, labeled cells were confined to deeplayers 12 and 15. Scale bar � 1,000 �m.


mon pool of tectal near-field neurons were not possiblewith our experiments.

Our results present a picture of a highly differentiatedsystem of tectal efferents within the intermediate layers:Although tectal projections onto GLv, GLd, IPC, ION, andthe ITP system all arise from the same tectal layer, eachtarget region seems to be innervated by a unique set of layer

10 neurons. This is further supported by a multiple sublay-ering of this major source of tectal output detected by meansof different calcium-binding proteins (unpublished data).

The layer 13–15 system

Our results revealed a substantial number of tegmen-tally descending cells to be located in the deep tectal

Fig. 8. Higher magnification of retrogradely labeled cells after medial tegmental injections labelingthe tectobulbar pathway: A and B depict all ipsi- and contralateral tectal layers. C and D show labeledcells in deep tectal layers 12–15 in more detail. On both sides, dendritic processes of the multipolarneurons were restricted to layers 12–14. Scale bars � 200 �m.


layers 13–15. Cells located in layer 13 were intermingledwith neurons constituting the exclusive source of tectoro-tundal projection, which is the major ascending visualprocessing stream in birds (Benowitz and Karten, 1976;Karten et al., 1997; Deng and Rogers, 1998; Hellmann andGunturkun, 1999, 2001). Tectorotundal and descendingcell populations exhibit a considerable difference in theirdendritic ramification patterns. After tegmental CtB in-jections, ascending dendritic processes did not reachhigher than up to layer 12. This is different from thelabeling pattern of most ascending layer 13 neurons.These cells are differentiated into at least five subtypescharacterized by dendritic arborizations within differenttectal laminae (Hellmann and Gunturkun, 2001) and by

projection to functionally specified subregions of the nu-cleus rotundus (Wang and Frost, 1992; Wang et al., 1993;Laverghetta and Shimizu, 1999; Hellmann and Gun-turkun, 1999, 2001). The overwhelming majority of rotun-dal projecting neurons has direct access to retinal input,with dendrites reaching up to retinorecipient layers 3–5.However, one of the ascending neurons (type II) displaysdendrites that do not ramify within the retinorecipientlayers. Insofar as the dendrites of the descending layer 13cells are comparably confined to the deep layers, the even-tual presence of bidirectionally projecting cells could belimited to these type II cells. Type II cells constitute onlya minor portion of the ascending tectofugal system andterminate in the nucleus triangularis, a dorsomedial ex-tension of the rotundus (Karten et al., 1997; Hellmannand Gunturkun, 2001). However, our double-labeling datashow that neither of the two descending systems overlapsin important numbers with ascending type II cells. Thedeep tectal layers 14/15 also contribute substantially toboth descending pathways. These multipolar cells sharedendritic features with layer 13 neurons but tend to havesmaller somata. However, these cells are restricted to thedorsal tectum medially, passing over into the nucleusintercollicularis. Because their dendrites do not ascendhigher than to layer 12, neither population receives directretinal input, suggesting that they convey indirect visualand/or multimodal information. It is not known whetherthese cells represent a functionally specialized subtype ofthe descending cell population different from layer 13neurons. Within the deep layers 14 and 15, ascendingtectorotundal cells are virtually absent (Hellmann andGunturkun, 2001). Thus, the deepest layers constitute

Fig. 9. After medial tegmental CtB injections labeling the tectob-ulbar pathway, neurons with thick radial processes directed to theouter tectal layers 2–4 were only rarely labeled. Scale bar � 200 �m.

Fig. 10. Photomicrographs through all layers of the lateral contralateral (A) and ipsilateral (B)tectum after CtB injections into the rostromedial mesencephalic tegmentum. Most cells were located inlayers 12–15 of the ipsilateral tectum; only few cell were labeled within layers 10–11 of the lateral andventral tectum. Scale bar � 200 �m.


Fig. 11. A: Photomicrographs of layers 8–14 of the right tectum afterbilateral TDA injections into the central thalamus and CtB injectionsinto the right lateral tegmentum. Red-labeled cells are those with as-cending thalamic projections, and green label indicates projections to thebrainstem. Although red and green cells intermingled within deep tectallayer 13, only very few cells contained both tracers (arrows). Numbersindicate location of tectal layers. B–D: Higher magnification of the boxedarea in A. The yellow and hence double-labeled cell in B (arrow) was

visible with the Texas red (C) as well as the fluorescein (D) filter set.E: Photomicrograph of layers 5–15 of the right tectal hemisphere afterTDA injections into the central thalamus (bilaterally) and CtB injectionsinto the left ventromedial tegmentum. Red-labeled cells are those withascending thalamic projections, and green-labeled cells project to thebrainstem. Although red and green cells intermingled within deep tectallayer 13, no double-labeled cell could be detected. Scale bars � 200 �m inA,E, 50 �m in B–D.


purely descending projections, which is comparable to thepattern observed in other vertebrates (Reiner, 1994).

Three conclusions follow from these observations. First,the avian tectum hosts multiple descending cell types thatare different from the systems giving rise to the ascendingstreams. Second, because of this fact, at least one furthertectal layer 13 cell type with descending projections has tobe added to the list of ascending tectorotundal cell types inthis lamina. Third, insofar as the ascending and the de-scending information streams are remarkably indepen-dent from each other, they seem to constitute information-ally encapsulated modules as postulated by Fodor (1983).By this line of reasoning, the multiple tectal cellular sys-tems might operate strictly in parallel with and indepen-dent from each other as special-purpose input–outputmodules with prespecified functions.

Visual field-dependent organization

The present study revealed that, apart from differencesin the proportion of ipsi- and contralateral populations,the descending cells are heterogeneously distributed alongthe tectal dorsoventral axis. Whereas tectobulbar projec-tions in the mallard and chick originate predominantlyfrom the rostroventral tectum (Shepherd and Taylor,1995; Tellegen et al., 1998), our lateral and medial teg-mental tracer injections primarily labeled cells in the dor-sal tectal segment. This overrepresentation of the dorsaltectum within the descending streams was clearly visible,although the ventral tectum has about twice as manyneurons as the dorsal tectum (Theiss et al., 1996). Thisdistribution is reversed from that of ascending tectorotun-dal cells, where, with the exception of type II, significantlymore projections arise from the ventral tectum (Hellmannand Gunturkun, 1999, 2001). Because the tectum is reti-notopically organized with the upper visual field repre-sented in the dorsal and the lower and frontal visual fieldin the ventral tectum (Clarke and Whitteridge, 1976;Hayes et al., 1987; Remy and Gunturkun, 1991), ascend-ing and descending tectofugal projections process informa-tion mainly from the lower and the upper visual fields,respectively. The predominance of ventral tectal projec-tions to the rotundus can explain why rotundal lesionslead to acuity impairments only in the lower frontal fieldof view (Gunturkun and Hahmann, 1999). Similarly, ourdata suggest the descending system to be specialized tothe upper visual field, represented in the dorsal tectum.The deep tectal layers receive input from the nucleusspiriformis lateralis (SpL), which conveys basal gangliainfluence to the tectum and hence participates in fore-brain control of the tectal premotor output (Reiner et al.,1982a). SpL is characterized by biochemical subdifferen-tiations (Reiner et al., 1982b), and the projection onto thetectum displays a rough dorsoventral topography (Reineret al., 1982a; Hellmann et al., 2001), indicating a visualfield-dependent modulation of premotor control. Thus, ret-inal information from the upper and lateral visual fieldseems to be preferentially analyzed with respect to pre-motor control, so the dorsal tectum might be specialized toallow reactions to objects emerging within this field ofview.

Segregation of ascending and descendingprojections

In comparing the efferent tectal cell types in tetrapods,the cells differ in 1) their receipt to retinal input, 2) the

localization of the cell bodies, and 3) the degree of axonalarborizations to different targets. During evolution, atrend for the segregation of all three features seems toexist. Whereas in salamanders all efferent subtypes re-ceive direct retinal input (Dicke, 1999; Roth et al., 1999),in amniotes the efferent cellular populations are segre-gated into cells receiving monosynaptic or polysynapticretinal input. This is accompanied by a different degree towhich the efferent cell bodies migrate into superficial lay-ers during development. Although ascending and descend-ing efferent cells are intermingled in the tectum of am-phibians, these cells exhibit a high laminar separation inthe mammalian superior colliculus. Different degrees ofseparation can be observed in sauropsids, with ascendingand descending cells partially located within the samelayer (Reiner, 1994; Belekhova et al., 2003). However,despite the different laminar organization patterns of themammalian superior colliculus and the optic tectum ofsauropsids, the spatial relations between the efferent pro-jection systems are very similar. In contrast to the case forthe avian tectum, retinal input enters the mammaliancolliculus via the third layer from the surface, the stratumopticum. The layer superior to the stratum opticum (stra-tum griseum superficiale) tends to have ascending projec-tions, although some cells located inferior to the stratumopticum also give rise to projections to several thalamicnuclei (Linke et al, 1999). Whereas neurons projecting tothe GLd are located within the upper sublayers of thestratum griseum superficiale, neurons projecting to thelateral posterior/pulvinar nucleus are located within thelower stratum griseum superficiale or within the stratumopticum (for review see Reiner, 1994), whereby their as-cending dendrites reach to the outermost sublayers (Majoret al., 2000). Thus, this organization resembles the pat-tern present in the avian tectum, where GLd-projectingcells are located more superficially than tectorotundalneurons, which likewise send ascending dendrites up tothe outer retinorecipient layers. Comparably to the tec-topontine and tectobulbar tracts of the avian tectum, thepremotor systems of the mammalian colliculus are segre-gated into a population descending within the tectopon-tine tract or within the predorsal bundle. Although mostcells are located below the stratum opticum, the tectopon-tine tract combines two cellular populations. The ipsilat-erally projecting cells are located within the stratum gri-seum superficiale or stratum opticum, and thecontralaterally projecting neurons are localized below thestratum opticum within the deep laminae (for review seeReiner, 1994). Again, this pattern is very similar to thearrangement of the avian tectal descending system, with asuperficial and a deep tectopontine population and a solelydeep tectobulbar population.

It is not clear how closely the laminar segregation oftectal cells is related to a separation of axons projecting todifferent targets. In salamanders, single cells display axoncollaterals to both ascending and descending fiber bundles(Dicke, 1999), and, in the turtle’s tectum, some tectorotun-dal cells display side branches to the GLd (Belekhova etal., 2003). The present data show that, in pigeons, ascend-ing and descending cells located within the same tectallayer represent completely separated cell populations.This parcellation might be related to the functional role ofthe tectum in visuomotor processing. Visuomotor integra-tion might be at least in part directly obtained by efferentcells in salamanders, but, in more highly developed sys-


tems, comparable functions seem to be mediated by direct orindirect contacts between sensory and motor neurons. Inmammals, the transmission of visual signals from the super-ficial to the deep tectal layers is mediated by interlaminarconnections (Mooney et al., 1988b; Isa, 2002; Doubell et al.,2003). In the avian tectum, reafferent circuits might sub-serve comparable functions. Apart from tectal interneurons,specialized nuclei are likely involved in such circuits. Twocandidates are the nucleus isthmi pars semilunaris (SLu), acomponent of the isthmic complex, and the lateral spiriformnucleus (SpL). SLu has reciprocal input with the tectum andprojections onto the rotundus and to the SpL (Hellmann etal., 2001). This connectivity pattern suggests that SLu plays

a key role in the topographically organized modulation of theascending projections and, indirectly, in cooperation with theSpL, also the descending projections of the tectum (Hell-mann et. al., 2001). Because SpL conveys descending fore-brain information to the deeper tectal layers, it is assumed torelay information regarding the bird’s ongoing and impend-ing movements in order to coordinate the bird’s body positionwith the location of objects in the bird’s visual space (Reineret al., 1982a).

The tectal mosaic

The present study shows that the tectum is composed ofmultiple cell types that constitute the diverse efferent

Fig. 12. Schematic outline of the tectal mosaic hypothesis. Wepropose that multiple cell types with diverse visual inputs at aboutthe same location of the tectal map constitute the tectal projectionsonto diverse areas. These projections are both retinotopically orga-nized and functionally specific. A pigeon brain with a highlightedoptic tectum is represented at upper left. An “unfolded” tectum with atwo-dimensional map of the tectal surface is shown in the center (theprocess of unfolding is described by Remy and Gunturkun, 1991).Small circles in different colors represent cells on the tectal map thathave descending projections within the tectopontine (TP, in yellow)

and the tectobulbar systems (TB, in orange) or ascending ones withinthe tectorotundal projection (cell types I–V). Each of these cell typesprojects to an area with the same color code. At upper right, a sche-matic cross-section of the optic tectum is shown, with retinorecipientlayers 2–7 depicted in dark gray. Tectal neurons with descendingprojections to the brainstem and ascending projections to the rotun-dus are shown with their main dendritic bifurcations in the superficialtectal strata allowing a cell type-specific organization of visual input.The complete arrangement explains how functionally specific tectalprojections arise from a retinotopically arranged organization.


systems radiating from the midbrain. These efferent celltypes can be distinguished from each other by severalmorphological criteria, such as laminar position, dendriticarrangement, and target structures. Because tectal lami-nae differ in their visual and nonvisual input (Miceli et al.,1987; Yamagata and Sanes, 1995; Karten et al., 1997), weassume that the cell type-specific dendritic bifurcationswithin only some of the tectal layers produce idiosyncraticmixtures of afferents and thus differential functionalmodes for each of the diverse tectal efferent systems. Atthe same time, each of these efferent cells is embeddedinto the retinotopic arrangement of the tectal surface,probably producing a retinotopic arrangement within alltectofugal systems. Thus, each point of the tectal spacemap hosts multiple cell types and the origin of multipleindependent projections onto different diencephalic andbrainstem structures subserving diverse functions. Thismosaic-like architecture, schematically outlined in Figure12, was first proposed by Hellmann and Gunturkun (2001)as a solution to the problem of how the retinotopic tectalmap could be transformed into a functionotopic arrange-ment within the nucleus rotundus (Wang and Frost, 1992;Wang et al., 1993; Laverghetta and Shimizu, 1999). Re-cently, Marın et al. (2003) could verify this proposal in afurther detailed analysis of the tectorotundal projection.

The present data extend this concept of the tectal mo-saic. Both the tectopontine and the tectobulbar pathwaysare heterogeneous with respect to the cell types involved,both originate from the complete extent of the tectal sur-face (albeit predominantly dorsal), and both pathwaysgive rise to terminal branches at different levels of thepremotor reticular formation (Reiner and Karten, 1982;Tellegren et al., 1998). At the same time, the crossedtectobulbar tract is involved in approach and orientationtoward desired objects, whereas the ipsilateral tectopon-tine pathway guides movements away from aversive stim-uli (Ingle, 1983; Ellard and Goodale, 1988; Dean et al.,1988). Thus, the principle of a transformation from reti-notopic to functionotopic coding as suggested for the tect-ofugal pathway could apply in exactly the same way forthe descending tectal projections. It is even likely thatthese descending projections can be further subdividedwithin the same framework. For example, injections intothe region of the nucleus interstitialis (Cajal) evinced la-beled cells that are located only in layers 11–15. Thus,these cells intermingle with other neurons descending inthe tectobulbar tract but vary from them in their predom-inance in the ipsilateral tectum and in their laminar lo-cation superficial to layer 13.

Taken together, our concept of the tectal mosaic positsthat, by virtue of multiple cell types with diverse inputfunctions at the same location of the tectal map, the pro-jections of the optic tectum onto diverse areas can be bothretinotopically organized and functionally specific. Insofaras this arrangement probably applies both for the ascend-ing tectorotundal and for the descending tectomesence-phalic pathways, it could represent a basic feature of theoptic tectum that also defines the architecture of the othertectal efferents.

LITERATURE CITED

Abramson BP, Chalupa LM. 1988. Multiple pathways from the superiorcolliculus to the extrageniculate visual thalamus of the cat. J CompNeurol 271:397–418.

Adams JC. 1981. Heavy metal intensification of DAB-based HRP reactionproduct. J Histochem Cytochem 29:775.

Beck PD, Kaas JH. 1998. Thalamic connections of the dorsomedial visualarea in primates. J Comp Neurol 396:381–398.

Belekhova M, Kenigfest N, Rio J-P, Reperant J, Ward R, Vesselkin N,Karamian O. 2003. Tectothalamic visual projections in turtles: theircells of origin revealed by tracing methods. J Comp Neurol 457:37–56.

Benowitz LI, Karten HJ. 1976. Organization of tectofugal visual pathwayin pigeon: retrograde transport study. J Comp Neurol 167:503–520.

Cajal SR. 1911. Histologie du systeme nerveux de l’homme et des verte-bres. Paris: Maloine.

Clarke PG, Whitteridge D. 1976. The projection of the retina, including the“red area” on the optic tectum of the pigeon. Q J Exp Physiol 61:351–358.

Crossland WJ, Hughes CP. 1978. Observations on the afferent and efferentconnections of the avian isthmo-optic nucleus. Brain Res 145:239–256.

Crossland WJ, Uchwat CJ. 1979. Topographic projections of the retina andoptic tectum upon the ventral lateral geniculate nucleus in the chick.J Comp Neurol 185:87–106.

Davila JC, Andreu MJ, Real MA, Puelles L, Guirado S. 2002. Mesence-phalic and diencephalic afferent connections to the thalamic nucleusrotundus in the lizard Psammodromus algirus. Eur J Neurosci 16:267–282.

Dean P, Redgrave P, Mitchell IJ. 1988. Organization of efferent projectionsfrom superior colliculus to brainstem in rat: evidence for functionaloutput channels. Prog Brain Res 75:27–36.

Deng C, Rogers LJ. 1998. Organization of the tectorotundal and SP/IPS-rotundal projections in the chick. J Comp Neurol 394:171–185.

Dicke U. 1999. Morphology, axonal projection pattern, and response typesof tectal neurons in plethodontid salamanders. I: tracer study of pro-jection neurons and their pathways. J Comp Neurol 404:473–488.

Dicke U, Roth G. 1996. Similarities and differences in the cytoarchitectureof the tectum of frogs and salamanders. Acta Biol Hung 47:41–59.

Doubell TP, Skaliora J, Baron J, King AJ. 2003. Functional connectivitybetween superficial and deeper layers of the superior colliculus: ananatomical substrate for sensorimotor integration. J Neurosci 23:6596–6607.

Ellard CG, Goodale MA. 1988. A functional analysis of the collicular outputpathways: a dissociation of deficits following lesions of the dorsal teg-mental decussation and the ipsilateral collicular efferent bundle in theMongolian gerbil. Exp Brain Res 71:307–319.

Fodor J. 1983. The modularity of mind. Cambridge, MA: MIT Press.Gamlin PDR, Cohen DH. 1986. A second ascending visual pathway from

the optic tectum to the telencephalon in the pigeon (Columba livia).J Comp Neurol 250:296–310.

Guirado S, Devila JC, Real MA, Medina L. 2000. Light and electronmicroscopic evidence for projections thalamic nucleus rotundus to tar-gets in the basal ganglia, the dorsal ventricular ridge, and the amyg-daloid complex in a lizard. J Comp Neurol 424:216–232.

Gunturkun O, Hahmann U. 1999. Functional subdivisions of the ascendingvisual pathways in the pigeon. Behav Brain Res 98:193–201.

Hardy O, Leresch N, Jassik-Gerschenfeld D. 1984. Postsynaptic potentialsin neurons of the pigeon’s optic tectum in response to afferent stimu-lation from the retina and other visual structures. Brain Res 311:65–74.

Harrell JV, Caldwell RB, Mize RR. 1982. The superior colliculus neuronswhich project to the dorsal and ventral lateral geniculate nucleus in thecat. Exp Brain Res 46:234–242.

Harting JK, Huerta MF, Hashikawa T, van Lieshout DP. 1991. Projectionof the mammalian superior colliculus upon the dorsal lateral geniculatenucleus: organization of tectogeniculate pathways in nineteen species.J Comp Neurol 304:275–306.

Hayes BP, Hodos W, Holden AL, Low AL. 1987. The projection of the visualfield upon the retina of the pigeon. Vis Res 27:31–40.

Hellmann B, Gunturkun O. 1999. Visual-field-specific heterogeneitywithin the tecto-rotundal projection of the pigeon. Eur J Neurosci11:2635–2650.

Hellmann B, Gunturkun O. 2001. The structural organization of parallelinformation processing within the tectofugal visual system of the pi-geon. J Comp Neurol 429:94–112.

Hellmann B, Manns M, Gunturkun O. 2001. Nucleus isthmi, pars semilu-naris as a key component of the tectofugal visual system in pigeons.J Comp Neurol 436:153–166.

Hunt SP, Brecha N. 1984. The avian optic tectum: a synthesis of morphol-


ogy and biochemistry. In: Vanegans H, editor. Comparative neurologyof the optic tectum. New York: Plenum Press. p 619–648.

Hunt SP, Kunzle H. 1976. Observations on the projections and intrinsicorganization of the pigeon optic tectum: an autoradiographic studybased on anterograde and retrograde, axonal and dendritic flow.J Comp Neurol 170:153–172.

Ingle D. 1983. Brain mechanisms of visual localization by frogs and toads.In: Ewert JP, Capranica RR, Ingle D, editors. Advances in vertebrateneuroethology. New York: Plenum Press. p 177–226.

Isa T. 2002. Intrinsic processing in the mammalian superior colliculus.Curr Opin Neurobiol 12:668–677.

Karten HJ, Hodos W. 1967. A stereotaxic atlas of the brain of the pigeon.Baltimore: Johns Hopkins Press.

Karten HJ, Revzin AM. 1966. The afferent connections of the nucleusrotundus in the pigeon. Brain Res 2:368–377.

Karten HJ, Keyser KT, Brecha N. 1990. Biochemical and morphologicalheterogeneity of retinal ganglion cells. In: Cohen B, Bodis-Wollner I,editors. Vision and the brain. New York: Raven Press. p 19–33.

Karten HJ, Cox K, Mpodozis J. 1997. Two distinct populations of tectalneurons have unique connections within the retinotectorotundal path-way of the pigeon (Columba livia). J Comp Neurol 387:449–465.

Laverghetta AV, Shimizu T. 1999. Visual discrimination in the pigeon(Columba livia): effects of selective lesions of the nucleus rotundus.Neuroreport 10:981–985.

Laverghetta AV, Shimizu T. 2003. Organization of the ectostriatum basedon afferent connections in the zebra finch (Taeniopygia guttata). BrainRes 963:101–112.

Lazar G, Toth, P, Csank, G, Kicliter E. 1983. Morphology and location oftectal projection neurons in frogs: a study with HRP and cobalt-filling.J Comp Neurol 215:108–120.

Lin CS, Kaas JH. 1980. Projections from the medial nucleus of the inferiorpulvinar complex to the middle temporal area of the visual cortex.Neuroscience 5:2219–2128.

Linke B, De Lima AD, Schwegler H, Pape H-C. 1999. Direct synapticconnections of axons from superior colliculus with identified thalamo-amygdaloid projection neurons in the rat: possible substrates of sub-cortical visual pathway to the amygdala. J Comp Neurol 403:158–170.

Luksch H, Cox K, Karten HJ. 1998. Bottlebrush dendritic endings andlarge dendritic fields: motion-detecting neurons in the tectofugal path-way. J Comp Neurol 396:399–414.

Luksch H, Karten HJ, Kleinfeld D, Wessel R. 2001. Chattering and differ-ential signal processing in identified motion-sensitive neurons of par-allel visual pathways in the chick tectum. J Neurosci 21:6440–6446.

Major DE, Luksch H, Karten HJ. 2000. Bottlebrush dendritic endings andlarge dendritic fields: motion-detecting neurons in the mammaliantectum. J Comp Neurol 423:243–260.

Manns M, Gunturkun, O. 1997. Development of the retinotetcal system inthe pigeon: a cytoarchitectonic and tracing study with choleratoxin.Anat Embryol 195:539–555.

Marın G, Letelier JC, Henny P, Sentis E, Farfan G, Fredes F, Pohl N,Karten H, Mpodozis J. 2003. Spatial organization of the pigeon tecto-rotundal pathway: an interdigitating topographic arrangement.J Comp Neurol 458:361–380.

Martinez-Marcos A, Font C, Lanuza E, Martinez-Garcia F. 1998. Ascend-ing projections from the optic tectum in the lizard Podarcis hispanica.Vis Neurosci 15:459–475.

Masino T, Grobstein P. 1989. The organization of descending tectofugalpathways underlying orienting in the frog, Rana pipiens. I. Lateraliza-tion, parcellation, and intermediate spatial representation. Exp BrainRes 75:227–244.

Masino T, Knudsen EI. 1990. Horizontal and vertical components of headmovement are controlled by distinct neural circuits in the barn owl.Nature 345:434–437.

Miceli D, Reperant J, Villalobos J, Dionne L. 1987. Extratelencephalicprojections of the avian visual Wulst. A quantitative autoradiographicstudy in the pigeon Columba livia. J Hirnforsch 28:45–57.

Mooney RD, Nikoletseas MM, Hess PR, Allen Z, Lewin AC, Rhoades RW.1988a. The projection from the superficial to the deep layers of the

superior colliculus: an intracellular horseradish peroxidase injectionstudy in the hamster. J Neurosci 8:1384–1399.

Mooney RD, Nikoletseas MM, Ruiz SA, Rhoades RW. 1988b. Receptive-field properties and morphological characteristics of the superior col-licular neurons that project to the lateral posterior and dorsal lateralgeniculate nuclei in the hamster. J Neurophysiol 59:1333–1351.

Nixdorf BE, Bischof HJ. 1982. Afferent connections of the ectostriatum andvisual wulst in the zebra finch (Taeniopygia guttata castanotisGould)—an HRP study. Brain Res 248:9–17.

Ogawa T, Takimor T, Takahashi Y. 1985. Projection of morphologicallyidentified superior colliculus neurons to the lateral posterior nucleus inthe cat. Vis Res 25:329–337.

Reiner A. 1994. Laminar distribution of the cells of origin of ascending anddescending tectofugal pathways in turtles: implications for the evolu-tion of tectal lamination. Brain Behav Evol 43:254–292.

Reiner A, Karten H. 1982. Laminae distribution of the cells of origin of thedescending tectofugal pathways in the pigeon (Columba livia). J CompNeurol 204:165–187.

Reiner A, Brecha NC, Karten HJ. 1982a. Basal ganglia pathways to thetectum: the afferent and efferent connections of the lateral spiriformnucleus of pigeon. J Comp Neurol 208:16–36.

Reiner A, Brecha NC, Karten HJ. 1982b. Enkephalin-mediated basal gan-glia influences over the optic tectum: immunohistochemistry of thetectum and the lateral spiriform nucleus in pigeon. J Comp Neurol208:37–53.

Remy M, Gunturkun O. 1991. Retinal afferents to the tectum opticum andthe nucleus opticus principalis thalami in the pigeon. J Comp Neurol305: 57–70.

Roth G, Dicke U, Grunwald W. 1999. Morphology, axonal projection pat-tern, and response types of tectal neurons in plethodontidsalamanders. II: intracellular recording and labeling experiments.J Comp Neurol 404:489–504.

Shepherd IT, Taylor JS. 1995. Early development of efferent projectionsfrom the chick tectum. J Comp Neurol 354:501–510.

Shu, SY, Ju G, Fan L. 1988. The glucose oxidase-DAB-nickel method inperoxidase histochemistry of the nervous system. Neurosci Lett 85:169–171.

Tellegen AJ, Karssen AM, Rietveld TM, Dubbeldam JL. 1998. A crossedprojection from the optic tectum to craniocervical premotor areas in thebrainstem reticular formation. An anterograde and retrograde tracingstudy in the mallard (Anas plathyrhynchos L.). Eur J Morphol 36:276–243.

Theiss C, Hellmann B, Gunturkun O. 1998. The differential distribution ofAMPA-receptor subunits in the tectofugal system of the pigeon. BrainRes 787:114–128.

Uchiyama H, Watanabe M. 1985. Tectal neurons projecting to the isthmo-optic nucleus in the Japanese quail. Neurosci Lett 58:381–385.

Uchiyama H, Yamamoto N, Ito H. 1996. Tectal neurons that participate incentrifugal control of the quail retina: a morphological study by meansof retrograde labeling with biocytin. Vis Neurosci 13:1119–1127.

Wang Y, Frost BJ. 1992. Time to collision is signaled by neurons in thenucleus rotundus of pigeons. Nature 356:236–238.

Wang Y, Jiang S, Frost BJ. 1993. Visual processing in pigeon nucleusrotundus: luminance, color, motion, and looming subdivisions. Vis Neu-rosci 10:21–30.

Wild JM. 1989. Pretectal and tectal projections to the homologue of thedorsal lateral geniculate nucleus in the pigeon: an anterograde andretrograde tracing study with cholera toxin conjugated to horseradishperoxidase. Brain Res 479:130–137.

Woodson W, Reiner A, Anderson K, Karten HJ. 1991. Distribution, laminarlocation, and morphology of tectal neurons projecting to the isthmo-optic nucleus and nucleus isthmi, pars parvocellularis in the pigeon(Columba livia) and chick (Gallus domesticus): a retrograde labelingstudy. J Comp Neurol 305:470–488.

Yamagata M, Sanes JR. 1995. Target-independent diversification andtarget-specific projection of chemically defined retinal ganglion cellsubsets. Development 121:3763–3776.


Tectal Mosaic: Organization of the Descending Tectal ... JCN 2004.pdf · Descending Tectal...

Documents

Transcript of Tectal Mosaic: Organization of the Descending Tectal ... JCN 2004.pdf · Descending Tectal...