* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Retinal target cells of the centrifugal projection from the isthmo
Survey
Document related concepts
Multielectrode array wikipedia , lookup
Clinical neurochemistry wikipedia , lookup
Subventricular zone wikipedia , lookup
Molecular neuroscience wikipedia , lookup
Synaptic gating wikipedia , lookup
Stimulus (physiology) wikipedia , lookup
Neuropsychopharmacology wikipedia , lookup
Synaptogenesis wikipedia , lookup
Development of the nervous system wikipedia , lookup
Optogenetics wikipedia , lookup
Neuroanatomy wikipedia , lookup
Axon guidance wikipedia , lookup
Transcript
THE JOURNAL OF COMPARATIVE NEUROLOGY 476:146 –153 (2004) Retinal Target Cells of the Centrifugal Projection from the Isthmo-optic Nucleus HIROYUKI UCHIYAMA,* KENJI AOKI, SHINJI YONEZAWA, FUKU ARIMURA, HIROSHI OHNO Department of Information and Computer Science, Faculty of Engineering, Kagoshima University, Kagoshima 890-0065, Japan AND ABSTRACT Although the avian retina has long been known to receive projection from a midbrain nucleus, the isthmo-optic nucleus (ION), the output of its target cells has remained obscure. We labeled the isthmo-optic (IO) terminals in the Japanese quail retina, by using anterograde transport of fluorescent tracer injected into the ION, and then labeled target cells for these terminals by means of intracellular tracer injection under direct microscopic observation. Somata of the IO target cells (IOTCs) lie in the innermost zone of the inner nuclear layer of the ventral half of the retina and have no dendrites but an axon. The axons run in the inner plexiform layer (IPL) for up to 6 mm and terminate densely in a round or elliptical terminal field, about 90 –290 m in diameter, of the outermost zone of the IPL. Longer axons (⬎2 mm) extend dorsally, but shorter ones (⬍1 mm) project ventrally or horizontally, so the terminals are distributed widely in both dorsal and ventral halves of the retina. The IOTCs cannot be classified into any of the five conventional major classes of retinal cells, including amacrine cells, and are thought to be “slave” neurons whose output is controlled by the neurons in the brain. Topographic separation between input to and output from the IOTCs by the axons might be essential for the overall topographic organization of the centrifugal visual system in birds. J. Comp. Neurol. 476:146 –153, 2004. © 2004 Wiley-Liss, Inc. Indexing terms: retina; centrifugal visual system; quail; retinopetal projection; tectum The isthmo-optic nucleus (ION) is a midbrain nucleus most of whose neurons project to the contralateral retina (Uchiyama, 1989). In turn, isthmo-optic (IO) neurons receive projections from the tecto-IO neurons in the ipsilateral optic tectum (Uchiyama et al., 1996). Thus, the efferent limb of the centrifugal visual system (CVS) in birds consists of three neurons in series, tecto-IO neurons in the optic tectum, axonal projections of IO neurons to the retina, and IO target cells (IOTCs) in the retina (see Fig. 1). Therefore, the IOTCs can be regarded as the output cells of the CVS. Ramón y Cajal described the association amacrine cells as the targets of retinopetal fibers (Ramón y Cajal, 1895, 1995), although he was not aware that the ION is the origin of the retinopetal fibers. The association amacrine cells of Cajal are cells that have no dendrite extending into the inner plexiform layer (IPL) but an axon that runs and terminates in the IPL, so the term association amacrine cells actually violates the original definition of amacrine cells by Ramón y Cajal himself (that is, “cells without axons”). Although many studies have presented supportive evidence for the association amacrine cells as the target of the IO projection, the methods used in much of this © 2004 WILEY-LISS, INC. research were collective axonal tracing, histochemical, or immunohistochemical methods (Catsicas et al., 1987; Morgan et al., 1994; Nickla et al., 1994; Fischer and Stell, 1999). Therefore, because of the technical limitations, the overall morphology of individual IOTCs, including axonal courses and terminal fields, has remained obscure. One of the authors of this paper along with colleagues previously tried to label individual axons and terminals of the IOTCs Grant sponsor: Japan Society for the Promotion of Science; Grant number: 15500199 (H.U.); Grant sponsor: the Japan Science and Technology Corporation/PRESTO Program (H.U.); Grant sponsor: Kagoshima University. Kenji Aoki’s current address is Computing and Communication Center, Kagoshima University, Kagoshima 890-0065, Japan. *Correspondence to: Hiroyuki Uchiyama, Department of Information and Computer Science, Faculty of Engineering, Kagoshima University, Korimoto 1-21-40, Kagoshima 890-0065, Japan. E-mail [email protected] Received 8 October 2003; Revised 20 February 2004; Accepted 26 April 2004 DOI 10.1002/cne.20225 Published online in Wiley InterScience (www.interscience.wiley.com). RETINAL TARGET CELLS OF THE ISTHMO-OPTIC PROJECTION by intracellular injection of Lucifer yellow in lightly fixed preparations, but the group was able to label the only proximal portions of the axons (Uchiyama et al., 1995). Thus, the description on the association amacrine cells or the IOTCs by Ramón y Cajal has been unconfirmed or unchallenged by other researchers for over 100 years. In the present study, we have succeeded in labeling the entire axonal courses and terminal fields of individual IOTCs by means of intracellular injection of neurobiotin in unfixed preparations. Thus, this is the first quantitative report on the entire morphology of individual IOTCs, ac- 147 complished by means of modern neuroanatomical methods. Single IO terminals contact single target neurons (Uchiyama and Ito, 1993; Uchiyama et al., 1995). The tecto-IO projection is organized in the same way (Uchiyama et al., 1996). The tecto-IO neurons are distributed widely in the optic tectum and have nonoverlapping, spatially restricted dendrites, indicating that the input to the system is topographically restricted, including the retinal origin. Thus the CVS seems to be organized strictly topographically and, as a system, covers most of the visual field (Uchiyama et al., 1998; Uchiyama, 1999). However, the IOTCs are concentrated in the ventral half of the retina (Uchiyama and Ito, 1993; Morgan et al., 1994; Nickla et al., 1994; Uchiyama et al., 1995; Fischer and Stell, 1999). The apparent displacement between input and output has evoked bold speculations on function of the CVS (Holden, 1990; Clarke et al., 1996). Topography of axonal courses and terminal fields of individual IOTCs revealed in the present study may present clues to the solution of this puzzle. MATERIALS AND METHODS Nineteen Japanese quail (Coturnix japonica) were used in this study. Animals were treated in accordance with the animal usage guidelines of the Society for Neuroscience, and the experimental protocols were approved by the local Committee for Animal Welfare of Kagoshima University. Anterograde labeling of the IO terminals Fig. 1. Diagram of the centrifugal visual system (CVS) in birds. The CVS consists of three neurons in series: 1, tecto-IO neuron in the tectum; 2, IO neuron; and 3, IOTC in the retina. IOTr, isthmo-optic tract; ON, optic nerve; Otr, optic tract; TITr, tecto-isthmal tract. Fig. 2. Fluorescence micrographs of IO terminals (A,B) and an IOTC (B,C). This is the same cell as shown in Figure 3. Arrowheads in A and B indicate the IO terminals, and the arrow in B indicates a round cell body of the IOTC that is contacted by an IO terminal. The focal plane of C is slightly shallower than the planes of A and B, so the cell body is seen as a halo in C. Arrowheads in C indicate a fine axon Animals were anesthetized with intramusclular injection of ketamine hydrochloride (3 mg/100 g body weight) and xylazine hydrochloride (4.6 mg/100 g body weight) and placed in a stereotaxic apparatus. All pressure points were anesthetized with lidocaine jelly. The skin was incised over the skull with the animals under local anesthesia, and a portion of the skull was opened with a dental drill. A 10-l Hamilton syringe whose needle tip had been extending from the soma of the IOTC. IO terminals are labeled with Fluoro-Ruby (A,B), and the IOTC is labeled with Lucifer yellow (B,C). Cone-shaped light areas on the left in B and C are caused by a micropipette containing Lucifer yellow, which is out of the focal plane. Scale bar ⫽ 50 m. 148 Fig. 3. Brightfield micrographs of an IOTC labeled with Neurobiotin. This is the same cell as in Figure 2. A: Entire axonal course. Because of the length, the photomontage is separated into three rows. Single and double asterisks indicate corresponding points. A proximal 500-m portion of the axon is very thin and is barely visible. In the upper left of the third row, another labeled axon and terminal belong- H. UCHIYAMA ET AL. ing to another IOTC are observed. Boxes in the first and third rows indicate the locations of B and C, respectively. B: Soma of the IOTC. There is no dendrite. C: Terminal arborizations of the axon. Numerous varicosities are observed. The terminal field is 100 –130 m in diameter. Scale bars ⫽ 1 mm in A; 50 m in B,C. RETINAL TARGET CELLS OF THE ISTHMO-OPTIC PROJECTION 149 Fig. 4. Brightfield micrograph of an IOTC labeled with Neurobiotin. This cell has a short axon. An asterisk indicates the cell body. Open arrow indicates a proximal portion of the axon. Solid arrows indicate basket-like terminations surrounding cells in the innermost row of the INL. Arrowheads indicate branching points of thin collat- eral branches. A proximal 100 –150-m portion of the axon is very thin and is barely visible. One of the basket-like terminations originates from a collateral branch and is outside the main terminal field. Scale bar ⫽ 200 m. sharpened was inserted stereotaxically into the ION from the dorsal aspect through the cerebellum. One microliter of 10% Fluoro-Ruby (Molecular Probes, Eugene, OR) in distilled water was injected (0.2 l/minute) into the ION of both sides. with a fixed stage (BX50WI; Olympus, Tokyo, Japan) equipped with a near-infrared image enhancer (Argus-20 with C2400-79; Hamamatsu Photonics). The preparations were perfused with oxygenated BSM at 25–28°C throughout injection sessions. Under direct microscopic observation through a water-immersion objective lens (⫻40) with a long working distance (3.3 mm), target cells for FluoroRuby-labeled IO terminals were penetrated with a micropippette. The tip of the micropippette was filled with 0.5-1.0% Lucifer yellow (Sigma, St. Louis, MO) and 2– 4% Neurobiotin (Vector, Burlingame, CA) in distilled water, and the rest of the micropippette was filled with 3 M LiCl. Membrane potentials were monitored with amplification and display equipment (Axoclamp 2B and pClamp; Axon Instruments, Foster City, CA). The IO terminals are calyceal in galliform birds (Uchiyama and Ito, 1993; Uchiyama et al., 1995; Fischer and Stell, 1999). In particular, quail IO terminals have an “acorn-cup-like” appearance and appear to grasp the somatic base of single target cells in the innermost region of Intracellular tracer injection into the IOTCs After 2–3-day survivals, animals were decapitated and enucleated. Eye cups were cut along the optic nerve disc and its dorsally projected line and separated into a larger nasal piece and a smaller temporal piece. Retinas were isolated in balanced salt medium (BSM) containing 124 mM NaCl, 5 mM KCl, 2 mM MgCl2, 1.25 mM NaH2PO4, 22 mM NaHCO3, 20 mM glucose (slightly modified from Chen et al., 1998). The medium was saturated with 95% O2 and 5% CO2, and the pH of the BSM was 7.4. The isolated retinas were flat mounted with vitreous side up on an agar bed. The preparations were placed in a recording chamber (Warner Instrument) on the stage of a upright microscope 150 H. UCHIYAMA ET AL. Fig. 5. Brightfield micrographs of terminal arborizations of an IOTC labeled with Neurobiotin. The two micrographs were taken at two different focal depths: outermost IPL (A) and innermost INL (B). Arrows in B indicate basket-like terminations surrounding cells in the innermost row of the INL. The terminal field is 200 –280 m in diameter. Scale bar ⫽ 100 m. the inner nuclear layer (INL; Uchiyama and Ito, 1993; Uchiyama et al., 1995; Fig. 2A). Therefore, quail IOTCs can be easily identified under direct microscopic observation. Iontophoretic injection of Lucifer yellow by negative current (–2.5 nA, 3 seconds, 0.25 Hz) for a short period after electrode penetration confirmed that the target cells were contacted by Fluoro-Ruby-labeled terminals (Fig. 2B) and made their proximal axon visible (Fig. 2C). Two fluorescent cubes with which the microscope was equipped were used for the observation and injection processes (UMWBV for Lucifer yellow and U-MWIG for Fluoro-Ruby). Once contact had been confirmed, the terminals and target cells were photographed with a high-resolution digital camera (2,560 ⫻ 1,920 pixels; DXC-S500; Sony, Tokyo, Japan) attached to the microscope, and Neurobiotin was injected iontophoretically by pluses of positive current (0.3–3.0 nA, 150 msec, 3.3 Hz) for 15– 60 minutes. Histology Two to five hours after the injection was finished, retinas were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer for 1 hour at room temperature (20 –25°C) or for 8 –10 hours at 4°C. After several washes in phosphatebuffered saline (PBS; pH 7.4), retinae were incubated in horseradish peroxidase (HRP)-labeled streptavidin solution (Histofine, Nichirei, Japan) including 0.3% Triton-X for 3–10 hours at room temperature, and HRP was visualized by conventional diaminobenzidine (DAB) histochemistry enhanced with 0.04 – 0.05% nickel ammonium sulfate. After several washes in PBS, retinas were mounted on glass slides with 0.5% gelatin in water and coverslipped with a water-soluble mounting medium (Aqua-Poly/Mount; Polysciences, Warrington, PA). Entire axonal courses and terminal fields of Neurobiotin-labeled IOTCs were plotted with a camera lucida. Long and short axes of the terminal field contours were measured and averaged. Digital photographs were processed in Adobe Photoshop 5.5 on Macintosh computers. RESULTS In the present study, 31 IOTCs in 17 retinas from 12 birds were labeled with Neurobiotin, 24 of which were successfully labeled up to their axon terminals. Somata of the IOTCs were about 10 m in diameter and lay in the innermost zone of the INL in the ventral half of the retina (Uchiyama and Ito, 1993; Uchiyama et al., 1995). The IOTCs have no dendrites, although rarely a short process was observed extending from the soma. An axon extended from the base of soma (Figs. 2C, 3A,B, 4). The proximal portion of the axon was very thin, and so was hardly visible in many cases, suggesting that its diameter could be finer than the optical limit (0.2 m). Distal to this bottleneck, the axons were uniform in thickness (diameter 1–1.5 m). The axons ran straight in the IPL for up to 6 mm, but most of them made at least one sharp turn before terminating. Occasionally, very thin collateral branches that might be developmental residues were observed at turning points (Fig. 4). Some of them were directed toward the vicinity of the terminal field of their main branches. The axons terminated densely in a confined round or elliptical area of the outermost zone of the IPL (Figs. 3C, 4, 5). Terminal branches had numerous varicosities, and, in some cases, the varicosities surrounded somata of a few relatively small pyriform cells in the innermost row of the INL, forming basket-like terminations (Figs. 4, 5B). The terminal fields ranged from 90 to 290 m in diameter (180 ⫾ 50 m; n ⫽ 19), which corresponds to 1–3° of visual angle. Displacement between soma and terminal field ranged from 0.5 to 6.1 mm (Fig. 6), averaging 3.1 mm (⫾1.7 mm; n ⫽ 24), which corresponds to 30 – 40° of visual angle. Whereas the IOTCs with a short axon (no more than 1 mm long; n ⫽ 5) projected either horizontally or ventrally, the IOTCs with a longer axon (more than 2 mm; n ⫽ 19) projected only dorsally. Terminals were distributed widely in not only the ventral half but also the dorsal half of the RETINAL TARGET CELLS OF THE ISTHMO-OPTIC PROJECTION Fig. 6. A: Camera lucida drawings of entire axonal courses of the IOTCs. Data for 24 IOTCs from 13 retinae are assembled on a left whole retina (nasal and temporal pieces). The drawings of somata and axons were placed at the appropriate coordinates when they were transferred into the summary drawing, and they were reversed leftto-right if they came from a right retina. An asterisk indicates the upper end of the optic nerve disc. Dots indicate somal locations, and circles indicate terminal fields. Crossed, gray, and black dots indicate 151 somata of the IOTCs with a short (distance between soma and terminal field is 0.5–1.1 mm), medium-sized (2.0 –3.1 mm), or long (3.6 – 6.1 mm) axon, respectively. Arrowheads indicate the IOTCs shown in Figures 2–5. B–D: Relationship between locations of somata (black dots in light gray areas) and terminal fields (white ovals in dark gray areas). IOTCs with a short (B), medium-sized (C), or long (D) axons. One cell with a medium axon took an atypical axonal course (indicated by a gray dot; C). Scale bars ⫽ 5 mm in A; 5 mm in D (applies to B–D). 152 H. UCHIYAMA ET AL. retina. A loose relationship was observed between location of the soma and the terminal field (Fig. 6B–D). DISCUSSION The IOTCs may be the same as the association amacrine cells described by Ramón y Cajal (1895, 1995), the type II proprioretinal cells described by Catsicas and colleagues (1987), and the parvalbumin- and nitric oxide synthase-immunoreactive cells described by Fischer and Stell (1999). Particularly, the association amacrine cells of Ramón y Cajal show a close resemblance to the IOTCs in their terminal morphology (Ramón y Cajal, 1895, 1995). However, the peculiar morphology of the IOTCs indicates that they do not fit the standard definition of amacrine cells, even though recent studies reporting many types of axon-bearing amacrine cells have expanded the definition of amacrine cells (Dacey, 1989; Sterling, 1998; Volgyi et al., 2001). Dendrites of neurons are generally sites for integration of input from more than one source. On the other hand, the dendrites of IOTCs are markedly reduced in prominence, and the somata of IOTCs are enveloped by the large IO terminals. Therefore, it is likely that the IOTCs function as “slave” neurons whose output is controlled exclusively by the IO neurons, although the present study does not rule out minor inputs to IOTCs from other sources. The IOTCs terminate in a confined area that corresponds to 1–3° of visual angle. This size is comparable to the receptive field size of the IO neurons (Holden and Powell, 1972; Miles, 1972a; Uchiyama et al., 1998). Electrical stimulation of the ION enhances visual responses of the retinal ganglion cells (Miles, 1972b; Uchiyama and Barlow, 1994), whereas cooling of the ION does the opposite (Miles, 1972c; Pearlman and Hughes, 1976). Therefore, single IOTCs may enhance visual response of the ganglion cells locally, in a confined area. Insofar as the IOTCs terminate distally in the IPL, they must not directly contact the ganglion cells, and indirect excitatory pathway from the IOTCs to the ganglion cells may relay the enhancement of visual response, probably via excitatory amacrine cells. Although the exact locations of cells in the retina generally determine the topographic representation of neuronal information, the IOTC soma is an exception, because its topographical representation is determined by the IO terminal that contacts it. Although the IOTCs are concentrated in the ventral half of the retina, the IOTCs may send their axon to an area determined by the IO terminals. Li and colleagues (1998) reported that microanesthesia within the ION reduces visual responses of only tectal neurons whose receptive fields overlap with, or are close to, those of IO neurons whose activity is blocked by the microanesthesia. Because the tectal neurons do not receive input directly from the ION, the IO neurons seem to enhance visual responses of the retinal ganglion cells whose receptive fields overlap with their own. Thus, the inputs and outputs of the CVS may be topographically well registered as a whole. Available evidence suggests that the retinopetal neurons of the CVS are excitatory and may function as enhancers of local retinal output (Uchiyama et al., 1995; Fischer and Stell, 1999; Hu et al., 2001). In groundfeeding birds, such as quail, pigeon, and chicken, each retina receives the output of about 10,000 retinopetal neu- rons. The receptive fields of the IO neurons have a wide suppressive surround, which extends over almost the entire visual field (Uchiyama et al., 1998). The appearance of an object in the visual field might activate a small number of IO neurons but might also suppress the visual responses of many other IO neurons, because the object also stimulates the suppressive fields of those neurons. Their wide suppressive fields may cover the “classical” receptive fields of many other IO neurons, so many IO neurons may compete with each other, even if their “classical” receptive fields are distant from one another. This implies that objects that simultaneously appear in the visual field would compete for the activation of CVS modules on a long-range scale (Uchiyama et al., 1998; Uchiyama, 1999). A most effective stimulus wins the competition, and responses to other stimuli are suppressed for several tenths of 1 second. Visual attention is a neural process that selects objects for further processing and/or visual orientation (Rensink, 2000). This selection process has been metaphorically expressed as an attentional spotlight (Treisman, 1986). Although it is not yet known how the selection mechanism is biologically implemented in neuronal circuits, the avian CVS is a possible candidate for biological implementation of the visual attentional selection mechanism. Insofar as the IOTCs project to areas almost throughout the retina, it is remarkable that they and their retinopetal inputs are largely confined to the ventral half of the retina. One possible reason may be developmental. For example, the IOTCs could be close together to enhance the probability that IO axons will contact them and obtain the target-derived brain-derived neurotrophic factor required for their survival (Clarke, 1992; Von Bartheld and Johnson, 2001). Alternatively, the IOTCs could be derived from precursors associated with the ventral optic fissure; this location, near the point of entry of incoming IO axons, would favor interactions of these axons with their target cells. In any case, the CVS is a highly organized and specific component of the avian visual system, which must be important for vision and survival. Its significance should be made clear by further multidisciplinary investigations. ACKNOWLEDGMENTS The authors deeply thank Dr. William K. Stell, University of Calgary, and Dr. Robert B. Barlow Jr., State University of New York–Upstate Medical University, for valuable comments on the article. Dr. Stell also carefully edited the English and kindly provided the paper of Ramón y Cajal in Spanish and in English translation. Taichi Kamihoriuchi assisted the authors in experimental procedures. LITERATURE CITED Catsicas S, Catsicas M, Clarke PG. 1987. Long-distance intraretinal connections in birds. Nature 326:186 –187. Chen Q, Olney JW, Lukasiewicz PD, Almli T, Romano C. 1998. Fenamates protect neurons against ischemic and excitotoxic injury in chick embryo retina. Neurosci Lett 242:163–166. Clarke PG. 1992. Neuron death in the developing avian isthmooptic nucleus, and its relation to the establishment of functional circuitry. J Neurobiol 23:1140 –1158. Clarke PG, Gyger M, Catsicas S. 1996. A centrifugally controlled circuit in RETINAL TARGET CELLS OF THE ISTHMO-OPTIC PROJECTION the avian retina and its possible role in visual attention switching. Vis Neurosci 13:1043–1048. Dacey DM. 1989. Axon-bearing amacrine cells of the macaque monkey retina. J Comp Neurol 284:275–293. Fischer AJ, Stell WK. 1999. Nitric oxide synthase-containing cells in the retina, pigmented epithelium, choroid, and sclera of the chick eye. J Comp Neurol 405:1–14. Holden AL. 1990. Centrifugal pathways to the retina: which way does the “searchlight” point? Vis Neurosci 4:493– 495. Holden AL, Powell TP. 1972. The functional organization of the isthmooptic nucleus in the pigeon. J Physiol 223:419 – 447. Hu J, Li S, Xiao Q, Wang SR. 2001. Tectoisthmooptic transmission in pigeons is mediated by glutamate and nitric oxide. Brain Res Bull 54:399 – 403. Li JL, Xiao Q, Fu YX, Wang SR. 1998. Centrifugal innervation modulates visual activity of tectal cells in pigeons. Vis Neurosci 15:411– 415. Miles FA. 1972a. Centrifugal control of the avian retina. II. Receptive field properties of cells in the isthmo-optic nucleus. Brain Res 48:93–113. Miles FA. 1972b. Centrifugal control of the avian retina. III. Effects of electrical stimulation of the isthmo-optic tract on the receptive field properties of retinal ganglion cells. Brain Res 48:115–129. Miles FA. 1972c. Centrifugal control of the avian retina. IV. Effects of reversible cold block of the isthmo-optic tract on the receptive field properties of cells in the retina and isthmo-optic nucleus. Brain Res 48:131–145. Morgan IG, Miethke P, Li ZK. 1994. Is nitric oxide a transmitter of the centrifugal projection to the avian retina? Neurosci Lett 168:5–7. Nickla DL, Gottlieb MD, Marin G, Rojas X, Britto LR, Wallman J. 1994. The retinal targets of centrifugal neurons and the retinal neurons projecting to the accessory optic system. Vis Neurosci 11:401– 409. Pearlman AL, Hughes CP. 1976. Functional role of efferents to the avian retina. II. Effects of reversible cooling of the isthmo-optic nucleus. J Comp Neurol 166:123–131. 153 Ramón y Cajal S. 1895. Sobre unos corpúsculos especiales de la retina de las aves. Acta Soc Española Hist Nat 24:128 –130. Ramón y Cajal S. 1995. Histology of the nervous system of man and vertebrates. New York: Oxford University Press. Rensink RA. 2000. Seeing, sensing, and scrutinizing. Vis Res 40:1469 – 1487. Sterling P. 1998. Retina. In: Shepherd GM, editor. Synaptic organization of the brain. New York: Oxford University Press. p 205–253. Treisman A. 1986. Features and objects in visual processing. Sci Am 255:114 –125. Uchiyama H. 1989. Centrifugal pathways to the retina: influence of the optic tectum. Vis Neurosci 3:183–206. Uchiyama H. 1999. The isthmo-optic nucleus: a possible neural substrate for visual competition. Neurocomputing 26/27:565–571. Uchiyama H, Barlow RB. 1994. Centrifugal inputs enhance responses of retinal ganglion cells in the Japanese quail without changing their spatial coding properties. Vis Res 34:2189 –2194. Uchiyama H, Ito H. 1993. Target cells for the isthmo-optic fibers in the retina of the Japanese quail. Neurosci Lett 154:35–38. Uchiyama H, Ito H, Tauchi M. 1995. Retinal neurones specific for centrifugal modulation of vision. Neuroreport 6:889 – 892. Uchiyama H, Yamamoto N, Ito H. 1996. Tectal neurons that participate in centrifugal control of the quail retina: a morphological study by means of retrograde labeling with biocytin. Vis Neurosci 13:1119 –1127. Uchiyama H, Nakamura S, Imazono T. 1998. Long-range competition among the neurons projecting centrifugally to the quail retina. Vis Neurosci 15:417– 423. Volgyi B, Xin D, Amarillo Y, Bloomfield SA. 2001. Morphology and physiology of the polyaxonal amacrine cells in the rabbit retina. J Comp Neurol 440:109 –125. Von Bartheld CS, Johnson JE. 2001. Target-derived BDNF (brain-derived neurotrophic factor) is essential for the survival of developing neurons in the isthmo-optic nucleus. J Comp Neurol 433:550 –564.